Introducing Telemetry in the Notation Software

• Dec 8, 2019 - 00:25

We're all aware that a lot of improvements need to be made to Musescore 3. In the past, our focus has been directed by conversations held on the forum or feedback sent directly by users. In addition to this, we have now started to regularly user-test the app on a wide variety of musicians (thanks to Tantacrul). This is helping us get a sense of who our users are and where they are succeeding or struggling with Musescore 3.

Feedback and forum conversations help us learn about bugs, missing features and underperforming functionality; testing allows us to see what parts of the app are working well and what parts are invisible to the user. However, there are a lot of other things we want to understand. For example, we currently populate a significant portion of our top bar with large buttons for the following: New, Open, Save, Save to Cloud, Print, Undo, Redo, Zoom Controls & 'Concert Pitch'. What we don't know is the percentage of our audience who use these buttons instead of the file menu or shortcuts. Tantacrul: "From my previous experience designing Paint 3D at Microsoft, I can tell you that our most used control was the Undo button and I would expect that to be the case here." However, we have a suspicion that print, save to cloud, zoom controls and 'concert pitch' are far less used and could safely be moved elsewhere to make space for functions that get a lot more day-to-day use.

The ability to look at telemetry data to understand what aspects of the interface are being used is a vital tool in helping us prioritise effectively. Take the inspector, for example. We all know that it needs to be rearranged to provide better contextual options but how should we go about doing that? Our first problem is trying to determine the 'value' of a certain functionalities over others. If we determine that a feature is niche and demote it to a secondary or 'advanced' menu, there are inevitably going to be complaints. The problem then is determining whether those complaints represent 0.1% of users or 10%? Telemetry can help us answer these types of questions.

Telemetry would also help us determine whether new designs are performing well. At the moment, there's an open question about how many people have used the new customisation features in the Palettes panel and whether it has made much a difference to the experience of first-time users. If we suspected that aspects of the new design are suffering due to poor communication, we could track small experiments to validate that suspicion. In some cases, a big impact can result from very simple changes like rewording a phrase or making an icon less ambiguous.

High level telemetry

Another goal of telemetry is to track why Musescore is being used. For example: how many of our users open Musescore to complete quick musical tasks, like converting one file format to another? How many use it for large-scale orchestration and how many export parts? We'd like to know how much time is spent in the app too: how many sessions are under 30 seconds and how many are over two hours? How many sessions include an export to a midi file? High level telemetry should help us measure retention and stability more reliably too. It would be very useful to know how many people who have downloaded the app in the last 6 months are still using it? What percentage of sessions end in a crash?

The dangers of relying on telemetry

Despite everything mentioned above, I think it's important to state here that we do not intend to rely on telemetry to make decisions. Telemetry can tell you that something isn't being clicked on but it can't tell you whether the problem is down to bad design or just general disinterest. All telemetry can do is provide the smoke. Feedback and user-testing will show us the fire.

Respecting privacy & being transparent

We are now looking into systems that will allow us to track how people interact with Musescore without collecting any personal information and we think the most open way to do this is to ask permission from users (probably via a dialog box) when they install a new version.

This is our thinking so far and we'd like to open it up to the community to get your thoughts. There have been plenty of open source projects that have collected telemetry in the past. The ones that have been relatively successful (Citra, TimeScale DB, etc.) were open and transparent about the process. However, we're aware that there can be a resistance to information collection of any kind and we'd also like to hear the arguments for not doing it, too.

Снимок экрана от 2019-12-05 08-54-21.png


Comments

Telemetry is a bad idea.
In any case spyware rumors come out.
I don't approve. In fact, in this case, I prevent this software from accessing the firewall through the Internet.

You can do a survey instead.

Don't produce nonsense ideas, ask users instead, they say what you need.

Who said the "Concert Pitch" button was rarely used?
Someone who wasn't a musician must have said that.

In reply to by Ziya Mete Demircan

As usual, Ziya has hit this nail on the head. This is spying. This will destroy the software's reputation. This makes me feel ill. I wish the mac had a way to prevent apps from accessing the internet at all (I'd gladly upload from the site as in the old days). Or build a build that did not spy. Or write my own editor or go back to pencil and paper at the organ and ink pen for good copy, as I did for decades.

In reply to by Anatoly-os

Because many users would reject it as an unacceptable intrusion on their privacy, and you'll get articles written that list this app among apps that spy. Often there is not a choice, and you'll get more information if you make it not a choice. When I use the new feature in MacOS to say that I trust it enough to have access to my folders, I am making a statement of faith. Spying would erode that faith. Spying is user-hostile, not user-friendly.

In reply to by BSG

The codebase is open. Anyone can check whether the software send some data if users declined the use of telemetry. There will be a dialog asking whether you want to help us improve the software. If a user chooses "No", we won't collect any data.
Why would MuseScore be listed as a spy software?

In reply to by Tantacrul

Agreed! It's complete nonsense to describe a feature gated by a crystal clear opt-in dialog as spying. I'm pretty astonished at some of the resistance shown here. Perhaps people somehow failed to understand the nature of the opt-in? Although even that was very clear from the screenshot.

I also think it's significant over-paranoia to assume that anything more than a tiny fraction of the community would opt-in and then somehow deliberately create fake data. I wonder what these people are expecting - sabotage from rivals in the Sibelius / Dorico camps?! Sounds very far-fetched to me, and I can't think of any other sensible motivation for such attacks.

I think it's a great idea. Keep up the awesome work!

In reply to by Ziya Mete Demircan

What do you mean by "do a survey instead"? Do we need to ask people about the approximate duration of the sessions, the frequency of using buttons and different features of the software? What about crash free? How would we know the number of users that install software and stop using it for some reason?

We know almost nothing about the way people use the software. We know a lot from the ones on the forum and in the community. But we loose the ones who tried to use the software, but stopped without writing anything here. User testing shows different patterns of the behaviour. We see how new users complete simple tasks. And we see that the software is overcomplicated for new users.

The question is whether we want to keep the software in untouchable caste or make it easily accessible for everyone who wants to write music.

PS. Regarding "Concert Pitch", we don't know exactly, because we don't have Telemetry in the editor. It might be No 1 feature for you, but not for 99% of users. Again, we don't know.

In reply to by Anatoly-os

Create a branch of the software that includes telemetry. Then do a randomized email to some portion of the userbase, requesting if they would like to help with the development of the software by using a version (for a period of time) with some telemetry to measure how they use some features. Have this done a while after every major release.

Have them run this version of the software for a week, recording data. Delete the data within a month of the telemetry collection period ending, recording only overall statistics.

That keeps it sufficiently "survey-like" and doesn't require bundling the main software with toxic telemetry.

In reply to by LuuBluum

We will show the dialog asking whether users want to improve software by sending anonymous data. If a user chooses "No", we won't collect any data. Anyone can check this since the codebase is open. Why would we need to make a separate bundle? If you don't want to send data, just say "No".

In reply to by Anatoly-os

I'm sure that local time signatures, plugins, horizontal frames, and figured bass are used by "less that 1% of the users", but that's me -- I couldn't live without them. The English (or Russian, let alone Mandarin) dictionaries contain mostly words that are only used by 1% of the users, yet great writers would be at a loss were they "trimmed out.".

In reply to by BSG

We will never remove the functionality.
Good example! Schools never teach students to use unpopular words. The words are hidden until you find them somehow in a poem or a written story. But MuseScore might use those unpopular words on the main screen. We don't know exactly.
MuseScore should become a software for everyone who wants to write music, not for the ones who read dozens of articles and the book.

In reply to by BSG

It is not about of taking rarely used features away, but to detect them (the smoke) and then possibly move them to a less prominent place (the fire).
I for one detected the cloud icon just very recently, I used File > Save online instead since years, so I could certainly live with it being removed from the basic or even the advanced workspace, to make room for some feature that is more commonly used.

In reply to by Anatoly-os

I don't have any issue with the current menus and toolbars ;-) If I had, I'd know how to fix it.
I am using the cloud icon, now that I had detected it, but would have no problem giving up on it for something that is more important, to others.

I do, however, like to stay as close as possible to the default settings, just to not have to change too many things if and when a factory reset is needed.

My name is Zöe, I am a mermaid and I live in Sparkleville - that's what you'll get from many users if you install telemetry. Most users will simply turn it off but a few % will actively seek to overwhelm the sytem with false data. An even smaller number (say 1:10000) will resent the idea so much that they will track down the initiater of the idea.

Enjoy the unseasonably mild weather tonight, the light rain and light winds, getting slightly brisker.

In reply to by mirabilos

As far as I know Debian also has its own kind of telemetry, also completely optional to a user and with similar goals declared.

But the question of disabling certain features in software packages is something to be decided by their maintainers, so if it is required by Debian to disable telemetry in software distributed by the project then it is indeed completely fine to do so.

In reply to by dmitrio95

Indeed. The method of implementation Anatoly described elsewhere (modal dialogue box at startup, explaining clearly what’ll be collected, so users can opt-in) is good and might even be suitable for Debian (I’d have to check), as long as it doesn’t connect anywhere without prior consent (we had to remove the Start Centre from MuseScore in Debian for that reason, it connects to the website, and the website embeds external trackers like Google APIs, Google Analytics and yandex.ru). I can’t promise that yet, but it sounds very acceptable.

I personally would not object for such data to be collected by a piece of sw that I have used and trusted for many years, as long as the items collected are clearly enumerated and kept under my control. As for new-comers, I suspect things can be quite different, though.

In reply to by HosAdeeb

I completely agree. In the same way that I don't sign anything without carefully reading it, I would not engage the telemetry without keeping a copy of everything it sends "home". Offering this possibility I think will increase the engagment of many cautious (or paranoid...) people.

If you have doubts on the general usefulness of a button in the menu...
Then the answer is to make the software more customisable, NOT to track users to adapt it to the 75% of users allowing telemetry and deceiving all the other ones.

In reply to by frfancha

Current situation is that 90% of users should adapt to the needs of 10% of the audience. If we can change the situation and meet the requirements of 80% of users, and let other 20% ones customise the interface, I would say, this would make MuseScore at least twice more popular than now.

I don't like the idea. But as long as I can switch it off, I don't care too much (this is what I do to any other software that uses this too). How, why, when and where I use any software is my business alone.

  • If you do this, make sure the dialog that asks for perimissions very clearly states what this is about, why, how, where and by whom the data gets used.
  • Make sure that this setting/decision can get changed later, and without needing a factory reset or be over-complicated (like don't hide it in the Advanced settings)
  • Make sure the dialog buttons are not "OK" and "Cancel", esp. not the latter, as that would cause missunderstandings ("Cancel? The installation?"), but rather "Yes" and "No"
  • Best make "No" the default, 2nd best (and least IMHO) don't use any default, forcing the user to make a decision rather allowing to "Return; Return; ..." over it

I really do appreciate you bring up this discission. And I can see where that data might be of value.

I (vaguely) remember a few softwares already doing that, and I'm OK with it.

As long as I can view the source code and make sure nothing other than collecting non-private usage information is going on :) Not that I don't trust you, it's just "for safety reasons". If it's necessary, we can probably also write a report explaining basically what we change and if users don't have 100% trust, tell them which commit(s) did this and what does the code mean (if not easy to understand).

In reply to by Howard-C

There is a problem with this: not everyone knows how to read the source code, and even if they know, it is difficult to detect exactly where and what information is really sent. Even experienced programmers can fail detecting all that can be happening when a complex piece of software is running (otherwise there would be no bugs, ever!). Another concern would be the possibility of some "third party" malware attack which could take advantage of the existence of that socket --to call it some way-- to deviate information to an unknown destination.

As telemetry to better understand usage is legitimate, I can thank you to have open this post for each people can give it's ideas.
To open this thread broader, I have open an post in the French forum, so you will be able to collect more thinkings.
I will relay here the thinkings of French people who have dificulty to speak English.
The post in French is here : https://musescore.org/fr/node/298284

In reply to by pskunt

Thanks for this post.
I think telemetry, even we can turn it off by choice, will give a bad impressive reputation, even the goal is acceptable.
Dont' forget the numerous instructors who conduct workshops all over the world and who know very well the way of final users, beginners or professional, use the software. A survey would be a good feedback IMHO.

It seems to me that the only way to find out how the software is used is to somehow track it. You could do a survey, but that seems hit an miss, to me. I have yet to use the Inspector, but I know it is useful. I use the keyboard for "undo", but I forget the short-cut for "redo". I undo a lot.
There are many things a survey could ask about that I might answer a certain way, forgetting that I really do something different in practice.
All we hear about on the forum is what people are having trouble with, and that's valuable. But we seldom hear about what they like, or use the most. Or even about some little thing that they wish was different but don't think important enough to mention. I use MuseScore to compose as a hobby. Arranging or things like that are of little interest to me. I am interested in playback and not so much being able to produce a score for real players. I suspect I am a minority. But MuseScore is becoming more appealing to me.
I have no problem with data on how I use MuseScore being collected.

Hello to all of you
First of all, I would like to greet MuseScore, who has the courage to put this idea on the table. Others, and among the most powerful, do not bother with such preliminary precautions: they install telemetry without saying it and by default. It is the temptation of the moment and no one escapes it.
It is for example very, very present in Windows 10 and if you want to avoid it you have to install special programs to counteract its effects.
But let's ask the question differently: Why do system manufacturers, program publishers need so much feedback? Quite simply to avoid disappearing, swept away by more informed and responsive competition. How many programs that we liked have disappeared because they have not kept pace with changing demands and technologies.
Let's admit that we are more and more demanding on performance, possibilities, options, settings, customization etc... etc...
We cannot ask developers to be always at the forefront and at the same time deny them the means to do so.
So that being said, here is my answer on the substance.
MuseScore is a program I use on a daily basis, and I greatly appreciate its capabilities, its free nature, its constant development and, finally, its very responsive community.
I am here in a climate of trustworthy and I would therefore accept, since it is also asked to me, that my behaviour on this program be observed and analyzed. The "yes" or "no" option being clearly visible and disengageable at my request.

In reply to by BSG

Sorry, but (and with the upmost respect) the idea that what Anatoly is proposing is spying is as ridiculous (and dangerous) as not hooking up Windows to the internet. As with any piece of equipment (Mac, PC, or Linux), or software, (MuseScore or Cubebase) you have to learn how to use it.

In reply to by bobjp

I've been programming computers, mainly professionally, for a half a century and know more than a little bit about operating systems (which I helped design), networks, and security, and "how to use computers". Why don't you tell me why keeping a single-application computer permanently offline makes it vulnerable to attack? If you prefer to be spied on, go for it. I don't, and apparently the designers of Debian don't either.

In reply to by BSG

Then you know there are more secure distros than Debian.
But again, I'm just not sure how collecting data on usage of aspects of a program is spying.

I assume you use a browser on your computer to post here.

In reply to by bobjp

I do, and it has queried me about whether I would tolerate its proposed invasion, and the answer was "no". Whether it's "spying" or not is up to the person being raided to determine, not the spy. A long time ago, I worked on mainframe time-sharing systems that had "metering". I even proposed and implemented some of it. But to me, "personal computer" means something different.

In reply to by BSG

Your comments are insightful… to say it with contemporary words, you rock! ☻

I’m not yet half a century old, but I’m more than my arms deep in the innards of operating systems often enough as well… might be nice to meet you at FOSDEM or so, if there weren’t the continental divide…

I have for many years wished we had that sort of data. It could help design better defaults for toolbars and other aspects of workspaces, but also help direct development (like, knowing the percentage of note input done by keybaord vs mouse could be an interesting stat). But of course, I also was sensitive to the idea that it wouldn't be very popular, plus I had no idea how to actually implement it anyhow.

So I think it's a good idea worth exploring, and I too am glad to see the open discussion of the topic, hopefully there will be some sort of consensus on the best way to make it happen without alienating users.

And as I think about it, I can also see value in some survey type questions.
Some of my thoughts would be about the reverb controls. Not enough documentation for my taste. As detailed as the reverb controls are (maybe too much control), I can't have different reverb (not just amounts) on different instruments. If there is, it's not in the manual.
Or how about making drum kits like any other instrument in the mixer and not something that pops up at the bottom of the screen. I know you guys that developed it understand it (again, mouse user, here), but it makes me not want to use drums.
Maybe using the mouse is slower (oh well), but there is more to it than speed. There is intent. Mouse input lets (or makes) me think about every note more. And again, I'm a composer not an arranger. So notes are not a given.

I may be paranoid, but...
When I hear the phrase 'telemetry in software' I immediately think 'keylogger' - which has (mostly) bad connotations wherever I read about it.

In reply to by Jm6stringer

And yet what better way for developers to stick to the demand and improve the program than by closely analyzing the behavior of users who agree to be watched for the benefit of the entire community?
There are two parts here, developers and users. Each party should try to integrate into its reasoning the other party's questions. This is the best way to move forward, especially when you are in a climate of mutual confidence. Which is my case with the MuseScore team....

In reply to by Papibois

There is no question that having 1984-type cameras and microphones in every dwelling would be extremely valuable, and prevent a lot of crimes. Then there is the other side. No, command usage telemetry is not Big Brother cameras, but the feelings and options of those whose activity is being monitored and reported is part of that difference.

Telemetry implementation has just been merged to master branch: https://github.com/musescore/MuseScore/pull/5525. Everyone is welcome to audit the code.

Telemetry module is switched off for developer/nightly builds.

@mirabilos "BUILD_TELEMETRY_MODULE" CMake variable is responsible for adding the Telemetry library. It is "OFF" by default and is going to be "ON" in release builds. You can either switch it off in Debian packages or remove the code introduced in the PR mentioned above.

After reading through the comments, I'm not so gung-ho with the potential effectiveness of telemetrical activity. The time required to 'analyze' user-data got by live connections from people who 'come and go' could potentially be -- instead of looking at statistics without knowledge as to stature of the user -- spent moreso on examining user comments, expectations, requests and their implementations or what have you (if the same people are programmers), and participating in bug fixes.

The people who find anything awkward about M.S. or think something particular ought to be changed, in theory, will involve themselves in discussion at the forums. It often takes quite some time before a person can become familiar enough with something like MuseScore before one can conceive of a 'better way' of an implementation or interfacing. To speak of gathering data without the judgments of 'better' and 'good' but merely as a statistical analysis of usage requiring those capable of such judgments to spend time on this rather than other said activities when it's known that notation and the software related to it isn't something that comes easily seems to miss the mark 'imho', but best of luck to those making these decisions. Whatever's best for MuseScore and its users.

In reply to by worldwideweary

OTOH, often someone on the forum is unable to describe what went wrong, how they ended up where they did, or whether what they did actually caused the problem. Users are not programmers. While programmers are users, they write things to be used a certain way. Users find the darnedest ways to cross things up, from keystroke combinations to system configurations.

In reply to by worldwideweary

Telemetry isn't a replacement for feedback, it's an additional tool that helps developers and designers make decisions. For example, one of the most common bits of feedback we get is 'MuseScore is too complicated'. Moreover, users don't (and can't reasonably be expected) to provide precise examples what they consider to be complicated and what they consider to be straight-forward on a button-by-button basis. We can definitely employ strategies to combat this type of issue, but without more information, guesswork and intuition are all we have to go on. And since users also rarely tell us when something is working as expected, we run the risk of hiding something important or promoting something relatively unimportant. Telemetry can tell us about all the parts of the app that people take for granted and don't comment on, which helps redesign work to be as painless as possible.

In reply to by Tantacrul

If this implementation is to be continued, I'm then curious or looking forward to see what specifically is the "fruit" of paying attention to user activity from within the program. Hopefully those who are going to be spending time looking at that data will be transparent and manifest their decision making processes and influences based upon whatever that is rather than merely give vague statements about "being able to make appropriate decisions".

In reply to by Tantacrul

The method by which changes would be implemented based on user-interaction data with the program is by its very essence vague to someone such as myself as it currently is stated due to not knowing exactly what is inspected and what methods will be used for correlating intention and priority with the data. As is supposedly the case, many people come and go with software, and to speak of priorities based on use case of recurrency may give a false impression. For example, I rarely if ever use the more buttons in the palette area because my palettes are set up usually the way I want, but that doesn't mean the more buttons ought to not exist. Not only this, but on my machine, the palette search function when deleting all its text suffers speed slowdowns, whereas an earlier version of my particular machine had no speed issues, but this problem won't be recognized by merely pulling statistical data from something like this. I might, as a new user, if I encountered this, get iffy and quit without mentioning it on forums. The statistics would give no indication as to what happened 'mentally' for the user, nor would the user's interaction-data provide appropriate information to accurately correlate excellence with priority.

Maybe it's not vague and I'm merely being wishy-washy. Maybe you guys could program something that manifested the exact data that was captured for analysis of the user in a local data file or something? That might help the skeptics be reassured if they could see what exactly was gathered, and not only this, but also, if this continues, again, declare openly what decisions are being made based on what data. If those two were fulfilled faithfully, then the main concern would shift over to the extra computing activity required for the data capturing/sending.

In reply to by worldwideweary

I don't really understand this stance (and I dislike the idea of invasive data exfiltration intensely). Assuming that users understand what data is being collected, how it is used, and why (the general terms of privacy legislation), and actively agree to it (or opt out) it seems that the issue ends there. Saying "I would be concerned if you used this data to remove, or make less accessible, some commands that I like" seems to me inappropriate, and the sole prerogative of the designers/implementors (a team which, in this case of open source software, you are free to join). Response to design/change proposals, or response to implemented changes ("Bring back the album feature!") are highly appropriate, but, as far as I can tell, not "I don't like the idea of your collecting data which might result in design changes i don't like." Just my "two cents" (fiftieth of a semitone). I repeat, I'm in the < 1% of users to whom figured bass is central, but I've been promised it won't go away for lack of use. Leaving aside the very substantial issues of privacy, trust, and reputation, knowledge and data cannot be a bad thing.

In reply to by BSG

No. One thing I am saying, in a sense, is that this data and its interpretations should be transparent and open to the community of "MuseScore" and to its users. This place has conversations about what should and should not be implemented with manifestations of support regarding feature requests and changes in general. Why not lift the burden of making these decisions slightly away from only a particular set of people and provide the community, many of whom like to give their support and lacks thereof, with participation related to this process? Of course, not everyone gives functional reasons for their wants, so that might muddy it up a bit.

Of course, to not trust in another's competence would require a blanket statement or a more intimate knowledge of the ... team :)

In reply to by worldwideweary

Try "responsibility of making these decisions" instead of "burden". Open source does not equal democracy. People who understand the product, its past, its implementation, and have the requisite experience in software engineering "make these decisions". The developers are not a cabal of brigands and rogues who have stolen power over the product from "the people" (pardon my hyperbole; maybe an elliptical parable would be less circular). I sense you feel that the power of developers to develop is élitist and insufficiently legitimate. Open source does not equal democracy, and (in my opinion), rightly so.

I'd like Marc to weigh in.

In reply to by BSG

The power to develop and appropriately harness language melded with functionality in particular systems and the ability to interpret user activity seems to be two separate areas of activity, albeit not entirely distinct. "Those who can, do. Those who can't..." or something like that. I don't necessarily equate power with legitimacy, but I have no negative feelings about developers who have abilities to implement and synthesize functionality. Far from it. I am curious, however, as to the importance of user-data and how the team would interpret this data knowing the larger possibility of its not being bound up with well-informed intentions.

One of the main points here is the desire for full disclosure of the information, or rather data, that is being "seen" by other people, so that the user who is "being watched" knows full well what has been collected. The user should be aware of that information if he should so want it. Either way, the results will speak for themselves in the sense that updates and new versions will indicate what was decided upon, and those in turn will trigger some desire to make forum posts involving user-input for further design/implementation changes so that the vicious cycle may continue >:D

In reply to by worldwideweary

How do you feel about the .com site, whose maintainers do not share their source, their methodology, their forthcoming designs, etc., and surely meter every activity that goes on? Or Amazon, Google, Apple, etc. doing the same? Why do the implementors of musescore have a special responsibility to share the results of their tools? (Again, assuming consenting users; we are not discussing the privacy issues here).

Why do you suspect "not being bound up with well-informed intentions"? It certainly sounds like lack of trust underlies this. You just said that you do not trust a priori that the developers are competent to interpret the data resulting from such instrumentation, or use it as an input to design. Or maybe you believe in principle that such data has no place as design input to any development process (that position is extreme, but defensible).

In reply to by BSG

I'm not concerned much with the .com site, so I have no feeling about it currently. It's merely my opinion that those in relation to this project should share what specific data is observed with those who may want this information. If this equates to "special responsibility" to you, I'm flattered that my should is contrived as such.

The phrase "not being bound up with well-informed intentions" was rather in reference to the users themselves, specifically new users not making well informed decisions within MuseScore, potentially 'muddying up' the data.

In reply to by Tantacrul

To use a software you need to work on it and specialize in it.
For example, if someone tells you that Ph0t0sh0p is too complex, you can't convert it into that can be used by elementary school children, can you?

The company should not make the mistake of removing important features (some Toolbar items, Concert-Pitch, etc) based on the abundance of transactions performed by beginners.
Because the operations that are frequently used by normal or expert users will be less than these. And beginners will use them after mastering.

I hope this telemetry feature doesn't cause MuseScore to be labeled as u n w a n t e d software by some operating systems and virus-scanners. As I stated in my first message: I hope this telemetry feature does not cause MuseScore to be labeled as unwanted software by some operating systems and virus-scanners. As I mentioned in my first message: If a rumor spreads among people, there is almost no way to fix it.

In reply to by Ziya Mete Demircan

I don't think you've read the proposal properly or understand what telemetry is. In my professional career, I've never worked on an app that doesn't collect some kind of information (eg. how many times does the app crash; How many times in a session does the user click on a particular button, etc.) The proposal made it clear that this information being collected is not personal data. It's anonymous. But you seem to be confusing it Facebook-style harvesting of personal information for some reason.

The suggestion that this makes MuseScore a virus is ridiculous. Just about every app you use (and the OS themselves) collect some kind of telemetry. Moreover, the proposal made it clear that people are going to be given the choice to OPT-IN - a courtesy most other apps don't bother with.

In reply to by Tantacrul

I try to prevent the software and OS that I use from sending data in many ways. I am also using "OOSU", and I block most software's Internet access from the firewall.
I don't even allow a crash report to be sent to M1crosoft. If I have a problem, I will use the M1crosoft support channel (forum).

I read Proposal and got it. I have no problem with that.

I wrote my comments on the UI according to this and the following sentence in the quote.

  • However, there are a lot of other things we want to understand. For example, we currently populate a significant portion of our top bar with the following buttons: New, Open, Save, Save to Cloud, Print, Undo, Redo, Zoom Controls & 'Concert Pitch'. What we don't know is the percentage of our audience who use these buttons instead of the file menu or shortcuts. Tantacrul: "From my previous experience in designing Paint 3D at Microsoft, I wouldn't expect that to be the case here." However, we have a suspicion that can be used to make a lot of day-to-day use.

  • "We are now looking into systems that will allow us to track how people interact with Musescore without collecting any personal information" *

As for the possibility of rumors and unwanted software: I have stated that I have thoughts about what might happen.

My writing is very positive. And I know I'm right in my warnings. You just need to understand and evaluate in your mind, You don't have to be agree with.

(Apologies about the length.)

What kind of data you collect is the most important decision point in this question.

As a corollary, please treat the list of data you collect as the most important part of the decision.

So please do not have "etc." in that list, and do not just casually list items.

Instead, list each collected data element/item separately, and give clear description to them. A title of a few words, and a sentence or two describing the data exactly, and an explanation of how/why it will be useful down the road for making decisions. Please make these titles and descriptions easy to understand (happy to help). Please allow each of these items to be enabled/disabled separately. When new items are added in future releases, please add new permission items for them, and ask permission for those (while also allowing changing previous permissions).

Please consider an opt-in feature, where developers can reach out to me individually if they have a targeted question to a small subset of users based on telemetry. This could be either a checkbox of something like "allow developers to contact me if they have a question", and then I could enter my email address, and then if a rare to reproduce bug happens and I happen to be one user who can reproduce it, I can be contacted and involved in the process by developers, for the benefit of the community.

Please allow postponing the decision (e.g. "I am considering, please remind me in [___] days"), in which case no (new) telemetry items should be enabled by default until the user decides otherwise, and the user is shown the dialog again after the time they requested elapsed (except if in the meantime the user has gone into telemetry settings in preferences and changed some settings).

I would also suggest features towards capturing the "here and now" moment of the session, to help reproducibility: I have an issue, I want a button or menu item I could use that somehow captures my experience, and relays it to developers. During this process I want to be able to opt in - from a list of checkboxes on the screen with detailed descriptions - to providing as much or as little data, as I feel comfortable with. To save the reoccuring support questions of "what OS? exactly what version of MS (commit)? what is your configuration of X option in MS preferences?", "can you provide a score?", etc. Allow me to take screen shots or capture a short video, demonstrating the problem. This could then channel into automatically opening a bug report on the proper forum, or produce a kind of "dump" file locally on my computer than I can then manually submit in a bug report and gets automatically processed. Kinda like the current "feedback" button on proteins.

Also, a textual search bar/field at a prominent place would be nice, ideally a kind of "answer to all": searching among all menu items/commands, dialog items, setting names, etc. (enriched with synonyms - see below), as well as a locally installed reference or tutorial material. For the case I did not find what I needed, provide an option among the search results ("I did not find what I was looking for, please forward my search to musescore developers"). If enabled in the privacy settings, please allow the option of search for X online - which would search in musescore.org in manuals, forum posts, etc. and it also would capture my search term/question in a central database at musescore.org. (Or perhaps these two could be unified into the same use case.) This database can be reviewed by developers later to understand what features, terminology/labelling, use cases, etc. users tend to have issues with, and provide an organic use of oft-used synonyms to add to search. Even more pipe dream is: allow me to see a list of my issues/questions I submitted through this, so that if/when I found solutions for those, I can feed the solutions back to go together with my question, crowd-sourcing building up a knowledge base/reference manual. After some curation these could in turn be added to the local documentation, for the search function to find.

"Private mode" (half-baked ideas): if I have telemetry enabled, allow me to turn it quickly off when I don't want it, and back on when I want it. Maybe a button on the toolbar? But what if I forget to turn in back on? Maybe an option so that I can choose to "ask me each time" (e.g. each MuseScore session starts with the question again, whether to enable telemetry for that session). Or make it per document or window. Or any number of permutations of these.

In reply to by bobjp

That's not the sort of thing the telemetry info is intended or designed for, though. It's more about getting aggregate data on what sorts of actions people do, not getting a blow-by-blow analysis of some particular problem with the sort of detail that would be needed to enable developers to reproduce a bug.

Hello! I am an arranger and music editor for publishing, digital, and printing. I moved to Musescore from Sibelius some years ago. I wonder if aside from telemetry or instead of it, wouldn't it be a good thing to adopt pro-workflows and to use workspaces suitable for specific tasks or letting you arrange and save workspaces, with a UI that lets you move the element and stack them into palettes... Like a worspace when writing for guitar which has it owns challenges, or sequencing, midi cleanup, or choir arrangement... etc. Professional composers or Music Publishing experts would give better insights than in many cases, non-pros, clicking here or there going nowhere and giving unclear statistics to the telemetry system. Just my opinion.

In reply to by jeetee

Yes, I did/I am. Sorry I didn't make myself clear enough. I meant, a flexible enough UI to enhance UX... on the style of the Adobe products, i.ex: Adobe InDesign. Maybe it is possible to get something similar inside Musescore but I haven't found it. Also having to open a window to organize UI elements doesn't seem as efficient as having a "live mode" where you could reposition the elements, undock the palettes, not within the main palette but as a separate element. That way you could get more satisfying workspaces, I am going here for a live redesign of the interface for specific tasks and not merely putting the buttons you use the most more visible. Also, factory presets for workspaces just like Blender. As a graphic designer myself, I am after what I consider would make the software easy to understand and customize. Makes sense? This is only a wish list, I am pretty happy with this amazing software, sadly I am not QML proficient enough to contribute with the software myself :-(

Im in the camp that telemetry is an invasion and unnecessary at that. I don't want to have to start looking for another notation program, but I will if this is going to be the path MS follows.

In reply to by Papibois

And how many times now has Facebook, Google, Microsoft (and who knows what other companies and organizations) have said, "Oh, we won't collect data if you don't want us to," and, lo and behold, it comes out that that was a lie.

Now look, I'm not saying you folks at MS would or are doing anything wrong. What I am saying is that there's already enough telemetry going on and we don't need anymore. It's really not that hard to find other ways other than having MS phone home with what 99.9% of us would never be able to verify by checking the "open source" reference to see what's being sent.

Just produce notation software, ask question, do videos, etc., but leave the intimate connection of what I do out of it. That's all I'm saying. And finally, the option to not send information is, as I've noted, not a trustworthy offer anymore these days. Again, that's not to accuse MS of any wrong-doing. Just stay away from telemetry is what I'm saying.

In reply to by BSG

I don't understand this suggestion that the MuseScore team are generally not trustworthy. This was presented to the community before it was implemented and it can be verified by anyone. Because it's an open source project, any abuse would be exposed immediately.

There's no pathway to evil here.

In reply to by Starkman

Ultimately, you have to trust software that runs on your machine, and reputation and web-lore and the use of packet tracers etc. has everything do with with that. When an vendor announces,, "we're going to have some Send Back to Mother Ship", that trust erodes, no matter what policy is announced with it.

In reply to by Starkman

It's OK that you don't know how, the beauty of open source isn't that you personally can do things with it, but that plenty of other people can and do. By your logic, no one else should have to develop MuseScore either, and yet, we do. If there were not a dedicated group of volunteers constantly examining and improving the software, it wouldn't exist. So I think you can safely forget the possibility that somehow the 100+ people who work on the software would somehow collectively decide none of us we care about privacy at all and just stand by idly if an actual invasion of privacy were to be proposed or implemented.

In reply to by Starkman

Starkman, it seems your concerns have been answered comprehensively and I think the explanations are more than reasonable.

If you opt in, MuseScore does not collect personal info, so there's nothing 'intimate' being turned over. Any developer can right now check the code and verify this for you. If you think this explanation is a lie, then that would make for quite an exposé story, don't you think? How dumb would it be to lie about this when so many people can immediately fact check it?

To reiterate the facts:

All users are presented with a clear set of options that allow them to opt in or opt out. MuseScore does not auto-opt in.

If you choose to opt out, no telemetry will be collected at all.

Can't really see any reason to be angry about this.

In reply to by Tantacrul

Really? No, I haven't been unreasonable, and neither has anyone else here objecting to the use of telemetry. And no, I have not been unwilling to consider the explanations given. I just don't find them compelling, and neither do others here.

I expressed, as several have, my concerned (and objections) to any kind of telemetry. I apologized for my harsh statement about looking elsewhere for notation software. I used organizations that have lied in regard to their use of telemetry—expressing that regardless of the open source availability of MS, I have no knowledge how to verify any telemetry sent to MS; just avoid using it, period is my point. Further, I did not accuse MS of abuse or any wrong doing. As to "intimate" information being sent to MS, the information MS collects IS intimate—however, I said nothing about personal information being sent. As to the information that would be sent, it would be very intimate in the context I'm using the word or it would be of no use to MS.

Finally, the option to opt of enabling telemetry out misses the point: stay away from the practice, period, is the point. As I noted, there are other ways (e.g., snapshots, video recording, etc.,) to garner information for enhancing MS.

So, if that isn't "soberly" engaging the issue, then pray what is?

In reply to by Starkman

Surveys, focus group testing, and what-not are useful, and we do those also. But they end up being pretty skewed, as you can't really collect nearly as much breadth or volume of data. Those methods are mostly useful for confirming specific existing hypotheses, not providing a database of information that can be used to answer questions we haven't even necessarily formed yet. There is no other known way to collect the sort of information really needed to learn the really interesting and useful things we'd really love to learn in order to improve the software in the way we'd like. So that's the question - is your personal squeamishness worth limiting our ability to improve the software for everyone?

BTW, not sure what you think is "intimate" about knowing which button you pressed, but fine, if that's intimate you, don't opt in. That simple. Fortunately, I have little doubt that lots of other people who aren't able to help with actual development will be willing to at least help us out in this small way.

In reply to by Marc Sabatella

Marc, you're a professional musician and composer, and I trust others working on MS are as well. Surely, you and the other developers have a clue as to how a piece of notation software should work, should act, should respond, and intuitively so. The forums have expressed well what people want and what they don't want. MS just hired a fellow (per your recent Cafe video) to take MS even further. Is it really necessary, then, to delve into the arena of telemetry? If you can't get it right based on what you know, then all the telemetry in the world isn't going to help. You already know, well, from what's already been said and observed, what musicians want in a notation software. How about looking at other software to see what they're doing and encorporating what you like into MS?

On the other hand, it's virtually impossible anymore to stop the move into data collecting all the way around. So, really, it's moot to object to the use of telemetry—the opt-out being irrelevant half the time. I just see no need to get to up-close-and-personal with MS software.

In reply to by Starkman

I have my ideas, the few dozen other developers have theirs, the couple of people working on design have theirs. But this is a statistically insignificant data set. It's relying too much on what a few insiders think that leads to interfaces that are hard for beginners to figure out. Forums tell us about the people who are sufficiently engaged with the product to create an account etc. They don't tell us anything about the people who downloaded the software, played with it for 15 minutes, couldn't get anywhere, and deleted it.

Again, no, of course, telemetry data isn't necessary. Improving MuseScore is not necessary. MuseScore is not necessary. Computers aren't necessary. Music isn't necessary. But if you think telemetry data it isn't valuable, then I can only guess you have no experience in software design on this scale. I'll make you a deal, though: I won't try to tell you what might be valuable for you in your life, and you don't try to me what's valuable to me in mine.

I have no idea what you mean about it being impossible to opt out, though. It's as easy as pressing one button, I promise. And it's more than a little insulting to have people accusing us of doing anything but exactly what we say we are doing. Again, I'll make you a deal: don't question our motives or integrity, and I won't question yours.

In reply to by Marc Sabatella

Marc, sir.
Apologies if you thought, in the least, that I was insulting, questioning MS integrity or anything like that. I've said, more than a few times, my issue is not a distrust with MS, leaked data, this, that or the other. My issue is with telemetry in general: I believe that only in very few circumstances should it be used.

My point with regard to not being able to opt out was not with regard to MS but just the whole technological movement of computers in general; one can't get away from telemetry, often even when opting out. Though now, after re-reading my post, I can see how such a statement could have been interpreted as a direct attack against MS, but it wasn't, not at all! I have full faith in MS. It was just a generic, "Oh well, maybe I should just let it (the issue of telemetry) go and be done with it, because it's here and that's that."

Finally, I FULLY respect everyone's right to their opinion about this issue. In the rush of time, my posts, however, have seemingly implied otherwise, which is not the case at all. And though I will choose to opt out of the telemetry feature, I will always be willing to help out by offering suggestions, participating in the Cafe videos contributing to the forums, and what have you (once I get through your course again!), and I look forward to doing so.

Thanks very much.

In reply to by Starkman

Fair enough. But I'm curious - you say you believe that telemetry should be used "only in very few circumstances" - what would those circumstances be, if "improving the usability of the software in ways that benefit all users" is not one of them? To me that is exactly the circumstance in which it should be used.

I can't speak to "the whole technological movement of computers in general". But I will say it's unfair to blame MuseScore for what anyone else does. Let's keep this focused on the actual implementation in MuseScore, not something else it reminds you of in some other application, website, or other unrelated business. So, I can't help you with Facebook issues. But I can say that if it pleases you to not provide this data to MuseScore, then don't opt in, and the data will not be provided, period.

In reply to by Marc Sabatella

Marc,

You said, “But I will say it's unfair to blame MuseScore for what anyone else does.” Never have I blamed MuseScore for anything or for what anyone else does. (I don’t even know how I could blame MuseScore in those ways.) Sorry you thought that.

As to times when telemetry is important:
* Software for home security: alarm, camera and other information can (selectively) be sent to the software company in order to alert the home owner of a (potential) dangerous intrusion.
* Software for reporting unauthorized use of one’s vehicle.
* Software that monitors and reports to a company one’s (various) health status.
* Software for monitoring a child’s use of his/her cell phone.
* Location software for tracking hikers, boaters, etc.
* Open source operating systems that receive crash reports and other selective information that can be verified by the community as keeping in check with what should be phoned home. (Microsoft uses telemetry and still there’s no way to know exactly what they are receiving, and they aren’t up front with any information about it.)
I could go on, but I hope this is enough to convey my thoughts about when it’s important to have information phoned home.

As to MuseScore, how much can the developers possibly know by the fact that I typed a quarter note, clicked a clef, added an articulation, opened and then closed the Inspector, the palette box or what have you? If I like inputting my notes by clicking, that doesn’t tell you anything except that I’m inputting by clicking. You can’t assume a reason why. If I don’t use the Inspector, you don’t know if it’s because I don’t like it, don’t understand it, or simply prefer another way to access the information.

On the forums, when it comes to problems that are difficult for a user to explain, one of the first thing people responding do is ask to have the user post a snapshot of the issue. I would contend, thus, that if Musescore developers can’t see my score and how I’m creating it and dealing with difficulties of using it, they sure can’t figure out what I’m doing wrong or what is not working correctly or why, for that matter, I’m doing what I’m doing. (To see my scores, be they simple as they may be, is out of the question, in my opinion, on principle.)

Better, in my opinion, would it be to just run some polls in the forums or on your videos: “Say, do you use the Inspector? Do you type or click your notes? Etc.? Send out a monthly email tip, add a 10-15 second video recording ability within MuseScore that will allow users to record exactly what they are doing so you can see the issue first hand.

The development of the program is far enough along now that it’s really only tweaking and adding features here and there. How users face off with MuseScore isn’t going to be solved by telemetry; again, something I think should just be avoided out of principle, in my opinion.

In reply to by Starkman

OK, you are defining telemetry a bit differently, obviously the data essential to the operation of a remote service needs to be sent by any app using that service. Telemetry in this context is more about usage data.

Anyhow, I get that you are not a software developer and are having trouble understanding how valuable this information actually is. So, I offer you another deal: I won't try to tell you what information is valuable to you in your job, and you don't try to tell me what information is valuable to me in mine.

As I've said, polls are useful too, but in a much different way, they don't provide nearly the sort of breadth of information. So here's another deal: don't tell me that X can substitute for Y in my job, and I will extend the same courtesy to you.

In reply to by Starkman

I would almost go so far as to say that surveys and the like are almost useless.

Here is what I see on the forum quite often. Someone has a problem. Something doesn't work. They post a picture and even a score. Someone tries the score and has no problem with it. There are many possibilities. Is the user not doing something properly? Is there something not right with their computer? The user has no idea. I followed one thread where the user couldn't get MuseScore to install no matter what he tried. This went on for a week. Turns out he was trying to install it to the desktop. This question might have been answered much sooner with the right data.
Sure, often the answer is in the manual. Not always. Sure MuseScore is coming along nicely. No software is ever perfect or really finished.Because it is open source and has an ongoing team working on it, I think uses need to give them all the help they need to make a better product. We live in an age where when people see something on the internet, then it must be free. MuseScore actually is free. I think we all need to do our part to make it the best notation software out there. Opt in or opt out, I don't care. Neither should anyone else. Let's get on with it. I'm happy to opt in. Maybe the Russians will be bored to tears with how many "un-do's" I do. I'm sure they already are.

In reply to by Starkman

Sorry, but you are comparing apples with giraffes. These are totally different types of data collection to Facebook/Google/Microsoft.

MuseScore does not have access to your phone number, location, home town, education, current job, photos, status updates, lists of friends, relationship status, sexual preferences, personal interests, or any other personal information. Not anything even remotely close. It even explicitly states this in the proposed dialog:

"We _do_not_ track personal data or sensitive information, such as location, source code, file names or music."

Here we are talking about things like how often you use the Undo button vs. the Edit Undo menu option vs. Control+Z, or how often your MuseScore crashes. Are you seriously saying that you are worried about this kind of data leaking? Even if you are, let's consider these two hugely unlikely scenarios:

a) The MuseScore developers deliberately decide to break their promise to keep the data anonymous.
b) The MuseScore developers or people hosting the telemetry database screw up and it gets hacked and the data leaks.

Firstly, there is zero upside / motivation for the developers to do a). There is only risk of damaging the credibility of the project if they got found out, and it would be extremely likely they would get found out because the source code is open for anyone to review. So I defy anyone to suggest a good reason why this would happen.

b) is (relatively) more likely since people do make honest mistakes, but it's still a very remote risk, because the leaked data would be utterly useless to anyone other than the developers. Noone else cares about this data! Not even the Dorico developers would care how often MuseScore users hit the Undo icon vs. Control+Z.

So this is worlds apart from the data Facebook/Google/Microsoft hold, which are goldmines of highly sensitive and potentially very damaging personal data. I'll happily agree that it's a risk trusting those big data giants with your personal data, but please don't have a knee-jerk reaction which prevents honest open source projects from progressing by politely asking users if they can collect some totally harmless non-personal data.

Telemetry is a great idea for MuseScore, and given how clearly and respectfully it has been implemented, I fully support it. Thanks and kudos to Anatoly-os, Tantacrul, and all the developers - keep up the great work, and I look forward to seeing the UI improve even further as a result!

In reply to by adam.spiers

Please, just stick to the issue that we who object to telemetry are raising: the use of telemetry, period, is, in our opinion, unnecessary and a step in the wrong direction. You are confusing the the matter of what data is sent verses data, period, being sent.

You said, "Here we are talking about things like how often you use the Undo button vs. the Edit Undo menu option vs. Control+Z, or how often your MuseScore crashes. Are you seriously saying that you are worried about this kind of data leaking?"

First, it's unncessary for MS to collect this data. Just make the software, interact with people on forums, videos, etc., to enhance the product.

Second, I don't like the idea of it even being suggested that MS can collect how often I use the Undo button vs. the Edit Undo menu option vs. Control+Z, etc. (And, again, opting out, in my opinion, misses the point.)

As to the matter of leaking information and what have you: irrelevant. I never brought that up. I simply don't like the idea of what I'm doing even potentially being broadcasted.

In reply to by Starkman

I think people have come to expect that local apps are their property once purchased or licensed, and work for them and them alone. There is a reason SOME people use local applications as opposed to web applications for SOME things. People, certainly I, am of the expectation that I can run an app locally and nothing about my activity at all, repeat, nothing about my activity at all is being reported to anyone. I think that's part of the local-app vs web distinction that people have come to expect, and ought be a bright red line. Yes, I can spell "opt out".

In reply to by BSG

I would say there are many reasons why any given person might choose a local app. Concerns about transmission of data are probably about 57th on the list overall. Also, not sure why anyone would "expect" local apps to not transmit any information when pretty clearly, a very large number of them do. So seeing a dialog asking about this should hardly come as any sort of surprise., Are these people also surprised every morning when their toast is crunchy?

But again, if you don't want to help us, that's fine, just don't opt in.

In reply to by Starkman

Of course it's not necessary. Nothing is necessary except to eat and breathe. MuseScore itself isn't necessary, new features or design improvements aren't necessary. But I think most people agree that MuseScore is valuable, and having new version with new improvements is valuable too. Similarly, it's extremely valuable to developers to have the sort of data that will allow us to better improve the software. Not having it will continue to limit us to only being able to collect that small and skewed amount of data that can be collected through other means.

So again, if you don't want us to provide us with the sort information that can help us better improve the software, fine, don't opt in. I have confidence that plenty of other people will pick up the slack and help us help you and everyone else by providing us with information that can enable us to better decisions on how to best improve the software. So you'll benefit whether you participate or not.

I suppose we could consider making some new improvements only available to the people who opted in and helped provide the data that guided the development (just kidding!!!!!!)

In reply to by Starkman

"Please, just stick to the issue that we who object to telemetry are raising: the use of telemetry, period, is, in our opinion, unnecessary and a step in the wrong direction."

In your opinion - that's the key bit. There are users (like myself) and developers who hold a very different opinion, therefore the only sensible solution is to allow people to choose individually. And that's exactly what they've done.

"First, it's unncessary for MS to collect this data."

It's not "necessary" to improve MuseScore either, but that doesn't mean it's a bad idea to do so. The rationale has been explained very clearly in the original post, so I don't think this claim holds any water unless you can suggest a better way of achieving the same results. Your suggestion was "Just make the software, interact with people on forums, videos, etc., to enhance the product." but gathering this data manually would be far less effective / efficient and would waste precious developer time.

"Second, I don't like the idea of it even being suggested that MS can collect how often I use the Undo button vs. the Edit Undo menu option vs. Control+Z, etc."

Why not? What are you worried about, really?

"(And, again, opting out, in my opinion, misses the point.)"

What point does it miss? I totally respect your decision to opt out of any data collection. So please respect my decision to gladly share my usage data with the developers so that they can make MuseScore even better.

"As to the matter of leaking information and what have you: irrelevant. I never brought that up. I simply don't like the idea of what I'm doing even potentially being broadcasted."

So just opt out. There is no need to try and enforce your personal position on everyone else.

In reply to by Papibois

Papibois,
No, I wouldn't run off and look for other software. I was just a bit upset when I wrote the initial post. I love MS and suppor it. I do NOT however, support any idea, whatsoever, of telemetry. So, forigive me, please, for being a bit harsh.

In reply to by Starkman

My final thoughts and post on this subject.

Somehow, when replying to a post (from my email), the post doesn't seem to go where I thought it would (in response to someone else's post), and when I come to this sight to see the posts, I can't find the post that was sent to my email, so if I've caused confusion, or my posts don't make sense (for lack of reference to the original), I'm sorry about that. Not sure what happened.

As to the telemetry issue, I've stated my thoughts: completely unnecessary for notation software development. But never have I meant to accuse MS of wrong-doing or of any mal-intentions. Not at all.
So, I'm finished.
Thanks.

In reply to by Starkman

So, here’s the thing. A bunch of brilliant coders get together and write an open source program. It works great for them. They know computers inside and out. Their computers run better than most of ours. The program they write works just the way they want on their computers. So, they put it out into the world. Sure enough, right away they start to hear about problems some folks have with it. These folks, bless their hearts, have no idea what went wrong and don’t know how to find out or describe just what happened. There are just about as many setups out there as people using them. Developers can’t rely on users to describe what went wrong because the user doesn’t know.
There are problems like this with software as well as OS’s. Just ask any number of people trying to figure out Catalina. And if you think Apple isn’t collecting and selling your data, think again. They don’t need to sell it to as many vendors because Apple already sells devices for three times as much as anyone else. They are fine computers, don’t get me wrong. But I digress.
I don’t understand what part of “don’t opt in” is so hard to understand. “I don’t like it”? “It’s not needed”? Really? If you go on the internet, your privacy is already shot. If someone wants your data, they probably already have it despite what you do. You must be smart, of course. But hackers break into government computers every day. What chance have we got.
I have no problem with MuseScore collecting data about how I use it.

Question for the team:
Will the people who spend their time analyzing other peoples' usage data be sharing the results explicitly as relates to what directly influences design changes or other activities based on that handling of data with version upgrades and/or message updates in the forums, or will this be a sort of passive-influence without recognition as to the direct influences? Thanks.

In reply to by worldwideweary

I like questions including OR statement. The answer is Yes :) Yes to the first part of the statement. Design decisions based on the data will be shared. Btw, almost all design decisions are based on some sort of data. For example, recent single click improvements are based on user testing activities. We are not allowed to share details of that interviews due to different personal data protection laws. Meanwhile, we will share as much info and insights as we can.

Hi everybody,

I justed wanted to update to the most recent version. Here are my thoughts with regards to telemetry:

I don't like it. And I don't want it.

"Another goal of telemetry is to track why Musescore is being used." Let me be frank: I think that's non of your business. As isn't almost everything else you are asking for with regards to telemetry.

"Everybody can check by looking into the source code". Wrong, I can't, I'm not a coder. It's a stupid argument.

"Do we need to ask people about the approximate duration of the sessions, the frequency of using buttons and different features of the software?" Ever heard of user testing in real life? Give me 10 users and in two days time we will find all your problems. Easily, it's what we did for years and it still works. It's called UX.

I left Mac and Windows 10 behind bc. of all the bs that is happening there, only to find myself in Linux land with companies trying to do the same there now.

"MuseScore does not collect personal info, so there's nothing 'intimate' being turned over." Wrong. If I turn on telemetry I'm not only sending data about how I use MuseScore. I'm also sending data about MYSELF, it's called metadata. It is possible to generate a complete profil of every user that has telemetry turned on, and I cant see anywhere that MuseScore have addressed this. Sorry if you did, I couldn't find anything, but I still do not want it. Now tell me, why should I trust anybody and be sure that he doesn't use these data for other purposes? Easy answer: I don't. And, come on guys, I registered with you and the first thing I come across in doing so is your USER PRIVACY PREFERENCES for cookies, with EVERYTHING OPTED IN. Why do I have to opt out in the first place?

Let me tell you this: I will no longer recommend Muse Score. I've been using it for years, I love it, but I will not update to the current version (I'm still on 2.1.0). I'm already thinking of going back to Mac, because as a registered Finale- and Sibelius-user I can use Noteperformer and other plugins, which Musescore still cannot.

Marc Sabatella: "I won't try to tell you what might be valuable for you in your life, and you don't try to me what's valuable to me in mine." And: "I won't try to tell you what information is valuable to you in your job, and you don't try to tell me what information is valuable to me in mine." I think you guys at Musescore are way too pushy for my liking.

And to the user who said: "By the way, the first time you use W10, you have the oportunity to turn off most data gathering aspects of the OS. And you can still do it now." And I still can't believe people are so uninformed. No, you cannot stop data of beeing transmitted to MS. Some, but not all of theme. Don't spread such bs please.

Thank you.

Let me tell you my opinion again: I do not trust anybody in 2020 who wants to gather data. It's data mining. Every time. And if MuseScore wants to take that route: Fine. Go ahead...

In reply to by peter franticas

So just say 'No' at that dialog and be done with it. Data will only be gathered and send after you explicitly agreed to it.

You don't need to be a coder to read code. You just may not be able to understand it...
The point though is, that this being Open Source there are many coders who can and do read and understand the code, so would scream and should if and when detecting privacy breaches

You do have a point though with that need to opt out for Cookies, this does not seem to conform with EU laws, which require an opt in, as of lately. However, this is not related at all to the telemetry discussed here

In reply to by Jojo-Schmitz

Here is what I'm at: I have three companies I no longer trust. Avid, MakeMusic, and now even MuseScore. I need printed scores. Since I am between a rock and a hard place now I'm back to square one where I can start from the beginning and choose again. It will not be MuseScore, bc. as I said, having Noteperformer and other PlugIns with Sibelius or Finale is superb. And there is also Dorico as a new contender, which does some things in a new and often better way than the others.

In reply to by Jojo-Schmitz

No, as I said I do not trust them either, but since all three are going "telemetry" I can choose the one that fits my needs best. As I said Noteperformer is quit an argument for me. And I think it has nothing todo with open source vs. closed source, I recently found out that Ardour is also calling home. And I do not like that either.

In reply to by peter franticas

So you are saing that Ardour is calling home and not saying so? That would be rather stupid, as it is Open Source, so would be quite easy to detect.
Microsoft IIRC indeed had to get caught in the act (and that happend regarless their closed source nature) before they admitted to it, and added means to disable (most of) this.

In reply to by Jojo-Schmitz

I'm not in the position to recommend anything, but I personally would prefere you just skip that whole telemetry thing. It will not help you in finding out what problems users have. The only thing that helps is to talk to them, watch them while they are using your software. And the old "please help us to better understand what you are doing" like everybody else is telling us, well, don't expect users to take that serious in 2020. We all have been cheated already often enough by software companies, be it on Windows, Mac or Linux. Rumors will spread that MuseScore is "spying", and I can't blame people. Look at the posts in this thread, it's easy to see, people dont like it.
I'm not a coder, but I'm into software development and software training since ages. I have yet to see a case where user tracking remotely, or "telemetry", has ever led to better software. It's data mining. I know.

Thank you anyway, MuseScore is absolut fine, I love it. But no.

In reply to by Marc Sabatella

Sure you're curious :)
As I said "I have yet to see..." It's very easy, it doesn't work. The problem is that you end up with way too much data. Imagine only 1.000 user giving feedback, what are you going to do with it? Even if you take only 100 user opinions into account, you will end up with 101 opinions on everything and the kitchen sink. There is a reason you only have 5 or 6, maybe 7 test persons in a user lab. After that the findings rarely change. If they do you know you messed up comletely.
Alright, you are not talking to real people, you are listening to what your software is telling you about us.

Since you asked: How are you at musescore going to process the data you achieved?

In reply to by peter franticas

There's definitly no spying, as that needs stealth. There is no stealth in Open Source, conceptionally impossible.
Espionage or spying is the act of obtaining secret or confidential information or divulging of the same without the permission of the holder of the information....The practice is clandestine, as it is by definition unwelcome. (WikiPedia, emphasis mine)

In reply to by Jojo-Schmitz

One of the problems we are talking about is that even if I give you permissions to track ("telemetry") I don't know what exactly you are going to do with my data. As I said before there are several levels on that you can obtain data. The ones that cannot be hidden at all is METADATA. Profiling. Sure, you are tracking when I hit a button or whatever. But you also get a profile. METADATA.

Right or wrong?

In reply to by Jojo-Schmitz

IP address is NOT Metadata, it is open data.

Tracking the time when I start musescore is also open data. But tracking the time when I start musescore over 6 months will give you a complete overview at what time I'm usually working. This is METADATA that if combined with tracking of IP can result in a complete overview that answers the questions: When is this user at home? Does he only work from home, or does he travel with his computer, since (IP!) he logs in via different providers? When does he sleep? Does he work for a longer time in a row (=he has a project running), or does he only work for some minutes at irregluar times (=drooling about just for fun, no pro, not earning money with the product). Is he mostly working at night (=family / no family)?

Even simple data can be sensitiv:

User has Macbook Pro=money.
User has stupid lenovo=no money.
User has several computers=money, maybe big family
User has always the same IP, but sometimes another guys with musescore comes to him and logs in (=same IP address!). BINGO: First connection made, user x is in contact with user y.

See how that works? Now imagine you have data from 3.000.000 users (Wikipedia, number of downloads until 2013!)

Again: Metadata CANNOT be separated from open data, Metadata is inherent.

What can you do with the collected metadata? You can sell it for marketing or you can use it for yourself. Even if you're not going to do this, maybe your provider would (the one that stores your data in the cloud).

Now tell me you guys at musescore don't know this.

In reply to by peter franticas

As no user name is collected or sent, nobody would be able to guess your financial situation nor whether you have friends using their computer at your home. And the computer type isn't collected either, only whether it is Windows, macOS or Linux, possibly. Leaves the data about possible multiple computers from same IP, but even that only if they are on different OS.

In reply to by peter franticas

See https://musescore.org/en/faq#faq-300080 :

  • Shortcuts - each event contains internal name of the command associated with the shortcut
  • Buttons - each event contains internal name of the command associated with the button
  • Menus - each event contains internal name of the command associated with the menu item
  • Inspector - each event contains the type of the element the property was changed for and the long name of the property
    • A property has been changed
    • A property has been set as a style
    • A property has been reset to the styled value
  • Crash Free - each event is just a number representing one of the following conditions in which the app may start.
    • First start of MuseScore (or after reverting to factory settings)
    • Start of MuseScore after update
    • Previous session was finished correctly
    • Previous session was finished incorrectly (crash happened)

No IP, no username, no start/end time, no OS

In reply to by Jojo-Schmitz

I'm not in the position to recommend anything, but I personally would prefere you just skip that whole telemetry thing. It will not help you in finding out what problems users have. The only thing that helps is to talk to them, watch them while they are using your software. And the old "please help us to better understand what you are doing" like everybody else is telling us, well, don't expect users to take that serious in 2020. We all have been cheated already often enough by software companies, be it on Windows, Mac or Linux. Rumors will spread that MuseScore is "spying", and I can't blame people. Look at the posts in this thread, it's easy to see, people dont like it.
I'm not a coder, but I'm into software development and software training since ages. I have yet to see a case where user tracking remotely, or "telemetry", has ever led to better software. It's data mining. I know.
Thank you anyway, MuseScore is absolut fine, I love it. But no.

In reply to by peter franticas

It's certainly fine to reject using a program even though it is free and suits your needs, and choose instead to use another program that isn't free. No one can stop you.

I just want to address one bit of misinformation in your post, the part about doing focused tests with small groups of users. Yes, this is valuable indeed, as we have said. We do this also. But it provides different types of information than telemetry. Yes, for many years in the past developers/designers were limited to the sort information you can get from focus groups, and they were able to make use of it as best they could. But now that is has become possible to obtain additional types of very useful data, and therefore possible to make improvements in a way that were not previously, it would be doing a disservice to the people who use our products to not to try to improve in this way. "Sorry, we know how we can make the product better, but we refuse to try". It would seem a shame not to at least ask people if they'd like to help us in this way. They are welcome to say no, and we respect that.

But again, sure, if you'd rather pay other companies money for their products - and then probably also give them your data, because they are surely less up front about it than us - that is your right.

In reply to by Marc Sabatella

Marc,
"It's certainly fine to reject using a program even though it is free and suits your needs, and choose instead to use another program that isn't free."

Are we now going back to "u know, it's free, so give us your data please"? Wow, haven't heard that in a while...

And as I said before, MuseScore does not fit my needs perfectly, as we all know Noteperformer cannot be used. All the guys I'm usually in touch with during a project, working on probably more than ten computers, are using NP, and it has become a common denominator, at least inside the group I'm working with. I'm the only one on Linux. Without NP. That made me think.

And sorry, I don't see "misinformation in my post", I only have a different view on some points.

"But now that is has become possible to obtain additional types of very useful data, and therefore possible to make improvements in a way that were not previously, it would be doing a disservice to the people who use our products to not to try to improve in this way."

Well, that's on the same level as "please help us better understand...". Right.

"But again, sure, if you'd rather pay other companies money for their products..."

Please see above. Sorry Marc, you are coming with the old "Linux is free why pay for other products?". Not going there, but there are a lot of reason to pay for software. Having a proper manual would be one. Support for VST/AU/AAX-Plugins another. And also: Go and run a recording studio. Within one day you will know that spending another 500 bucks for Software is your least problem.

Guys, let's stop it here, I'm not eager to discuss the matter on a (too) personal level. MuseScore is a wonderful software, and I just cannot understand why you are going down that route, but well, time will tell. I'm not even angry.

I'm disappointed.

In reply to by peter franticas

To be clear, I am saying nothing remotely like what you are attributing to me. I intended my words to be taken at face value. When I say you are free to pay for other products, I meant that and only that. You truly are free to do this; I am fully supporting your right to make that choice. Nothing else implied.

Whether or not the MuseScore happens to want to serve its users by doing everything we can - including collect telemetry data - in an effort to improve has nothing to do with whether we charge for it or not, nothing to do with whether you personally want to help us in this endeavor or not. We would want to do everything we can to improve the product if it weren't free, and we continue to want to do everything we can to improve the product even if some small percentage of users are uncomfortable sharing the data that might help.

If we believed the majority of users would not be comfortable with us helping them in this way, then there would be no point in trying. But the fact that MuseScore is open source in addition to merely being "free" is something that we hope gives people confidence about why and how we are doing this. It remains the case that anyone who feels uneasy about this can simply not opt in. We hope that our complete transparency on this issue will encourage enough people to help us help them, though, to make it worthwhile for everyone. And in any event, even those who choose not to opt in will benefit from the improvements we make. So everyone wins whether you opt in or not.

The misinformation I referred to was the claim that focus groups can provide everything anyone would ever need to know regarding usability. That's just factually incorrect. Focus groups are useful indeed, but there are things they cannot tell you. This isn't a difference in viewpoints, it's a matter of fact. Whether or not opting in is worth the risk one perceives, that's a viewpiont, and you have the right to have a different viewpoint on that than me or anyone else. But whether or not telemetry data can provide information that cannot be obtained through focus groups is not a matter of viewpoints, but of facts.

In reply to by Marc Sabatella

Hi Marc,

so what you are saying is that these points down below need telemetry and can not be cross checked in a user lab? Well, okay then.

Let's leave it like that. This discussion helped me to make up my mind about data collection, telemetry (what a word) or whatever you may call it. Not buying.

Best
Cheers


-Shortcuts - each event contains internal name of the command associated with the shortcut
-Buttons - each event contains internal name of the command associated with the button
-Menus - each event contains internal name of the command associated with the menu item
-Inspector - each event contains the type of the element the property was changed for and the long name of the property
-A property has been changed
-A property has been set as a style
-A property has been reset to the styled value
-Crash Free - each event is just a number representing one of the following conditions in which the app may start.
-First start of MuseScore (or after reverting to factory settings)
-Start of MuseScore after update
-Previous session was finished correctly
-Previous session was finished incorrectly (crash happened)

In reply to by peter franticas

Of course you can collect a limited quantity of this kind of data in a controlled setting, but it will strongly biased by how you construct the test environment. For example, what kind of users do you select - completely new to notation software, experienced with some other application but new to MuseScre, somewhat experienced with MuseScore, expert MuseScore user? These are actually all interesting cases and worth studying individually, but you won't get the big picture data that is accurately representative of users as a whole. Another important variable is the kind of guidance you give users in your test. Eg, do you tell them, "hey, just play around and do whatever you like", or do you say, "see if you can figure out how to create and edits parts for the following orchestra score" or do say "choose one of your own files (assuming an existing user) and work on it as you normally would"? Again, each is interesting in itself and worth studying on its own, but none will give you the big picture.

So, controlled tests are good for relatively focused data, but they won't paint as accurate a big picture of usage as a whole. The great thing about telemetry data is that it eliminates these sorts of biases by automatically giving you an accurate sense of all users across all usage patterns, and you can use that data to help put the more focused data into a broader perspective.

That's my take, anyhow. Others may see the value of the data to them personally in a different way. But this is what my three and a half decades of professional experience in the field has taught me.

In reply to by peter franticas

I’m trying to understand the anti-telemetry argument. Near as I can tell, it boils down to this: It invades my privacy and doesn’t give any useful information.
It seems to me that if you believe that you are not interested in telemetry for those reasons, great.
If you believe that I shouldn’t use telemetry for those reasons, that’s different. What I agree to is up to me.
Opt in. or not. Seems simple to me. So many posts making this complicated.
Personally, I’d rather see better sound fonts. Not NotePerformer, though.

Just a suggestion about this. Can you please provide (perhaps as a sample at opt-in time or, better yet, as a permanent option in Preferences) some means to dump to a human-readable local file, the actual data (or a sample) being sent to analytics?

I've read the PR and seems okay to me (from a privacy point perspective), but I'm not totally clear about what events are being recorded for the telemetry as of today (every key presses? every menu selections?)

I see that the thread title is "Introducing Telemetry." So, has this been decided? It sounds like it is coming whether people want it or not.

I don't know if it matters what % of the users use something. Musescore is set up perfectly for beginners to use. The buttons on top tell the beginner all of the most important and most used functions, and they give hints of shortcuts. I don't think it's good to change that panel. I don't mind talking about what I use and how I use it, but I prefer information to be gotten from me and given from me through conversation, and not through mass collection of metadata. I should hope that this mass collection of data issue should become a priority for general users of digital tech in the future, so it would be harmful to Musescore in the long run to be going down that road.

It might be an option to allow the users to make more custom panels, including the main top panel. I don't use Concert Pitch. It does not annoy me to have it there, but I might want to remove it if there were a customizing feature.

In short, I don't think that it's necessary for anyone else to know how I use Musescore. I've requested a daily Log button that was never developed. Why? Because some % would want it and some % wouldn't? Why not make the button and allow me to decide for myself whether I would replace Concert Pitch with a daily log button. And what difference would it make whether someone else used it or not? Or what % used it? A Daily Log button is a great idea for composers who collect scraps all day. So far the idea has been shelved. So why would I want you to collect info from me when you have not developed the one thing that I precisely indicated I would want?

I think the best information you are getting is from the forum.

In reply to by Rockhoven

I think that there is a whole lot of misinformation on what telemetry is and what it is for.
It's already here in new versions of MuseScore. You can turn it of in Preferences.
It's not at all just about adding or deleting features, or seeing who does what. Someone was afraid they might loose the figured bass function because he was sure that he was one of the very few using it. Probably every music student using this software in college is using that function. OTOH, have no idea what the log function you talk about might be, so I don't know if I would use it or not. There are lots of things that I would like to see happen that I'm pretty sure won't. I don't think that's the point.
What matters to me more is when something goes wrong, and it will, is a way to help figure out what happened. I can't see under the hood. Somebody needs to be able to look at some data and just maybe point to something that they see happening a lot and maybe fix it.
Privacy? Heck, China probably already has your info :-)

In reply to by bobjp

That's right. China does have it, and they are getting it through projects like this one. As long as I have an option out of it, I'm OK. I don't want to be uncooperative, but I don't like all the data collection that's going on. I'm Old School. I believe that information should be communicated in conversations with the consent of all parties.

I looked in Preferences for the option. It says that it is anonymous. Tell that to China!

In reply to by bobjp

Maybe it's no big deal. I've left it for now. the problem I have is that one party can collect it, but another party can steal it from the party that collected it. You see what I'm saying. I'm just hyper about this because of the bif Windows 10 fiasco last week. Plus the FB problems with third parties accessing data, throwing elections and wotnot. I want to help. I just think that collecting metadata is sort of dehumanizing. Isn't the forum info much better? I know. It's not fast and convenient? Well, you're getting the data you need anyway because most people don't know that you're doing it and it's a long way from starting out as a beginner with Musescore and getting to even know that there's an option in Preferences.

Q. You don't need every users data, do you?

In reply to by Rockhoven

I think people engage in much more dehumanizing behaviour than collecting metadata. I don't believe for a minute that there was a big fiasco with the update that caused some uses to have their profile information hidden. I believe that it happened to some and that there was probably a good reason for it. There are 6 W10 computers in my house. I think that in all the years we've run Windows, only one of them had a minor glitch caused by an update. When you read the articles, they can only say that there have been several reports from different sources. No hard numbers, or even guesses, as to how wide spread problems are. So people panic. Some years ago, there was a big scare about some bad trojan that was going to hit. Sure enough, when it did hit, lots of systems went down. Turns out, Microsoft had already issued an update to protect against it. But many IT people and other users who were sceptical of updates didn't install it and took a hit as a result.
Sure, be hyper, but be sensible.
Face it, the only way to truly be secure is to never use the internet, ever. No online shopping, no browsing for information, no GPS, no email, no cellphone calls, no credit cards.

I said that I was against this. But I also opted in when I setup the latest version a few days ago. And I have not changed those setting. So why be against it? I think it hits a highly charged issue. I don't think many people are willing to be perfectly honest with you.

It looks like you were going to install this no matter what. People said they didn't want it, yet someone here apparently knows better. So why did you bother asking?

Marc: This project is not free. It is costing me an arm and a leg. Not this project in particular, but the whole Internet is costing us big time. This project is only complicit with the rest of the web.

BSG: "However, we're aware that there can be a resistance to information collection of any kind and we'd also like to hear the arguments for not doing it, too."

Someone is asking, and that's the reason for this discussion. The thread was made in the spirit of transparency. So, in that same spirit of transparency, who made this decision?

In reply to by Rockhoven

I have no idea what definition of "free" you are using, but by the normal one found in the dictionary and in the world of "free software" in particular (where there is a very specific technical definition that differs from the dictionary definition), Musescore is absolutely free. If it suits to talk about some aspect of MuseScore other than the ones normally implied by the word "free", feel free to provide a clear definition of what you mean, and then maybe we can agree on a better term. But the word "free" is already taken.

In reply to by Rockhoven

Jojo - I didn't bring it up, but so far BSG and you have mentioned "democracy," "election" and "free speech." I said "privacy." I'll let you guys read the news these days and figure it out for yourselves. I'm interested to know what people from France think. Telemetry is insignificant, but so was the straw that broke the camel's back.

In reply to by Rockhoven

This is a privacy issue
Talking specifically about the data collected by MuseScore here, is it though?

It's certainly perceived as such by yourself, but so far all your examples have pointed towards other types of (persona linked) data collection and profile building; both are quite irrelevant to the data being collected by MuseScore.

In reply to by Rockhoven

Rockhoven. It seems to me that MuseScore is free to ask people if they want to enable telemetry. And people are just as free to say no. Just as you are free to use the internet, or not. If you believe that your privacy is threatened by telemetry, you are free to believe that. People that believe otherwise are free to do that also. You can't argue from a standpoint that you are right and everyone else is wrong. All you can do is state your side (which you have so stated) and be done with it.

In reply to by Rockhoven

I don't have a "monopoly" on the dictionary, but you won't find any widely-accepted dictionary that provides a definition of "free" that supports the claim that MuseScore is not free. It is indeed a privacy issue, though. Some people don't mind sharing giving up that tiny bit of privacy in exchange for the help it would provide in improving the tools for everyone's benefit. Others feel that giving that that tiny amount of privacy is not worth the benefit to themselves or others. That's fine. If you would rather not help us help you, that's fine. I don't see the point of complaining here that we are asking for help though. You are free to just say no and leave it at that. Luckily plenty of others will help, and you will still benefit from the generosity of others even if you yourself decline.

Anyhow, none of this changes the fact that MuseScore is free under both the relevant definitions (free as in beer, free as in speech, as is commonly said in the software business), which was the point of my previous response.

Telemetry is installed and was going to be installed regardless of what anyone had to say about it. The title is "Introducing Telemetry." The OP states all of the benefits. The purpose of this discussion was stated at the end of the OP "However, we're aware that there can be a resistance to information collection of any kind and we'd also like to hear the arguments for not doing it, too." So what are your arguments against telemetry? I can think of many reason against it, and against collecting information on this forum also. But I'll leave that to you. You are just as smart as I am.

In reply to by Jojo-Schmitz

I opted in when I set up the latest version of Musescore this week. The OP wants us to voice our discontents. I'm doing my job. What about you? They gave all the good reasons and wanted to hear the flip side. So, I'm ready to make up a buncha stuff in order to cooperate.

The purpose of this discussion is to voice any objections. I cannot object without evaluating, and I can't evaluate without complete information. I've looked at this page https://musescore.org/en/node/300080

It says "What kind of data do you collect with telemetry?" Is this a complete list of the data that is collected? I'd like to see screenshots of all the data collected on a user.

In reply to by Rockhoven

As I said some time ago, I would be glad to enable telemetry as long as I know exactly what is being sent. To be really transparent requires that the user have access to a file or copyable report containing the infomation to be sent.

In am only halfway reading through this thread, which is more than most people will do. I will finish the discussion at another time.

My opinions so far are :

1) The real ethical issues have not been fully explored yet, in this discussion.

2) The people who are submitting the data should have access to the data, they should evaluate the data, and they should make recommendations for design changes.

My reasons: When you say in the OP that "WE can help YOU" - this I don't appreciate. I could just as well say "I'M helping YOU." What should be happening is that WE are helping US. The situation as it stands is uncomfortable. The people who are submitting the data should have control over that data. Otherwise the community is threatened by a sense of US and THEM. YOU can see MY data? But I cannot see my own data? And I cannot view YOUR data?

Can I install telemetry on the administrators of this forum and monitor their every action? I mean, so that I can help you become better administrators, right? I need to know every single move that you are making in your administration of this forum and in your use of my telemetry data.

No, this is NOT a JOEK!

In reply to by bobjp

Yes. It is simple. Opt in and I'll examine the data. Are the people who implemented this perfectly comfortable with the rest of us receiving and examining their data? As I said, the ethical issues have not been addressed in this discussion. I'm opted in. So I'll examine the data, right? I'll examine the data of those who implemented the telemetry program, right? If you don't agree to this, then opt out of the option.

In reply to by Rockhoven

In other words, are Anatoly and the other administrators who implemented this, and their supporters, including yourself and Marc, comfortable with submitting their telemetry data to the rest of us? I think the real issues here have not even been barely outlined. I doubt that the class of technocrats online is always honest, just and rational.

In reply to by bobjp

OK. So what if someone were to come in here and say "I'm not going to opt in until you give me access to the data. Furthermore I am going to advise other people here not to opt in until the data is made available. I am leading a boycott of the telemetry." No, what are you going to do with that person? You have their number, right. You have forum data on them. How are you going to use that forum data to control the conversation? We would need to install telemetry upon the administrators to prevent and even prohibit them from controlling the conversation. I believe you are already getting too much information from people's online activity here.

There is very real resistance to this out there. They just can't speak up. I have no idea whether the Yes and the No buttons are actually connected. What's to stop you from one day connecting both buttons to telemetry opt in? I have no idea what you are viewing and why you need it. I have no access to the info and therefore cannot recommend what changes should be made. i am extremely leary of technocrats taking these actions and making these decisions. Is there some reason why YOU cannot submit YOUR data to ME? Or do you choose to opt out of that choice? Because that would inform the rest of us accordingly, wouldn't it?

To illustrate the positions of users and administrators. An administrator might be a technically competent to develop this program. that does not mean that they are competent to make decisions about it's uses. That would be like allowing a television repairperson to choose what television shows I should watch and when.

I have been opted in for over a month now and still have not had my sound problems addressed. Can't telemetry alert you? What does telemetry tell you about the playback problems? f telemetry can help individuals here with specific problems, I'm fine with that. But we, the opt-in should have access to the data. Is there some objection to that? You can always opt out. Simple solution. YOU opt out of that. How many people here want to opt out of that choice? Would you opt in or out if ALL of US could view the data? Who would opt in and who would opt out?

What happens when the administration changes here. People do die. They would be replaced. So someone tales over and they have all of that data, and I still don't know what exactly what all of that data is. Could be location data. Anyway, they say to themselves "Hey I want all the data and I'm not going to tell anyone. I'm connecting the No button to the Yes circuit. I'm going to add in all kinds of stuff and not tell anyone about the changes."

Don't keep saying "Opt out!" I'm opted in and I want access to the info. We who are providing the information will examine it and make the recommendations. This is 2020, not the year 2000. There is about to be a rude awakening on the internet.

In reply to by Rockhoven

You say you have no idea if the yes and no buttons are hooked up. Do you you really distrust the open source community that much? If so, then why would you trust anyone who simply says "there is no telemetry at all"? Or who says, "running this program won't erase your disk drive"? At some level you need to trust people, if you are going to use the software they produce. But at least open source makes that especially easy, because if anyone tried lying, they'd be caught. Maybe not be you, but by someone, and you'd hear about it.

I don't get the TV analogy. The people who develop software - well, more the ones who design it - are exactly the folks who benefit from data about how the program is used, because it helps them make better design decisions. Is there someone else you'd rather be using that data?

In any case, telemetry data is collective and anonymous, not individual. No one seems your data individually, and no one seems mine either. In theory, I suppose the collective data could be made public, but then people would need to opt in to that, and I think they'd be much less likely to agree. It's one thing to agree share telemetry data with the folks who work hard to produce a program you love and use and want to see improved, especially when the people you are agreeing to share it with are actually in a position to do something constructive with it. Why would anyone agree to share that data with random strangers who have never done anything to help anyone and wouldn't have the means to do anything constructive with the data in any case? I think the fact that the data is not shared outside the MuseScore company is probably pretty important to making the average person feel OK about it.

As for any sound problems you are having, best to start a new thread about them. I see some older ones of your but most seemed to be simple misunderstandings. But even if there are bugs involved, I don't see any way telemetry data could possibly help with that, though.

In reply to by Marc Sabatella

Although it is clear to me that the uses of this "phone home" data are benign (and I do not opt in), one should keep in mind very widespread uses of phone-home data that are not benign (although their perpetrators would say "This allows us to better support your tastes and viewing preferences and offer you a customized stream of (whatevers)!") New TV's phone home to say exactly what shows or movies you watch, where you change channels, and so forth. This data is richly valuable to advertisers, and is sold and resold, to develop, usually completely automatically, "profiles" of your most intimate and sensitive personal details to offer "products we think you''ll like". The story of a woman who was offered baby supplies before she herself knew she was pregnant is legendary. And let's not forget law enforcement uses of search history, such as dragnetting people who looked up a recent victim. These things are not benign, directly impact the power balance between individuals and larger actors, and are deeply troubling as models of the social contract evolve in the adolescent digital age.

I don't think collecting MuseScore command usage data anonymously is dangerous or problematic. But I want to make the case for those who are horrified by the notion of phone-home "metering" (or "surveillance", as one prefers) and are understandably insensitive to arguments about "degree".

"No pets" means no hamsters as well as no vipers.

In reply to by BSG

"These things ...directly impact the power balance between individuals and larger actors...in the adolescent digital age."
Well said.

Prior to the digital age, any nascent republic would solicit its citizenry: "We're from the government. We're here to help."
(Many still do to this day.)

As we are all observers of, and participants in, history (no opt-in or voluntary enrollment is available), I can understand people's inherent distrust.

(BTW: I'm still running ver. 3.3.4 ;-)

In reply to by Marc Sabatella

The first objection on this subject is mine.
And I write here again -for the record-.

Telemetry cannot be used for things because it's easily recognizable.
For example, let me tell you what you can get from Palettes (without seeing the data):
These are rarely used:

  • double sharps and double flats
  • C clefs, F clefs on the 3rd and 5th lines
  • 64th, 128th notes.
  • The rest of the ornaments, except for the first four (turns, trill and mordent).
  • soft accents, fade in and out, volume swell, sawtooth
  • four or more dynamics eg: pppp, fffff
  • slides an doit, scoop etc.
  • short ceasura, salzedo
  • metric modulations
  • feathered beams
  • Noteheads palette
  • tick .n. span barlines

I can also say what is used frequently from the menus and other places (there is no need to be psychic {or telemetric}, it is very easy to know).
for example: The most used menus:
File Menu: "New", "Open", "Open Recent", "Save", "Export"
Edit Menu: "Undo","Copy", "Paste"
Help menu: Click and Nothing :)
Most used key: Spacebar (for Play and Stop)

Q: So if it's easy to know, then what is telemetry used for?
Answer: For deeper things!
Q: What could these be?
Answer:

  • In which sequences does the software crash?
  • How long does an average session take for a score?
  • Shortcuts and note entry sequences used on the keyboard

Q: Is the list so small?
Answer: Nothing more comes to my mind. What else can you do except move / drag your mouse, click somewhere or press some keys on the keyboard?

Q: So this telemetry thing is unnecessary
Answer: Perhaps keystrokes or mouse-moves are important to someone, or they have time to waste(!).
Or they want to put something else on an important part of the main panel at the top, and they want to evacuate it, using the telemetric data as an excuse.

Q: Really?
Answer: No, I'm just guessing. This has nothing to do with reality.

In reply to by Ziya Mete Demircan

Hi there. My name is Martin. I'm the head of design at MuseScore. Just thought I'd weigh-in here to give a practical example of where telemetry is useful to us.

First - you are right that certain things are obvious and that we don't need telemetry to justify them. The case where Telemetry is useful is where we literally have no idea of usage. For example: there is an option in near the top of the Inspector (next to the offset settings) where people can turn on a 'grid' function. Since I am currently reorganising the inspector, I'm asking myself "does this feature deserve to be so prominently placed?". The way I determine this is by asking users and our community first. The answers I get are mixed. Some people swear by it and others never even realised it existed. The problem is we have no idea what the reality is, once you take our millions of users into account. Is it used by 30% or people or 0.2%? My own feeling is that we could make much better use of this real-estate but I'm leaning on my own biases here. Bear in mind that there are 100's of different options in the Inspector where we are asking this question.

This is where telemetry is very useful. It gives us a factual understanding of exactly how much usage each function in the Inspector is getting. In some cases, we discover that things which are quite hard to find are used quite a lot... which means that they are not positioned in the appropriate place. In other cases, there are options right at the top of the inspector which very few people bother with. This could either be due to poor descriptions or obscure iconography. These are the kinds of things users don't complain about because they don't know they are problems. People don't say things like 'this button is fine where it is' because things that are fine don't need to be commented on. Similarly, people don't tell us 'I was unable to discover X feature' because they don't know it is there to be discovered.

UI organisation is a complex problem and you often have to take a utilitarian approach by promoting the most commonly needed tools and providing alternative methods for advanced users to get what they want. It is impossible to get a high-level look at how everything is performing by conducting interviews alone and we want to make sure that our changes disrupt people's experience as little as possible.

One last thing I'd say about this is the Telemetry is only one tool we use. Our primary method of gathering information is user testing. These two methods of collecting information are not the same at all and one does not make up for the other. I understand you have an aversion to the idea of Telemetry but I can assure you that it is very useful to us and there is a reason why just about every proprietary app in the world uses it. The only difference is they almost never ask.

Do you still have an unanswered question? Please log in first to post your question.