Number of MuseScore languages listed on Wikipedia article

• Sep 24, 2018 - 08:29

(This could have gone to the Translation or Documentation forum. I chose the former, since it affects the translation effort more)

I have started doing some fact checking and clean-up on the MuseScore Wikipedia article,
https://en.m.wikipedia.org/wiki/MuseScore
According to it, the desktop software is “Available in 48 languages” (which is qualified by a disclaimer of “although some translations are incomplete at this stage.”)

I just linked the Transifex Musescore Dashboard (https://www.transifex.com/musescore/) as a citation on the MuseScore languages available section on Wikipedia. The Dashboard lists 64 projects (Languages).

From the Transifex site some statistics (%s only looking at translated):
* 11 are listed as ready for use (= 100% translated)
* 4 more have 66 or less strings to be translated (= 99% translated)
* 3 more have 330 or less (=95% translated)

My Question: what do we consider translated and or general purpose usable? 100%, 99% / 95%?

Stating ”Available in 48 languages” (or 64, e.g. anything that looks at current translations in progress rather than useable ones), regardless of the disclaimer, feels like a misrepresentation to me.

For perspective (looking at Wikipedia articles), Sibelius and Dorico each list 9 languages, Finale 8. Even if we only list the 100% complete ones (11), MuseScore still exceeds them, if that is a factor.

I suggest we change to
* “Available in 11 Languages” and only list the (currently) 100% translated ones, “with many more in progress”.
* “although some translations are incomplete at this stage.” would no longer be necessary

This will also make it also easier to maintain the Wikipedia article and the shortened list of languages.

I could also add to the MuseScore Wikipedia “Talk” page how we calculate the “Available in x languages”, if it is not glaringly obvious from the % or criteria that we decide on.


Comments

MuseScore is available in 64 languages. Anyone can help in completing those that are not already, language updates are available 'on the fly', no updates MuseScore needed, so this is a moving target. The number of complete translation won't shrink, only rise.
I've just changed that 48 to a 64, I don't think more needs to be done

In reply to by Jojo-Schmitz

Sorry, bat that's a rather unconventional interpretation of the word 'available'.
Next time you go to the bakery your baker hands you some flower, a cup of water and "... salt is on sale at the supermarket!" ...
List the 11 (?) complete (available) languages and be proud. When more are finished, add those.

More arguments against 48 (or worse, 64) Languages being considered "available"

Questions (most of which I thought of as rhetorical, but some of which we apparently do not agree on the answers):

  • If a feature is in the 3.0 Alpha, would we consider it generally available, fit for use, since anyone who can code can improve it? We would not. That would be absurd.
  • If a translation has 1300 strings missing (80% of 6500 total strings), would users consider that language "available"? I would not. At best, it depends on which of the 1300, and which parts of the UI the user is using.
  • Are
    a) having a language in the Transifex framework created (one of 64, with a completion rate of anything between 0% and 100% translated) and (a.k.a. the "participation" criterion)
    b) any reasonable end user’s definition of what is available or useable (anything between 11 and 22 languages that are 90% or up) (a.k.a. the "achievement" criterion)
    the same? I would say no.

Why are we so massively lowering the bar for the translations? Are we encouraging anyone to participate and get to 100% (even if it includes the 609 tutorial strings)?

Lasconic and Jojo-Schmitz . You are some of the heaviest heavyweights on these forums. I implore you to reconsider your position - please let us not use the fact that the software is open source and the translation open to anyone who knows a second language (hint - this is not how you define “translator” but this is what we make do with) as an excuse to make the English word “available” lose any semblance of meaning within the context of the MuseScore translations. The vast majority of users have no interest in improving a translation. They just want to use it (in their native language if theirs is not English, being a bonus). Creating the expectation that a language is available (in the conventional meaning of the word, which I assume implies functionally complete if not 100% complete) when it is not, will leave a bad taste in the mouth, and create a negative perception of MuseScore's localisation initiative in specifically and MuseScore in general, of over-promising and under-delivering.

As a minimum, to look at what translation is truly done (of the 6512 strings), the MuseScore UI (4771), Instruments (1132) and Start Center (7) would have to be included. That leaves the YouTube tutorial (602). No easy way to split it off, and having to go through each language to see which resources they are really done translating. From experience, translating the tutorials is a breeze, compared to some of the UI strings without the proper context of their use (and I mean more than the “context” property in Transifex). If a language has 602 strings outstanding I can bet that the majority of those are on the UI side.

regards
Riaan

In reply to by Riaan van Niekerk

My personal experience with other opensource projects is that different people can have good-faith different ideas about what "available" means, which is seldom affected by extensive argumentation. In fact, I've met opensource advocates who genuinely believed that all it takes for a program to be considered "available" in a language is for the user to see his own language upon opening the program, even if it's just the first-level menu items.

Personally, I do not consider a product to be "available" in a language unless a reasonably recent version of the user manual or help pages is also translated (unless it is a product without any such resources). But I acknowledge that others may feel that a product is "available" if all or at least e.g. 95% of the menus and dialogs are translated. That said, I don't think it useful to argue about the semantics of "available".

I don't think there is a single answer to this problem, but my approach does not centre on the semantics of "available".

People involved in the volunteer translation field will know that if it says that a program is "being translated into X languages", it doesn't mean that none of those languages have been completed yet, but a general reader may well interpret such a statement as merely a marketing promise, which will not reflect well on the product. On the other hand, general readers may not be familiar with software products that get released before they are "fully complete", so I grant Riaan's point that saying "available in X languages" can create the wrong impression with the wrong kind of reader.

Since it is not possible to somehow link the Wikipedia page to Transifex so that it always reflects the latest information, the information given on Wikipedia should reflect some kind of general truth. On the other hand one must ware against adding information to Wikipedia that may quickly become out of date.

Also:

The number of languages listed on the Wikipedia page forms part of the infobox, and infoboxes on Wikipedia typically contain abbreviated or generalised information. The purpose of an infobox is to give a visitor a brief overview of the subject, and not to provide the user with up-to-date and reliable information that he can base critical decisions on. If really necessary, more comprehensive (and complete) information can be given further in the article. The number of languages listed in any software infobox reflects the situation on a certain date. As far as I know, you can add a date or version number to it, but I would not consider that necessary unless the information is wildly out of date.

"with many more in progress" sounds like marketing language and will eventually be removed from the Wikipedia article by someone.

The actual number of languages varies almost from PR to PR if you count 100% as the standard for "Translated in # languages" At this moment there are 10 languages at 100% and 19 at 95% translated or higher and 5 have less than 100 untranslated strings.

This is also a skewed stat since there are a number of PRs that have been merged into the master branch is these numbers do not accurately reflect the latest release version.

"MuseScore is available in 65 Languages" is a bit deceptive on the surface because it is less than 50% translated in 24 languages (I don't count American English since it's the original language of MuseScore).

The real questions is how many languages can MuseScore be used in if you don't speak English. English is the fallback language for words that are not translated into another language. For example, if you look at the Japanese forum you will see a lot of English words intermixed in most conversations concerning MuseScore.

In reply to by mike320

There's also a question of which strings are or are not translated. I could probably make a case that MsueScore would be extremely useful with only 50% strings translated, if you choose the right 50%. Like, most people could live without the translations for the bagpipe embellishment tooltips :-). For that matter, tooltips in general are less important than menu items and dialog labels, etc. I get the sense we don't have any such priorization?

In reply to by Marc Sabatella

You are correct that many of the strings are less important than others in translating MuseScore. It's not easy from my knowledge of Transifex to determine which category of strings (tooltip [less important], dialog labels [very important], menu items [very important] etc.) are not translated by language. What priority do you place on screen reader text? hhpmusic probably puts a high priority on it, 99% of the other users not so much.

I think my point about which translations are useful to non-English speakers is the most important gage of whether MuseScore can be considered translated to that language. This is very subjective and cannot be measured by a percentage unless it's 100%.

We had a very constructive discussion on the MuseScore translation forum (https://t.me/MuseScoreTranslation , which I strongly recommend that anyone involved with the localisation effort, joins).
and agreed to list the the languages fully translated (or within 1% of), and then another 12 languages those languages with 72% translated and up (2000 or less strings outstanding).

To see the results: https://en.wikipedia.org/wiki/MuseScore -> Infobox on the right, click [show] to expand List of languages.

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