Too many Crimean Tatar keyboard layouts added for Romania

I've notified Bhavani Shankar and Reşat SABIQ because they introduced the four new keyboard layouts for Romanian Crimean Tatars and they should know best how to proceed with this issue.

Bhavani, Reşat what is your point of view regarding this issues. Please let us know!

Looking at the wikipedia article about Crimean Tatars, we should add the layouts to Ukraine, as Ukraine is hosting the majority of Crimean Tatars

http://en.wikipedia.org/wiki/Crimean_Tatars

In Romania we have a small number of tatars and if we are going to add specific keyboard for each minority we will end up with a list of about 30 layouts.

Anyhow, grouping layout by Contry is not the right thing to do and layouts are associated with languages and not countries.

My vote is for having the Crimean Tatars layout for Ukraine.

Cheers,

+1 for moving the keyboard layouts to Ukraine, since there are 10 times more Crimean Tatars than in Romania.
Please keep in mind that the Romanian keyboards are also meant for use in the Republic of Moldova, where is no [known] Crimean Tatar minority.

I think you both got it wrong: xkeyboard-config groups layouts by country.
In our country there are many ethnic groups.
We should make their life easier and help them write in Romanian and their native language without much effort.

What I think is wrong is the number of keyboard layouts for 24.000 Tatars - one for every 6.000, and if we consider Linux users are 1% of computer users and all consider that all Romanians use a computer (which is evidently false) we have one keyboard layout for less than 60 Tatars.

I think we can accommodate all 240 Linux using Tatars in one single keyboard layout.

Thanks for forwarding this issue upstream and engaging with suv on it. I am of the feeling that this is really rather an upstream decision, and not of high enough criticality (compared with X crashes and so on) that we need to take action in the distro, thus I'm marking it wontfix. We will take whatever solution is arrived at upstream when we merge in the latest xkeyboard-config package. Thanks for working on this issue and helping to make xkeyboard-config optimal for everyone.

As I was one of those, who triggered the discussion, even if not directly, I am compelled to make some clarifications. So: ‘t's about Romanian users being exposed to layouts for languages they don't have any interest in seeing (or even, apparently, which they would rather not see)’. No, this is NOT a political issue, it has nothing to do with tolerance, with human rights and with minority issues. It has to do WITH CORRECTNESS of a linguistic issue connected to a keyboard layout, nothing more.
As a linguist and a person who speaks several languages, and—even if not familiar with Altaic languages—I feel at home with linguistics at large.
To a large extent, this issue is connected to the entirely wrong approach of keylayouts in Linux (am a Mac OS user, and creator of several keylayouts for Romanian—several variants, Old Italic, Old Church Slavonic, Czech. So, I feel competent to speak about this issue (see http://www.unibuc.ro/ro/cd_sorpaliga_ro , the English page seems to be under revision). Linux distros deal with political entities, and then places keylayouts under countries / states, an entirely wrong view, even if, of course, covering a reality. But what happens with those situations when there is no state there? What about Old Church Slavonic, Old Italic / Etruscan, Gothic etc. etc. Shall we randomly put them under a certain pretext?
The four or five Crimean Tatar keylayouts under Romanian is a total absurdity. It is true that there is a Tartar minority in Dobrudja, respected as human beings (I know some of them personally, my wife worked with them for 2 years). What has this to do with putting those unhappy keylayouts under Romania?
Do something, find something more intelligent than this really bad solution. Be constructive, use your inteliigence in order to achieve a superior way of thinking and do remove those keylayouts from there. They are entirely useless. And note that they are badly put there not because I / we do not want to see them there, but because their location should not be there. As Tartar is an Altaic language, until you find a more generic and more flexible solution, their place may be under Turkey or Mongolia or anything similar. This is not an ideal solution either, I agree, but it has at least a minimum of a logical approach.

And, very briefly: the long-standing improvised solution of keyboard layouts in Linux, so rigid and so impossible to modify (as in Mac OS, using either Alex Eulenberg’s site or John Brownies’s UKELELE) is the main reason why linux cannot be efficiently used in foreign language teaching and never in dealing with special situations like old languages like Old church Slavonic, Old Italic / Etruscan, Gothic etc. etc. because is a tantalizing cause to try to create your keylayouts in linux. This is, in fact, the essence of the discussion, not those badly put Tartar kaylayouts under Romania, disregarding who had in fact this eerie, whimsical idea.

Does anyone read this? Is there a reaction to the ideas exchanged here? Is there an admin of this discussion? Just to know whether it makes sense to continue watching this thread.

1. you should comment on the upstream bug: http://bugs.freedesktop.org/show_bug.cgi?id=23544

This is not a Ubuntu specific bug and it just tracks the status of the one upstream. Also fixing this in Ubuntu is not recommended as hardly any other distro will get the fix from Ubuntu.

2. The upstream maintainer agreed that 4 layouts may be a bit too much for a community of less than 25000 people making < 0.1 of the population and I tried to find some Tatars to see if one of the layouts was good enough or if they can be amended to contain the Tatar-specific letters and the Romanian ones so that they'll be able to write correctly in both languages without needing to switch between keyboard layouts.

Unfortunately I only found one Tatar ethnic and he wasn't writing anything in Tatar on the computer and could not help me in this regard. To be noted: he wasn't even aware of keyboard layouts for Romanian either and was not using Romanian diacritics to write Romanian texts.

Sorin: you mention having some Tatar contacts. Could you please contact them and see whether we can define a single keyboard layout that would either:
a) be enough for both their Tatar and Romanian texts
b) be enough for their Tatar texts

For starters, Crimean Tatar isn't really related to Mongolian. Mongols in related territories got assimilated, and their language hasn't had a major influence on Turk(ic) languages: although some words from Mongolian have been adopted and are in use. Now to the issue at hand.

> To a large extent, this issue is connected to the entirely wrong approach of keylayouts in Linux (am a Mac OS user, and creator of several keylayouts for Romanian—several variants, Old Italic, Old Church Slavonic, Czech. So, I feel competent to speak about this issue (see http://www.unibuc.ro/ro/cd_sorpaliga_ro , the English page seems to be under revision). Linux distros deal with political entities, and then places keylayouts under countries / states, an entirely wrong view, even if, of course, covering a reality. But what happens with those situations when there is no state there? What about Old Church Slavonic, Old Italic / Etruscan, Gothic etc. etc. Shall we randomly put them under a certain pretext?

> And, very briefly: the long-standing improvised solution of keyboard layouts in Linux, so rigid and so impossible to modify (as in Mac OS, using either Alex Eulenberg’s site or John Brownies’s UKELELE) is the main reason why linux cannot be efficiently used in foreign language teaching and never in dealing with special situations like old languages like Old church Slavonic, Old Italic / Etruscan, Gothic etc. etc. because is a tantalizing cause to try to create your keylayouts in linux. This is, in fact, the essence of the discussion

I agree with these 2 quotes. I have expressed similar, if not the exact same, opinions in this regard. The issue is not with Crimean Tatar layouts, but with the design pushed by 1 or 2 people (without any discussion). The design is based on political entities, rather than linguistics, thus stirring unnecessary inter-ethnic conflicts. No other OS has such a politicized keyboard layout approach. The only reason it is still flying in Linux, is because there hasn't been any discussion about the design: the decisions have been made by 1 or 2 people without any open public discussion. Well, it is obvious that political entities have no place in software: so they have to be removed, leaving only linguistic considerations.

I personally do not accept the superiority of an opinion of 1 maintainer (who is in fact a subject of a country that has been torturing languages for centuries). If 1 such person is stonewalling a proper design for keyboard layouts, it would be totally OK from my viewpoint to do any of the following:
a. overrule such an opinion
b. provide patches for Ubuntu and see what the Ubuntu community thinks about it: if it goes through in Ubuntu, there's a chance other distros will pick a more logical approach as well (this would be OK initially, but in longer term would probably necessitate c.)
c. branch xkeyboard-config project into a new open source project that is free from political entities and considerations, and based solely on linguistics and informatics

The alternatives to fixing the real issue mentioned here, involving deleting Crimean Tatar layouts, would be an unacceptable bigorty from my viewpoint, and will only waste...

1. I have mentioned this upstream earlier, but here's an example of a logical locale-based approach to layouts:

-Language1
 |__ Common Layout1
 |__ Common Layout2
 |__ Common Layout3
-Language2
 |__ Common Layout1
 |__ Common Layout2
 |__ Country1
      |__Country1-specific layout1
      |__Country1-specific layout2
 |__ Country2
      |__Country2-specific layout1
      |__Country2-specific layout2
+Language3

This is a mockup of a tree, with 2 recursively expanded nodes, and 1 collapsed node. E.g., by Common Layoutx, i really mean Langy Common Layoutx, but it was getting too verbose, so it's implied...

And, in terms of implementations, as i've mentioned before, i believe locale-based files would be ideal:
lang1
lang2
lang2-Country1
lang2-Country2
lang3
and so on.

Furthermore, the UI implementation could provide a country-list for filtering. If somebody needs US-specific layouts, the filter could display all languages with US-specific layouts (filtering non-US specific layouts for a matching language out, maybe with an option of filtering them back in)...

2. Alternatively, there could simply be a single list of locales, just as the implementation files above, and once a locale is chosen, all layouts for it are shown, including common ones, if they aren't overriden. This would also be a very clean approach.

Anyway, i don't have much time for this: this is just a few quick thoughts on the matter. This is the kind of discussion that needs to take place. Mac and Windows need to be analyzed to take the best from them, and improve further...

Regards.

So:
1. People commenting here basically agree with the idea that the politically-based architecture of Linux keyboard layouts is not good, and that it should be done more flexible. If this is not obvious if referring to modern, contemporary languages, it all becomes clear if we wish to add extinct and / or classical languages, which are not politically-based keylayouts, as the ones mentioned.
2. Something must be also done with (and for) such isolated keylayouts like Crimean Tartar even if, frankly, I still cannot see the argument for placing them under Romanian! (Crimean Tartar is indeed an Altaic language akin to Mongolian and to all other Altaic languages, going as far as—very probably—Korean and Japanese as a very remote expression of an easternmost expansion of the Altaic group). Note: I do not have close contacts with Tartars in Dobrudja, I once had, but—frankly—I doubt any of them, even if close to me in this moment, may be of any help here. But I assume that those who created those keylayous must have had some contacts, otherwise I cannot imagine a scenario that they put the keylayouts there without having any contact with the issue. Or am I wrong here? Or adding keylayouts in Linux is like a political dance? This problem is NOT a political one and I regret noting that, in time, Linux developers lost the initial philosophy: simple, free, flexible.
3. Adding too many variants under a basic, reference variant is not a good idea, because to many people this may be confusing. There should be 1-3 variants at the most, exceptionally more, but adding additional keylayouts should be simple and flexible, ideally Mac OS style or any similar simple method. Also, a simple to use keylayout creator, similar to UKELELE, should be implemented, together with a simple way for installing the keylayouts. THIS would solve many issues, allowing users to create / adapt existing keylayouts for specific needs.

Let us move at least one step forward with solutions. I for one have my solutions for Mac OS, am not technically competent here, but am ready to discuss and suggest methods based on what I know.

FYI: the politicized "By country" tab has been dropped in the latest xkeyboard-config code, and a patch[1] has been submitted to show Crimean Tatar layouts in a language-centric manner, consistently with how it's now done for most other layouts. Among other things this patch ensures that Crimean Tatar layouts are not mixed with Romanian layouts.

[1] https://bugs.freedesktop.org/show_bug.cgi?id=35787

Thanks. Unfortunately UI freeze has already passed for Ubuntu natty, but will look forward to seeing these changes included in oneiric when we update to a new xkeyboard-config release.

The code fixes are on their way. The last thing left to be fixed is the casing in the subject of this bug. :)
I look forward to seeing language-centric keyboard preferences UI in stable releases as well.

Thanks all.