New languages? Awesome!

As Impersonator said, it's because Tatoeba follows ISO 639-3 for its definition of a language.

Can we have this feature rolled out to all input fields? Specifically Add Sentence and Change Language (for a sentence). In both, I am restricted to the languages in my profile, and this is causing problems.

For Add Sentence, it can lead to situations like this: http://tatoeba.org/eng/sentence...38210#comments

For Change Language, I find that even as a Corpus Maintainer, I cannot fix the flag of some languages because they are not in my profile.

I like the idea of grouping languages according to profile, but I think the current system is restrictive to a fault.

This seems to be a problem with DirectWrite rendering, as I see this problem on Google Chrome and Internet Explorer. I don't use DirectWrite on my Firefox, and it renders just fine: http://i.imgur.com/geoHM96.jpg (I use MacType for font smoothing)

Edit: Yep, re-enabled Direct2D on Firefox and the text changed to the larger ə. If I'm not mistaken, font hinting is at play here. I wonder if there are any settings for DirectWrite's font rendering anywhere.

Going to second that, very good change.

I use Jameel Noori Nasta'liq myself. Makes me cringe a bit, because it makes all Arabic text rendered in Nasta'liq, but I make do because Urdu is the only language I read in the Arabic script.

I agree with tommy_san. I'm not an expert at HTML5, but I do think this should be possible with web fonts and some scripting to limit it to Urdu (and Kashmiri, Western Punjabi, Ottoman Turkish etc. as the need may arise). We could also use web fonts for other scripts that are not very widely supported, especially ancient ones.

مبارک ہو طاہر جی!

پھر سے خوش آمدید!

Considering almost all the Marathi sentences are my own, it would be safe to say that all my Corpus Maintainer actions have been outside of Marathi.

I became a corpus maintainer to keep an eye on South Asian languages in general because there are a lot of spam or mistaken sentences added by confused users, especially in Malayalam, but also Urdu and Hindi. These are usually sentences written with the assumption that Tatoeba is a translator website, I'm guessing. I frequently had to leave notes and keep bumping them, and eventually I decided to step up.

I often correct typos and other minor errors in the languages I know well: English, Marathi and Hindi, and frequently create links between these.

> "Because I wish to make such corrections so as to ameliorate the quality of this website, I requested to become a corpus maintainer myself, but my request was denied, on the basis that as I am a native Macedonian speaker, I could only take care of the Macedonian corpus (for which there is evidently no need, since I constitute almost 100% of the corpus already, such that I can manipulate it as I see fit even as an advanced contributor)."

I find this very unusual. I was not aware of any such rule, and I was certainly not told of this when I became a corpus maintainer. Like 123xyz, I am responsible for nearly 100% of the sentences in my native language, Marathi, but I wasn't denied CM status for it. In fact, I don't recall 123xyz's request to be a corpus maintainer was put on the wall at all.

I would be very interested to hear why I could become a corpus maintainer, but not 123xyz.

なんてかわいいこえ！よみさんはきっとせいゆうだろう？ｗｗ　アニメやVNからリップしたオーディオみたいとおもった。

ありがとうございました、よみさん！

I was wondering this myself. +1

As for the distinction between those claiming one native language vs. those claiming multiple native languages - I don't really the difference. If a person claiming more than one native language is considered more unreliable, then, as sacredceltic suggested, a person claiming a particular native language may also be considered unreliable, as it's possible they are lying.

I am against the idea of tying users' self-declared ability in a language to the quality of their contributions.

While Userscripts died, its role is now largely performed by Greasy Fork: https://greasyfork.org/

I suggest you also upload the scripts there, as it's becoming a replacement for Userscripts.

I noticed Tatoeba mentioned on a well-known tech blog when a language-learning startup started using its API. That may have caused an influx of new users.

Link: http://thenextweb.com/apps/2015...st-ios-update/

This sounds like a good idea. Maybe there can be a "See Members" button next to the languages on the page.

It could also be handy to include stats the way CK's native speaker page did. For example, a quick horizontal bar chart showing what percentage of sentences the user has added in that language. This way, users will know at a glance, what the pattern of the user's contributions is.

+1 I've always wondered this myself.

Unicode all day, every day. There's no need to reinforce outmoded practices just because they are widespread. I imagine Myanmar is lagging behind because they only opened up communication to the internet recently? They will eventually have to get used to Unicode then, as similar non-Unicode solutions were once all across South and SE Asia, and had to be weeded out by the efforts of many app-makers, web designers and transliterators.

As for adding Myanmar, I believe Alan's current policy is to only add languages for which we have native contributors ready to consistently contribute sentences in that language.

I like it a lot, especially the Details field. I agree with CK's suggestion.

Edit: I can't think of what the "0: Almost no knowledge" level can be useful for. I can think of half a dozen language I have "almost no knowledge" of, but stating those really wouldn't serve much purpose here on Tatoeba. After all, I have almost no knowledge of them.

If a user is actually learning a language (or studied it at some point), then "1: Beginner" sounds good enough.

+1

I noticed one of the sentences was in Javanese script and brought it up in the previous thread. Not sure if it's the only one, but here: #3937561

I'm not sure how we should handle that. Should we enforce Roman? If we allowed both scripts, that would essentially allow double sentences. Ideally we'd use a dual-script solution like for Georgian, Chinese and Shanghainese, keeping Roman as primary and Javanese script as secondary.