Talk:جدا

Definition from Wiktionary, the free dictionary
Jump to: navigation, search

Disccusions with Stephen over the past months has had me thinking about a way to help overcome the problem with certain scripts such as Arabic and Thai being too small for him and other people. I didn't like the inline HTML font setting he experimented at the time but recent stuff with IPA and Cyrillic stress marks has given me some ideas.

I'll use this page to test the Arabic:


Arabic[edit]

Template:ARchar (djíddan)

I’m still trying it out, and so far it looks pretty good. However, now that it’s large enough to see clearly, I see a problem...but I don’t know if the problem is with the Wiki software or if it’s something on my system...perhaps my browser (IE 6) or my OS (Win2K). The problem is with compound diacritics...I can type them correctly in the edit window and they look correct, but after saving they are mangled. In the Arabic name of this article, you should see a double fatha (like an equal sign '=') above a shadda (a sort of 'w'). That’s how it appears before saving, but after I save it they are misplaced, and the = is inside and below the w. It is possible to have the = below the w, but this is a different sound (in versus an).
If this is not a bug in my browser or something like that, then I think we should avoid this combination of diacritics altogether.
Testing: Template:ARchar (- over w), Template:ARchar (- under w), Template:ARchar (waw over w), Template:ARchar (= over w), Template:ARchar (= under w).
How do these five examples appear to you? —Stephen 08:56, 12 Jun 2005 (UTC)
  • I saved this, then viewed it, and it is definitely being changed upon saving. In order to work properly, the shadda (w) always must be typed before the vowel mark. That’s how the GPos tables in the Arabic fonts are set up. But after I save the page, the order is switched so the the vowel is first and the shadda comes last. Let me try deliberately mistyping...I’ll type the vowel first and then the shadda. If it gets switched, then it should appear correctly:
Reverse order: Template:ARchar (w under -), Template:ARchar (w over -), Template:ARchar (w under waw), Template:ARchar (w under =), Template:ARchar (w over =). —Stephen 09:25, 12 Jun 2005 (UTC)
  • Nope, no difference...the order is still backwards. Apparently Wiki does not allow compound Arabic diacritics to be saved in the correct order. At least, I cannot save them using IE 6 in Win2K. —Stephen 09:30, 12 Jun 2005 (UTC)
It's probably one of two things. 1) The font in the edit window is different to the display font, and they have different quirks. 2) A couple of months ago, the Wikis started to do Unicode normalisation. This caused some problems with Hebrew because many well-known Microsoft Hebrew fonts were designed for an order of vowels and dagesh opposite that dictated by Unicode. Microsoft are apparently updating their fonts to overcome this problem. Perhaps there is a similar issue with normalized ordering of Arabic vowels, shaddas, sukuns, etc.
I recommend you try playing with the settings in Template:Arabic fonts. I'm on my work computer which is not set up for Arabic and I certainly don't know the script or the Unicode quirks as well as you but I am trying to learn. I also recommend you get the feel of Template:Arabic font size. Both of these can be overridden just for you or by any other user by customizing User:xxxxxx/monobook.css — Hippietrail 14:39, 12 Jun 2005 (UTC)

UniPad[edit]

The order is actually changing as I save the page. I can copy the saved Arabic word (from either the edit window or the display window, no difference) into MS Word, increase the size and select from any of my numerous Arabic fonts, and then I can insert a space between each of the Arabic glyphs...and the order is clearly reversed from my original "shadda-vowel" to "vowel-shadda." I don’t think messing with the Arabic fonts will do any good, because every one of my fonts requires the "shadda-vowel" sequence.
And in fact, this is the logical order. For any word with a double consonant, you have C (consonant), V (vowel), then CC (double consonant), then V again...like this: j-i-dd-a. The first consonant of the pair doesn’t receive any vowel, obviously...the vowel is only for the second of the pair...so you have to type the double consonant first (which is what the shadda does), and later put the vowel.
I suspect they got confused in the Normalization process because of right-to-left issues. They probably meant to have it save the shadda first, but got the whole thing bass-ackwards.
It could be that Microsoft was rushing to get things working before the nitty-gritty was finalized with Unicode, or they hadn't got as far as thinking about normalisation at that point. (Normalisation is important for searching as sorting but I'm sure you know all this stuff already.) The only was to see if it's different fonts rendering differently or normalisation having different ideas to font designers, is to check the exact sequence of Unicode codepoints both before and after saving the wiki page. A very good Unicode editor for this is SC UniPad available for download here: http://www.unipad.org/download/ - I would try to test it myself but I think you're the only contributor here who truly understands Arabic script and font design issues, but please report your findings.
  • I downloaded UniPad and tested جِدًّا , both before and after, and it showed what I already knew...before saving, I have j-d-{shadda}-{double-fatha}-alif. After saving, it turns into j-d-{double-fatha}-{shadda}-alif. In essence, it means we’re now typing the first consonant, then adding a vowel after the second consonant, and only then typing that second consonant after the fact...and the Arabic fonts do not allow this. Both consonants (meaning the consonant + shadda) must be on the page before any vowels are added after them.
Stretching things out a little to separate the diacritics for you, here is how I type the word "jissan": ( Template:ARchar ) ... and this is how it gets switched: ( Template:ARchar ). The font’s GPos table tells {double-fatha} what to do when it follows {shadda}, but there are no instructions for the reverse order, so one glyph overlays and obscures the other. —Stephen 11:00, 13 Jun 2005 (UTC)

I've just tried asking about the normalisation problems on IRC but the right people are not there now. I have found a very basic discussion of the Hebrew case in English here, and an extensive one in Hebrew here - is your Hebrew as good as your Russian? — Hippietrail 12:56, 13 Jun 2005 (UTC)

I filed a bug report on this if you're interested in following it. In fact you might be the best to field questions from the developers: http://bugzilla.wikimedia.org/show_bug.cgi?id=2399

CSS[edit]

By the way, I’m very interested in the User:xxxxxx/monobook.css page. Is there a template that I could use, or instructions somewhere? —Stephen 15:01, 12 Jun 2005 (UTC)
It's annoyingly hard to find good info. There are various pages on meta, you can do a search there or on Google. If you have a look at my file you'll see these important lines:
/* test my experimental font-size template for certain scripts */
.AR { font-family: Courier New; color: green; }
.TH { font-family: Courier New; color: blue; }
I only put the "color" stuff in for quick visual feedback. You'll be most interested in "font-family" and "font-size". I don't know how much CSS you know, I'm just learning myself but the standard docs are for free download from w3c.org. — Hippietrail 17:33, 12 Jun 2005 (UTC)

There is a Thai test page here: Talk:ตุ๊กตุ๊ก

The Azeri entry[edit]

Is it correct to use the Perso-Arabic script for the Azeri language? As far as I know the Perso Arabic script has not been used for Azeri since 1918 so I'm not sure if it's correct to use the script as an entry? Shouldn't this entry be named "Old Azeri" or something similar to seperate it's script with the one currently used today? I'm mentioning this because it might get confusing for some people when they see the Azeri language written in Perso Arabic script and more commonly with the Latinized script. For example the Ottoman Turkish and the Turkish entries are more clear in their distinctions. Placebo 12:07, 15 July 2010 (UTC)