I think the key here is, when using Windows, what version of Uniscribe you are using, not the browser (also, it may depend on what font you are using - some fonts have opentype tables for diacritic positioning, others don't).

Multiple diacritics display fine for me in Hebrew, I can use the Nafees Nastaleeq font (for Urdu), bn.wikipedia and ml.wikipedia are rendered properly (bo.wikipedia renders properly in an edit window, but not viewing).

One problem is that FireFox doesn't allow the degree of control that IE does where assigning specific fonts for specific Unicode ranges is concerned. Instead, it seems it selects automatically whatever the first font it finds that supports it at all (Arial Unicode MS "supports" Tibetan, but it doesn't really - it should preferably use Tibetan Machine Uni or as a secondary choice TibetanUchen). And if a page has a language header, it will use that and ignore any specific font tags on the page - try viewing http://zh-min-nan.wikipedia.org/ in FireFox (make sure you have Chinese fonts installed first, or you won't be able to replicate this) and rather than displaying in a nice Roman font like Tahoma like it should (and does in IE), it displays in the default Chinese font. This is annoying, because although it is readable, it is EXTREMELY ugly and a bit difficult to read fluently.

I would expect with FireFox that I could have even better control, perhaps even by entering custom unicode ranges and then selecting the fonts to use for them (IE is limited to a list of Unicode ranges, which doesn't include all of them).

Mark

PS

Speaking of people simply not knowing, I got my mom using FF, but she switched back after it gave her problems - her job at the time required her to log in to the same website with multiple IDs at the same time. I asked many different people how to do that, I did what they said, but nothing worked. But in IE, it's very easy. And it seemed like people were like "Oh, did you try this? Didn't work? Oh well." which can hardly be a good thing for FireFox.

On Sun, 12 Dec 2004 04:07:19 -0500, Olve Utne utne@nvg.org wrote:

Hello Gerard M., On Sun, 12 Dec 2004 08:52:03 +0100, you wrote:

...

Just saying that Firefox is better is plain stupid when the argument IS that it does underperform. [...] So I am fully aware why there is a such a strong case for IE and against Firefox when used with Farsi and propably Urdu, Arabic and Hebrew.

I agree, and I can confirm your suspicions regarding Hebrew. I have tested recent versions of Firefox and Mozilla with Hebrew, and neither of them can handle complex Hebrew text with combined diacritics and cantillation marks. I believe it is a case of varying degrees of support for glyph-substituting in complex combinations. Opera is best on this, with IE as a good #2. A problem with Opera in Windows seems to be that it corrupts pictures when downloading them from Wikipedia. Therefore, I personally use Opera for most things, and IE when needing to download Wikipedia images.

(Another problem with Opera is that it is in some ways less up to it when it comes to CSS than IE is. In that respect, Mozilla and Firefox does a bit better than Opera, but they are both actually outperformed by IE when it comes to scrollbar properties, block-align by css in table cells, etc.)

That does not by any means mean that I am particularly fond of IE...! It is just that it is one of the two Windows browsers I know of that can actually handle Hebrew properly. Not because I am stupid or uninformed, but because I need something that Firefox and Mozilla (and sometimes also Opera) do not in fact provide.

-Olve

---

Olve Utne http://utne.nvg.org

---

Wikipedia-l mailing list Wikipedia-l@Wikimedia.org http://mail.wikipedia.org/mailman/listinfo/wikipedia-l