- From: David Woolley <david@djwhome.demon.co.uk>
- Date: Tue, 19 Oct 2004 21:47:18 +0100
- To: www-html@w3.org
> Edward Lass <elass@goer.state.ny.us> wrote: > > should prepare a number of different versions of a large document, just > because it happens to contain a couple of phrases in a couple of > different languages? This is just the reason we need a mechanism like > "<alt>" that isn't tied into the external objects. Generally, if a document has phrases in a non-dominant language, they are intended to be viewed by users of the dominant language. > <alt> > <span>石川 雅康 (ISHIKAWA Masayasu)</span> > <span>ISHIKAWA Masayasu</span> > </alt> Firstly a computer language issue. This syntax basically breaks any attempt to write a straightforward formal grammar that enforces any sort of contents model, as you are basically allowing an ANY element as the content for every single body element type. Also looking at this example: - if this were implemented, I would either think the person's name was Shi Chuan Ya Kang, or guess that it was Japanese and have no idea how to pronounce it. Having the font doesn't mean that you speak the same language. In this case a (simplified) Chinese font would display the characters correctly, but a user in the PRC would understand the characters to have the pronunciation that I've indicated (using Pinyin rules); - a browser might not know the capabilities of a remote terminal or printer; - I think this example would be presentational, as both pieces of information are valid information (many Chinese people use a westernised name in latin script, so that couldn't be deduced from their Hanzi name, but even here it tells a Chinese user how to pronouce the name). In practice, I think that the authors of this page were showing off, and given that the language of the page was English, the Kanji version should have been parenthesized, not the latin alphabet version. > display it (I don't think there are many people who have a font > installed that contains all the Unicode characters). I don't know what I imagine that all Windows XP users in the far east are most of the way there (I think that the CJK fonts are optional for western installs, although they are still on the distribution media, so there are no big downloads invovled). Indic fonts seem to be standard for all XP installs. I'm pretty sure that Cyrillic would have displayed correctly on Mozilla or IE on any standard install of XP. > you mean by saying "a user agent couldn't be expected to reject the > Cyrillic letters". The user agent has no other alternative, if it > doesn't support Cyrillic, and it is craving for instructions by the > document author of what to do. You are thinking presentationally. Structural XHTML (i.e. XHTML 2.0) are not about telling the browser what to do, but about describing the structure of the information. > # Better internationalization: since it is a World Wide Web. I think one can assume that the next generation of desktop system will have full Unicode fonts installed by default.
Received on Tuesday, 19 October 2004 20:47:25 UTC