W3C home > Mailing lists > Public > www-international@w3.org > October to December 1996

Re: Accept-Charset support

From: Drazen Kacar <Drazen.Kacar@public.srce.hr>
Date: Thu, 5 Dec 1996 22:17:10 +0100 (MET)
Message-Id: <199612052117.WAA22144@jagor.srce.hr>
To: mduerst@ifi.unizh.ch (Martin J. Duerst)
Cc: Drazen.Kacar@public.srce.hr, erik@netscape.com, Alan_Barrett/DUB/Lotus.LOTUSINT@crd.lotus.com, www-international@w3.org, bobj@netscape.com, wjs@netscape.com, Chris.Lilley@sophia.inria.fr, Ed_Batutis/CAM/Lotus@crd.lotus.com
Martin J. Duerst wrote:
> On Thu, 5 Dec 1996, Drazen Kacar wrote:
> To decide whether an UI for Accept-Charset is needed, we mainly have
> to examine software that does the right thing. Adding an UI to broken
> software is not usually the right solution.
> What I would immagine a good browser or similar software to do is
> to examine the available resources, in particular the fonts and
> maybe translation tables. On X11, this would mean doing something
> like xlsfonts, checking the encoding part of the long font names,
> and concluding from that that e.g. a font for iso-8859-2 is available,
> or for a more sophisticated browser, whether the display resouces
> for iso-8859-2 can be patched together somehow. If a browser finds
> such a font, it can safely add iso-8859-2 to its "Accept-Charset"
> list.

NSN does check. When you want to select perticular font for some charset,
you won't be able to select if NSN can't find fonts. That's quite
logical. But this is Unix. I don't know about other environments.
Can application on Windows or Mac find out for which code page is a
particular font made?

> Just a small comment: Latin 2 is a synonym of ISO 8859-2, and should
> not mean anything else on any platform.

I know that, but I don't know which term to use to group "all code
pages that contain certain characters for a particular language".
Nobody here uses Latin 2 in other sense. Practicaly everybody use numbers.
Novices will say "that thing used on Windows" or "that thing used on Unix" :)

> There is of course much need for accept-charset, as opposed to
> a UI for it, where I don't see the need, because the software
> should be able to decide what it can accept by its own.

Depends. You are obviously thinking about GUIs only. Next version of Lynx
will have Unicode support. The code is written, but needs a little
testing. Most of the tables are written, but not all. In particular,
mapping of Cyrillic to Latin and the other way round is missing.
Environmnt on which Lynx is usually being run is not able to show the full
Unicode. Browser will be able to translate from one code page to another,
but it can't decide which q value is the best in all cases.
This is not a problem, anyway. Lynx has UI for this and will continue
to have it.

> At the Sevilla workshop, I was rather depressed than impressed
> when Eric showed the list of "charset"s that Netscape is currently
> (or will be) accepting. Sorry I didn't speak up at that moment.
> The list was long, and it contained lots of what I would call
> garbage. I cannot blame Netscape for trying to accept whatever
> servers seem to send. But in any case, please show this in a

The server will send whatever appears to be in the document. Most of the
garbage today is produced by HTML editors.

Life is a sexually transmitted disease.

Received on Thursday, 5 December 1996 16:20:18 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 21 September 2016 22:37:16 UTC