[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#104413: patch] Cannot read language links for my mother tongue by lynx

Hi,

At Thu, 12 Jul 2001 11:16:57 +0200 (CEST),
peter karlsson <peter@softwolves.pp.se> wrote:

> > Though it is a good idea to file a bug report against Lynx,
> > the bug report will be "wishlist" because it is NOT a bug of Lynx
> > that it doesn't support "entity reference".
> 
> Well, it *does* support parts of them, so it really should support all of
> it. All that is needed is recoding tables from Unicode to EUC-(JP,CN,TW,KR).
> If Lynx doesn't support it, it *is* a bug in Lynx.

1. There would be discussion on what approach should be taken.
   Use raw conversion table?  The tables are very large and someone
   may dislike this approach.  Moreover, there are no trustable
   common Unicode <-> JIS X 0208 conversion table.  (There are a few
   versions.  This situation would not be fixed because of cold war
   between Unicode Consortium members like MS and IBM.)  How about
   iconv() approach?  It might not be portable enough.  Use libiconv
   or other libraries?  ....

2. Even if we assume that we succeeded to improve Lynx till tomorrow,
   the newest version of lynx won't spread all over the world in one
   or two days.  At least, Woody would not be released till this
   November.

3. Lynx might not be the only web browser which does not support "entity
   representation".  For example, I imagine w3-el does not support it
   because emacs does not support Unicode yet.  So far, Debian web pages
   are not using UTF-8 because we think it is too early to do so.

Why do you dislike an idea that 

 1. Apply my patch NOW
 2. Improve Lynx
 3. Wait for the improved Lynx will be included into "stable" version
    of major Linux/BSD distributions
 4. Remove my patch


> > And more, now we have an easy and straightforward way to solve this
> > "cannot read" situation. Why not take this way?
> 
> Well, for one thing, it breaks some other parts of the logic, since adding
> these strings (even the "(Russian)" that is there today) will also see them
> added in places they should not be added. Perhaps we could add the
> language's name in the page language (so that it would say "<nihongo>
> (japanska)" on the Swedish page).

I don't think such strictness is needed.  "Japanese" in Swedish would
be "japanska".  My idea can be modified to 

    #&26085;&#26412;&#35486;&nbsp;(nihongo)

instead of

    #&26085;&#26412;&#35486;&nbsp;(Japanese)

because my focus is on using ASCII character.  Either will be OK.

---
Tomohiro KUBOTA <kubota@debian.org>
http://www.debian.or.jp/~kubota/
"Introduction to I18N"  http://www.debian.org/doc/manuals/intro-i18n/
Reply to: