according to the Israeli standards, the "last resort" for Hebrew points and
accents is to display nothing. This means that in your table the mapping of
0590 to 05cf should be "ignore".
The rational is that in Hebrew, points and accents are always optional and
never affect the meaning of the word.
> -----Original Message-----
> From: Otto Stolz [mailto:Otto.Stolz@uni-konstanz.de]
> Sent: Wednesday, December 08, 1999 2:20 PM
> To: Unicode List
> Subject: Re: Support for Multilingual Documents
> Am 1999-12-06 um 9:57 h hat firstname.lastname@example.org geschrieben:
> > In the meantime, Web developers need to be aware of the
> "standard" level of
> > support (which means code page support, for example, on
> Windows) in order
> > to get the result that they expect. And realize that while NCRs
> may encode
> > the data in a lossless manner, the end user will probably see a question
> > mark, underscore or black block at display time.
> > Someone might argue that characters that you can't display are
> probably as
> > meaningful as displaying characters the user can't read... but in most
> > cases I find that users read that your page is "broken" somehow.
> For a character that cannot be displayed due to lack in the locallly
> available script support, browsers (and btw. any other program) should
> display the pertinent glyph from a Last Resort font, such as discussed
> in <http://fonts.apple.com/LastResort/LastResort.html>, rather than a
> question mark, or a rectangle.
> This would give the user an indication of what is missing, locally;
> it would avoid the impression that the page is "broken" somehow.
> Anyway, if the user cannot read a particular script, the pertinent
> last-resort glyph would be as meaningful to him as the original character
> (which cannot be displayed locally), in most cases -- the exception
> being that the page contains a discussion of particular characters.
> As an additional goody, the implementor could provide a context-menu to
> download, or install, the missing script support.
> Best wishes,
> Otto Stolz
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:56 EDT