Usenet 1994 October/usenetsourcesnewsgroupsinfomagicoctober1994disk2.iso/std_unix/volume.28/text0094.txt

Text File | 1992-08-17 | 1.0 KB | 27 lines

Submitted-by: [e-mail] (Peter Wisnovsky)

In article <[e-mail]> [e-mail] (Dave Decot) writes:
>What kind of OS support do you envision is required?  Generally, applications
>should be written using libraries that are codeset independent anyway.

At the meeting, more elaborate algorithms for collation sequences than
those that exist in existing standards were discussed. Also, for doing
string matching one needs to "normalize" a Unicode string, because
there are a number of different representations of a string that
displays the same, as well as some normalizations that affect display
but may be desirable anyway (display Kanji numerals as Arabic ones,
for instance). Also, if you need to calculate the display width of a
string you would want to have an OS function to do it, since the
rendering algorithms can be quite complicated.

Peter Wisnovsky
--

UUNET:   [e-mail]
US Mail: ShareBase Corporation - 1919 Addison Ave. - Berkeley, CA 94704
Phone:   [phone]


Volume-Number: Volume 28, Number 96