Re: Registration of new charset "UTF-16"

["Martin J. Duerst" <duerst@w3.org>]

At 16:22 98/05/19 +0900, MURATA Makoto wrote:
> Larry Masinter wrote:
> > So far, I'm not aware of _any_ implementation of a browser _or_ a server
> > that support UTF-16, much less two implementations of each that have
> > been tested to be interoperable.
> 
> Today, I made a UTF-16 HTML document in Japanese.

The requirements on servers is obviously extremely low. The only thing
they have to be able to do is to spit out some binary data, and label it.
Spitting out works since HTTP 0.9. Labeling data shouldn't be a problem.
If somebody claims that it doesn't work, s/he probably hasn't tried
hard enough.


> MSIE4.01 and Netscape  
> Navigator 4.04 are able to display it successfully.

If we ignore for the moment the differences between UTF-16 and UCS-2
(which are not relevant as long as a document doesn't contain any
characters outside the BMP), then Tango
[http://www.alis.com/internet_products/browser/browser.en.html]
can also do it.


> I had to choose some combination of big/little and CR/LF/CRLF, though.

What do you mean by this sentence? You had to choose a particular
combination because only this was supported by the tool with which
you made the file, so you cannot claim that all combinations are
supportend on the browser side? Or you had to produce a particular
version in order for the browsers to work? If the second, then which
combination was this?


Regards,   Martin.