Re: [xml-dev] Ten Years Later - XML 1.0 Fifth Edition?

[Jonathan Robie <jonathan.robie@redhat.com>]

Elliotte Harold wrote:
> Jonathan Robie wrote:
>
>> I think most implementers use the Unicode support in their 
>> programming languages or standard Unicode libraries for parsing 
>> Unicode, because it's too much work to roll your own. It's odd to 
>> tell someone who relies on newer versions of Unicode that although 
>> the characters of their language work fine on their computer and can 
>> be displayed in their word process, we require XML parsers to check 
>> each character to ensure that they do not support these characters.
>>
>
> Really? In the case of of characters added in Unicode 3.0 and later 
> there's very little if any support for them in the major operating 
> systems. Possibly you can add it in, but it ain't easy. They certainly 
> do not "work fine on their computer and can be displayed in their word 
> process"


Perhaps Unicode 3.0 would have been the sweet spot, but that's not what 
XML currently uses.

On my operating system, Fedora 8, they seem to work fine in some word 
processors, and display fine in Firefox (at least for some languages, I 
haven't tested them all, and you admittedly do not have a standard 
keyboard driver to make input convenient in these languages). The ICU 
libraries are in wide use for applications that need to support Unicode, 
as is the Gentium font. I believe that libxml supports some, but not 
all, of these characters.

The fact that many operating systems do not support these and many other 
languages well is one of the big reasons that XML frameworks should have 
good language support, not a reason to limit support in XML.

> XML has much better support for more languages than any word processor 
> or operating system I've ever seen. (Here I do mean the end user 
> definition of an OS such as Windows, Mac OS X, Ubuntu, etc; not just 
> the CS definition of OS)

That's a feature, not a bug ;->

Jonathan