OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

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

Re: Two different sets of experiences about non-English identifiers

7/13/01 2:30:34 AM, XML Everywhere <host@xmleverywhere.com> wrote:

>As of yet, nobody has made a good argument
>as to why XML names are so restrictive.
>Who hasn't been surprised and a
>little miffed that values for "ID" attributes
>can't start with a number?

Seriously, the main reasons are historical.  One of the earliest design goals of XML was that it in 
fact be a profile of SGML, primarily so that existing SGML software tools could process XML 
documents.  This meant using a fixed SGML declaration for XML, which in turn required "freezing" 
certain things that were choices in full SGML, and one of those things was the set of name 
characters.  At the time, full Unicode support was still rare in SGML processors, so there had to be 
compromises in these choices.

This is actually typical of many real-world systems; they wind up with constraints that wouldn't have 
existed if the systems had been developed from scratch in a self-contained environment.  This is off-
putting to a lot of people, especially those who are just out of school because most engineering/CS 
curricula seem to leave students with the impression that the primary job of a programmer/engineer is 
in fact to create new systems from the ground up, rather than to incrementally modify existing 
systems (the former is certainly sexier than the latter, I'll admit).  But in Real Life, most _de 
novo_ creations without historical ties fail, largely because they can't get enough adopters.