xml-dev - RE: power uses of XML vs. simple uses of XML

[Matt Sergeant <matt@sergeant.org>]

On Sun, 9 Jul 2000, Simon St.Laurent wrote:

> I didn't mean to reopen the great 'are namespaces good for anything?'
> debate and I'll try to close it here quickly.

Sorry Simon, couldn't resist replying ;-)

> At 12:43 PM 7/9/00 +0100, Matt Sergeant wrote:
> >I think maybe you missed my point about "a variety of ways". If you're
> >only using XML in one domain and you don't need namespaces then there's
> >no need to gratuitously implement them. But work in different
> >areas/domains and with different technologies and eventually you're going
> >to run into a need for namespaces, unless you expand your tagset/DTD
> >beyond all sensibility.
> 
> A lot of developers only work in one area or domain or even only one
> XML-based project.  That's some of what I mean about power usage (which
> folks devoting their thoughts to XML are interested in) being different
> from simple usage ('just trying to get the job done, now get out of my way,
> please').  Power users are interested in Matt's suggestion, but simple
> users may not care, period.

This is all true. And it applies to other tools too. Take Perl as an
example. A lot of Perl "programmers" are just CGI hackers, blissfully
unaware of Perl's object oriented features, its networking API's, or its
ultra powerful tie() interface. And thats just fine because they use what
works for them. Like I said though - a perl programmer who uses perl in a
variety of ways will eventually have a need for those more advanced
features.

> Trying to explain that "you don't know where your documents will be in
> fifty years" doesn't have a whole lots of impact, either.
> 
> I've heard from a few developers that the mere fact that DTDs weren't
> namespace-aware led to a brief period of soul-searching in which they
> decided to ignore both namespaces and DTDs.  Now that's a side effect I
> didn't expect!

I'm not suggesting they should care right now. In fact thats the beauty of
it - they don't have to. I was merely saying that I don't think namespaces
should be considered in the same class of specs as say XLink and
XInclude and schemas, in fact XML 1.0 and XML Namespaces should probably
be one document (I'm assuming XML 2.0 will merge the two), with namespaces
part of the optional features of XML.

> >> I think namespaces are really only useful in a limited number of cases
> >> (where arbitrary embedding is common) unless they are being used 
> >> gratuitously.
> >
> >I feel the opposite. Namespaces are incredibly useful for modularisation,
> >and should be used to ensure that your DTD for one application is
> >sensible and re-usable.
> 
> Maybe best practice is to declare a default namespace on the root element
> of every document you're creating where namespaces "aren't being used".
> Oddly, though, I don't think I've ever heard this suggestion put forward,
> even in my own books.  (It's Sunday morning, no coffee, so I'm probably
> wrong.)
> 
> I'll have to think about that one.  While it fits well into generic
> processing infrastructures, I'm not sure its a good idea in general.

I don't see how it can hurt. A non-namespace aware processor sees an extra
attribute, which it probably just ignores. It's the X in XML.

-- 
<Matt/>

Fastnet Software Ltd. High Performance Web Specialists
Providing mod_perl, XML, Sybase and Oracle solutions
Email for training and consultancy availability.
http://sergeant.org | AxKit: http://axkit.org


***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************