Re: [syndication] Re: Thoughts, questions, and issues.

[Eric Bohlman <ebohlman@netcom.com>]

On Wed, 16 Aug 2000, zac wrote:

> Aaron Swartz wrote:
> > You also may want to start using a real programming environment. ;-) Just
> > kidding!
> 
> Kidding aside I think this is a actually a bit of an important point. I
> would hope that when people look at extending XML and XML based formats that
> they stop to think of the effect that those changes will have  people what
> don't use C or Java or link to the MSXML DOM object.

You keep talking about "C or Java" yet there are plenty of scripting
environments (I already named just three of them) that are *far* easier to
use and learn than such system programming languages, and which have
complete XML parser support.

> I think comments like Eric's observation that
> 
> > so it's only the *very* desperate Perl
> > hacker (i.e. the one doing CGI work on a super-cheap Web host that won't
> > do more than bung a copy of an old Perl executable up on their site) who's
> > hurt by this
> 
> is actually quite worrisome.  I don't mean to imply any ill will or negative
> connotations on Eric's part but people have to realise that these formats
> (and XML in general) are being used by people that either don't have access
> to the same sort of tools that you do or who don't want to be bothered with
> having to use / understand / implement something like namespaces when all
> they want to do is parse a list of news items and display it on the screen.

I'm not sure what this "don't have access to" means.  All the tools I've
mentioned are free (both in the monetary and the Stallmanian senses) and
widely used.  As for not wanting to be bothered with learning namespaces,
that's a rather short-sighted attitude, since a little learning up front
can result in a big reduction in repetitive work down the line.  The
problem is that a lot of people want to do more than just display a list
of news items on the screen; if that were all anyone wanted to do, there
would be no reason to even use XML; a very simple flat-file format would
suffice.

I seriously hope that you aren't hack-parsing XML, because hacked parsers
generally omit certain details and as a result they become dependent on
things that are actually variable (e.g. assuming that an opening and
closing tag will be on the same line, or that attributes are in a certain
order) or can't cope with legitimate constructs (for example, defining and
using entities).  XML parsers that truly grok XML are widely available; in
the very worst case, you can use a command-line PYX processor (available
in Perl from Matt Sergeant and in Python from Sean McGrath, and possibly
now implemented in other languages) to transform the XML into a
line-oriented format.  But trying to parse RSS the same way that Netscape
and MSIE parse HTML lies on the road to disaster.