[XML-SIG] Tim Bray on XML

Mark Bucciarelli mark@easymailings.com
Wed, 19 Mar 2003 07:42:52 -0500

On Tuesday 18 March 2003 10:49 pm, Uche Ogbuji wrote:
> > On Tuesday 18 March 2003 2:01 pm, Uche Ogbuji wrote:
> > > I hardly think there's any deep insight there for developers.
> >
> > The insight I got was that when you are planning an application that will
> > have to parse _large_ xml files, it's not efficient to have to load the
> > whole tree into memory.  This makes good sense and is a common-sense
> > counter to XML-hype.
> >
> > > Yes, Tim
> > > sometimes goes beyond that, but he's always been very careful to point
> > > out that when it's time to do full, general XML parsing, he uses a
> > > proper parser, just like anyone else.
> >
> > That's not what he says in this article.
> I just had a bit of a more careful read to see if Tim Bray could realy have
> said such a thing, and I'm not sure where you got that he did.

The way I read it, the majority of his recent XML work is processing large 
input streams of XML and that regexp is the best solution he has found.

I didn't mean it as doomsday, rather as a data point that xml api's could be 

You're right that my reaction was a bit "intense."  The opinion he offers 
jibes with my own experiences and flows counter to XML marketing hype.  I've 
made some decisions and recommendations at work recently that run counter to 
a similar kind of marketing hype, so I was excited to read his perspecitve.

I'm reading up on how libxml2's XmlTextReader does it ...