XOM API for Processing XML with Java

["O'Brien-Strain, Eamonn" <eob AT exch.hpl.hp.com>]

> -----Original Message-----
> From: Elliotte Rusty Harold [mailto:elharo AT metalab.unc.edu]
[...]
> Validating and non-validating processors alike must report violations 
> of this specification's well-formedness constraints in the content of 
> the document entity and any other parsed entities that they read.
[...]
> Given Sparta's failure to satisfy this requirement, the comparisons 
> on its web page to Xerces are unfair.  Gaining speed by not 
> performing the mandated well-formedness checks is cheating.

I need to understand more clearly how well Sparta does in doing the mandated 
well-formedness checks.  Running and analyzing the results of the Oasis tests 
should be a good first step.  If Sparta is indeed lacking in this regard I 
will either document it explicitly (including when doing performance 
comparisons) or fix it.  I will do this before a "stable" release.

> More 
> importantly, it's extremely damaging to the XML ecosystem. The one 
> thing that's guaranteed in XML is well-formedness. It is the 
> foundation on which everything else rests. If you're letting 
> malformed XML through your parser, then you're polluting the XML 
> ecosystem.

Hmmm.  An interesting viewpoint.  I guess it is a legitimate concern when you 
look at the historical experience of HTML: over-tolerant web browsers 
resulted in quite a mess in the definition of the de-facto acceptable HTML 
format.

> I can't recommend that anybody use a parser that 
> deliberately chooses not to check for well-formedness. In fact, I 
> actively recommend that they don't.

Sparta was originally designed to run on very small devices like PDAs, 
watches, electronic jewelry, and small appliances.  In that context it seems 
reasonable to trade-off code-size and performance against some error checking 
for conditions that do not effect the resulting DOM model.

         - o - o - o -

By the way Elliotte, thanks again for all this feedback.  As a result I have 
added a bunch of feature requests in the SourceForge tracker:

 https://sourceforge.net/tracker/?atid=493365&group_id=60189&func=browse

__ 
Eamonn O'Brien-Strain 
HP Labs 
eob AT hpl.hp.com