Performance improvement for xml loads (+comments)

[Christopher Browne]

On Thu, 07 Dec 2000 16:13:06 +1100, the world broke into rejoicing as
Tyson Dowd <trd@cs.mu.OZ.AU>  said:
> On 06-Dec-2000, Derek Atkins <warlord@MIT.EDU> wrote:
> > Nobody is suggesting going back to the old binary format.  I'm
> > certainly not.  I *AM*, however, suggesting a NEW binary format.
> 
> Any new binary format will have to be at least as extensible as XML.
> After all, there's no point writing a nice tight binary format today,
> when tomorrow there will be another field that needs to be added.

XML has the merit of being easily serialized; it wouldn't be too
difficult to use it for THAT purpose, and using other formats that are
isomorphic to it for more direct access.

One thought would be that the IronDoc object database system was
designed with isomorphism with XML as one of the secondary goals. [See
<http://www.ntlug.org/~cbbrowne/oodbms.html> for linkage to IronDoc
info; it's vaporware, but characteristic of the sort of DB that might be
appropriate...]

> The other thing to consider is that I've heard you can generate a
> near-optimal binary representation automatically from a DTD.  If you are
> suggesting an approach like this for generating a binary format, then
> that would be just fine, because it can be extended and maintained
> semi-automatically.

I don't think this is the right answer; the "semi-automatic" part causes
me concern, in that it doesn't guarantee extensibility.

> Personally, I'm not convinced that performance of the XML routines is
> going to be a long term problem.  Besides, a lot of people feel more
> comfortable with XML (or compressed XML) than being "locked in" to a
> binary format (even if the source is available).  I'd much rather see
> improvements to the XML based system than a completely different system,
> because there's a lot of synergy to be gained by going with XML.

Indeed.
--
(concatenate 'string "cbbrowne" "@ntlug.org") <http://www.hex.net/~cbbrowne/>
As of next Monday, COMSAT will be flushed in favor of a string and two tin
cans.  Please update your software.