RE: Efficient XML - request for input

[Matthew MacKenzie <matt@xmlglobal.com>]

At 08:32 PM 3/8/00 -0800, Don Park wrote:
> >To that end, it seems that more and more people are favoring
> >element-normalization (keep data out of attributes, use attributes
> >(sparingly) for meta-information).  Most new standards are also
> >favoring elements over attributes (XSLT, XSchema, etc.).
>
>Interesting observation I made recently was that XML parsers
>parse SML files faster than XML files.  In some cases, as
>much as 40% faster.  Considering that SML versions of XML
>files are about 6% larger, performance gain is rather
>interesting.

Interesting, but completely understandable.  I am assuming you mean SMIL, 
not SML - SML is a programming language to the best of my 
knowledge.  Parsing that I have done is always slower on data that contains 
more attributes, CDATA and content, probably as a result of the in-memory 
structures that need to be created as these portions are discovered.

I advocate, from our experience parsing every sort of document imaginable, 
the use of simple, non-cryptic XML with as few attributes as needed.  I 
find attributes add weight to even an event based parse.

my $0.02, $0.034 after GST.

--
Matthew MacKenzie
CTO/VP R&D
XML Global Technologies, Inc.



--
Matthew MacKenzie
XML Global Technologies, Inc.