[
Lists Home |
Date Index |
Thread Index
]
The fast and usually correct answer is: it depends.
And that ain't no help. Off the top:
DOM - usually the worst solution, IMO. It leads
to an excess of exposed scripting, lots of memory
overhead, etc. I'm not sure when DOM is the
right solution. I'd like to hear opinions on that.
Data Binding - expensive. Fast. Proprietary
in many cases (could lead to lots of one-off
code) but well understood and keeps as much code
hidden as possible.
XSLT - the best of the worst. One assumes a
processing framework exists, is adequate, is
tested, is fielded ubiquitously. We had transforms
before XSLT, but we didn't have HTML. HTML is
still the binding force of the web universe.
I'm not sure what pipelining adds except
more stuff. When does one need it? That
is another "opinions welcome' topic.
len
From: Jeff Lowery [mailto:jlowery@scenicsoft.com]
> See this insightful article from Sean McGrath:
>
> http://www.propylon.com/news/ctoarticles/Zen_and_the_Art_of_Mo
> torcycle_Manuals_20020822.html
Yeah, I read this. In this case, though, the markup decisions have been
done. The question is how best to process the markup. And there is no one
answer, but maybe there's a trend.
I can see Sean's point about XSLT not being hte One True Way. I tried
number 3 already (but in reverse): if you're docs are in any way similar,
then you have one helluva XSLT script to write. On the other hand, what's
better? DOM? Data binding?
I see pipelining as an enabling technology rather than a solution in itself,
but maybe I have blinkers on.
-----------
> 3. A transform (from proprietary serialized format to XML interchange
> format)
|
