cocoon-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Benjamin Dixon <>
Subject Re: Multiple Transformations
Date Sat, 23 Sep 2000 21:04:57 GMT

Wow! The cahcing is a major improvement (at least for 60 seconds :>) but
I'm still mystified about how I should implement this stuff. I think our
ideas are along the same lines, I want to provide a certain set of tags to
other developers so that they never have to worry about logicsheets etc,
all the need to know about is XSL and the tags at their disposal. I don't
know how to accomplish that AND use esql. I took the XSL step out of my
pages and I noticed that the XSP/esql combo always returns an HTML
doctype. I suspect that is the source of all my problems. But if I want to
dynamically select things out of a database, process some of them (format
dates, phone numbers, etc) BEFORE they arerive at the stylesheet, I have
no clue how to go about it. I wish there were examples of more complex
setups lying around.


On Sat, 23 Sep 2000, Donald Ball wrote:

> > 1.) Is esql generally a lot slower than the regular sql? I understand
> > maybe there's an extra processing step, but 10 times slower?
> no, it should run at about the same speed if not somewhat faster. it would
> seem that your stylesheet is messing up somewhere. also, if you're looking
> to maximize performance, you should really implement the
> hasChanged(Object) method in your xsp pages. for my database pages, i have
> a timeout method that caches stuff for 60 seconds. almost two orders of
> magnitude improvement in performance - cos you not only lose the xsp
> execution step, you also lose anything (e.g. xslt transform) that occurs
> thereafter. do a search in the archives for 'cache madness' if you want a
> complete example of how i did this for one site.
> to debug your stylesheet problem, i'd lose the xslt process step and see
> how the sql results differ from what your stylesheet expects.
> > 2.) My goal is to separate content and logic as completely as possible so
> > that I can write the logic and provide a sort of taglib for others to
> > write up the formats. What other options would I have besides
> > XSP->esql->XSL? Would it be beneficial for me to use more logic and simply
> > skip the esql step? 
> i generally provide my authors with a set of elements in a site-specific
> namespace, e.g.
> <goheels:articles subject="Basketball" number="10"/>
> which gets transformed by the goheels logicsheet to a bunch of elements in
> the esql namespace (or whatever), and so forth and so on. i really like
> this approach because the elements that authors work with mean something
> at a very high level. here, for instance i'm saying i want the ten latest
> articles on basketball. the source and retrieval mechanism are specified
> by my (centralized) goheels logicsheet. the one caveat to this approach is
> that if you're using namespace-prefix invoked logicsheets, _all_ of them
> must be declared by the initial page - which is a big liability since the
> initial page shouldn't _have_ to know anything about how the data is being
> generated. this _is_ a known liability in the cocoon1 xsp implementation
> and, i'm fairly certain, has been removed in cocoon1.
> cocoon1's xsp implementation is kinda hokey in another respect which i'm
> sure you're realized - tying logicsheets to namespace _prefixes_, not
> namespace _uris_. i believe this has been rectified in cocoon2 as well.
> - donald
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

Today's Random Quote--------------------------------------

 Dig a well before you are thirsty. 
   Chinese Proverb

View raw message