[uf-discuss] Re: DOM scripting as an alternative to include-pattern?

[Tantek Ç elik]

On 6/5/06 3:48 PM, "Michael Leikam" <leikam at yahoo.com> wrote:

> --- Ryan King <ryan at technorati.com> wrote:
> 
>>> To me include-patterns seem like a subset of DOM and
>>> both
>>> seem less to do with the data format itself than the
>>> inherently procedural transformation from one format to
>>> another.  What is the difference between defining a
>>> data
>>> format and defining what people do with that data
>>> format
>>> (i.e., what that data format is used for)?
>> 
>> The difference is this: with the current <object>-based
>> include  
>> pattern, *I*, the writer of the consuming agent get to
>> decide how its  
>> implemented� if we include document scripting as a
>> solution to the 
>> include pattern, then I have to run your code on my
>> server, and  
>> that's just not reasonable.
>> 
>> -ryan
>> 
> 
> Good point.  I hadn't given enough thought to the burden it
> would put on consuming agents.  Even though I think it's
> far more important to make microformats easy for content
> producers/authors v.s. easy for consumers/parsers;

Michael, it *is* more important to make microformats easy for content
producers/authors v.s. easy for consumers/parsers, and in fact, procedural
content makes it HARDER for producers/authors, not easier.

E.g. validating that procedural content is actually correct and does what
you expect it to is impossible (c.f. halting problem as Ryan pointed out).

OTOH, validating declarative content is a solved problem.

And most authors don't write procedural code.  They write (X)HTML+CSS.
That's it.

In general, to simply communicate data/information/knowledge, if you find
you need to embed procedural code, you're probably making a mistake
somewhere.


> the
> consumers, search engines and syndication are vital in
> making this whole effort worthwhile.  What's the point of
> marking up my contact information if agents can't
> aggregate, republish, reformat and use it?

Indeed.  The key here is to do so in a way which fits into the behavior
already being widely adopted by web designers/publishers (thus making it
easy for *non*-programmers to do so), which is to use semantic class names,
rather than invent new ways of doing so (XML silo files, random new URI
technologies/schemes) with new syntaxes etc.  This has determined much of
the approach that microformats has taken in contrast to other approaches.

Thanks,

Tantek