Re: [syndication] Thoughts, questions, and issues.

[Jonathan Eisenzopf <eisen@pobox.com>]

Paulo Gaspar wrote:
> 
> Thanks for your continuous insistence on staying SIMPLE Dave.
> 
> The more I see talk of 2 zillion XML related standards that are supposed to
> be respected at all time the more I think that XML is gradually being taken
> by the complexity that plagued SGML.
> 
I sure hope not.

> It seems that the only difference is that XML standards are getting complex
> on a piece-by-piece basis instead of all-at-once.
>
This is the result of the evolution based on user requirements
(usually).
 
> That is happening with:
>  - the XML "core" standards (I can not get enthusiastic about stuff like
>    the Infoset);
>  - the Syndication related standards;
>  - SOAP.
>
I agree to some degree, but some solutions are complex because
they solve complex problems.
 
> XML main advantage over SGML was supposed to be simplicity... or did I get
> it wrong?
> 
One of the advantages certainly. But XML was also a sneaky way to
get people to adopt SGML.

> Why not build XML standards as simple cores that allow for extensibility
> instead of trying to cover all the possible and impossible functionalities?
> 
Have you read the spec? Read section 4. The RSS 1.0 proposal is
exactly what you describe above. RSS 1.0 is simpler (in the # of
elements); it's a small lightweight core that's extensible via
namespaces. You can create a simple RSS 1.0 document with as
little as 9 or so elements. Could it be simpler or more
lightweight? RSS 1.0 was the result of user feedback and the
request for extensibility while maintaining simplicity.

> Namespaces can be convenient if one wants to build extensions... but why
> not making namespace use optional - only for those cases?
> 
Namespace use is optional (except for the default namespace).
It's used when extending the default namespace.

> I would like it so much that standards would respect a bit more the KISS
> principle even at the cost of the Holly Graal of completeness.
>
Well, RSS 1.0 is only a proposal, so suggestions are welcome. How
would you have solved the problem? What can we do to make it
simpler?
 
> Sorry if I go partially out of topic, but this obsession with completeness
> (at the cost of simplicity) that menaces RSS is a general trend.
>
I agree, sorta. But whose obsession with completeness are you
referring to? We abandoned the idea of trying to meet everyone's
needs and instead created a small extensible core so users could
extend it to meet their specific needs. RSS 1.0 was created to
meet the requirements of simplicity and extensibility at the same
time. 

-- 
Jonathan Eisenzopf    |  http://motherofperl.com    
eisen@pobox.com       |  http://perlxml.com
Perl Hacker           |  http://dc.pm.org