|
Why not use dcterms:modified instead of modified? It makes no sense to put an element in the Atom namespace if it is already defined in another people are familiar with.
It is funny how people want to make the syntax as simple as possible when at the end there will be tools to abstract it all away. I don't care about a simple syntax, I want what is right. If that means making Atom a little more complex and based on RDF and Dublin Core, so be it.
XML is great. One of the biggest problems people find with XML is that when it was being passed around, people are inconsistently naming and structuring data. So specifications and recommendations have been released in hopes that people would start using consistent terms. RDF and Dublin Core come along and do just that and then people complain that it is complex. Of course it is going to be complex! We are talking about trying to build consistent vocabularies and terms here!
I recently switched over one of my projects to RDF/DC and I am really pleased with the consistency it gives the metadata. I am still working out the quirks to take full advantage of RDF but once I run my documents through the RDF validator, the advantages become clear: consistent terms and lots of tools.
My final piece of advice is: don't cheat. Some XML parsers cheat and actually allow non-XML behavior to sneak in. We cannot risk cheating in implementing XML as well. Namespaces help us define our terms so let us take advantage of this feature. Let us not be scared of complexity. Chances are when a developer talks about complexity they mean the secretary-at-my work-won't-understand kind of complexity which in fact isn't complex for real developers. If it is going to take complexity to do something right then don't worry, your average Joe Blogger won't be writing his Atom in Notepad, he will be using a blogging tool.
-Nasseam
http://www.myspotter.com
http://www.opensec.org
|
Previous Message
