0xDECAFBAD

It's all spinning wheels and self-doubt until the first pot of coffee.

RSS 2.0 namespace versus RSS-Data, Part 3: Electric Boogalee

So, for the same of argument, yesterday I threw together examples of what a use of RSS-Data might look like alongside what the same data in an RSS namespace extension might look like. I promised code, but never got a chance to circle back around. Fortunately, Phillip Pearson connected the rest of the dots for me with two examples:

I was just a little surprised by his results, since I expected the code to handle an RSS namespace to be at least a bit more complex than the RSS-Data example. But, as Phillip observed later, the scripts were pretty much equivalent in length, complexity, and ease of construction.

Then, this morning, I saw that Danny Ayers had posted an example in RDF of this same data. It doesn't differ very much from my namespace extension example, except that the few differences there are enables his example to flow through RDF tools (as well as, usually, XML tools like XPath and XSLT).

In a comment on one of Phillip's posts, though, Roger Benningfield makes the point that this example is a bit biased:

I agree that there won't be a ton of difference between a struct full of strings and plain ol' XML. But that's kind of a stacked example, since SDL would allow a lot more than that... arrays, integers, and arrays of integers inside structs.

What I did could be obscuring some work. I just took an existing schema from Amazon, which gave me some initial work already for free. (Though, there's something to be said for that in and of itself.) The structures were already established, and the schema was created with XML representation already in mind. This could have placed RSS-Data at an example. While I really don't think that XML-RPC serialization offers more flexibility in expression than XML itself, I could be wrong and I don't want to be tilting at straw men.

So, while I doubt that I'll have a whole lot of time today, I think for the same of completeness, someone should go through the parallel processes of going from problem statement up through data modeling and on to production and consumption of RSS-Data and an RSS namespace extension. While doing this, capture the work involved in both.

I could see shortcuts taken on the RSS-Data side, since you don't have to be concerened with various bits of XML tech like DTDs or schema or whatnot. You can jump right into coding up an example usage and come up with your data model on the fly. Whether this is a good thing or not, I'm sure many will disagree. Also, I'm sure others would go through this differently than I would. Again, your mojo may exceed mine.

At this point, I can see the benefits of RSS-Data in rapidly cobbling together scripts, but I lean toward having a decently defined data model first. You can do this in your scripts, but using the existing XML tech forces you through some specific processes. On the other hand, I can see where some busy developers don't have time or spare brain cycles to absorb all the XML tech. It could be made easier at that end of things, which is where I'd rather spend my effort.

Anyway, I'm really interested in seeing where this goes, because this comparison of RSS-Data, RSS namespace extensions, and even RDF seems like another very concrete, non-theoretical way to demonstrate the benefits and drawbacks of these ways of thinking about data and interoperability.

shortname=rss_data_versus_namespace_3