March 17, 2006

Genesis III

I'm arguing with Danny Ayers again.

My latest take on tools vs. formats.


Oh well, I'm getting repetitive, but still, as everyone's still OPML-bashing. :-)

Getting data out of OPML and RSS 2.0 is SynWeb rather than SemWeb, because the decision as to what the information "means" is decided by the scrapers (following usage conventions) rather than by explicit URIs added by the data's creators.

Tools trump formats and processes. The funny thing is not simply that this is *true*, but that the entire history of computer science can be interpretted as one long war between pragmatic tool builders and idealistic format / process builders. And the tools win every time. And the idealists *still* haven't noticed.

A format without a tool is like a technology without a business model. It has no interface to the wider, user, consumer community. It is not ecologically "fit". It can't feed itself on user attention or money but must be kept on life-support by optimists.

The SemWeb people are great programmers. They spend all their time writing code. How come they never actually produce any "tools"? Because all the tools they produce are explained like this : "this is a great tool because it uses a better format / process". That's not a "tool" or an "application". Because tools / applications have an inner-world of their technology, and an outer-world of their usage. And they have a story that mediates between the two. That explains *why* this technology is necessary for that usage.

SebWeb programs have an inner-technology *and* an outer usage. But they have no story that joins the two together. A story that says "this technology is *also* for that usage, but better (by some internal criteria)" isn't the same thing. This is why they must be parasitic on the story provided by real tools like the OPML Editor and various RSS generators, consumers.

3 comments:

Scribe said...

Alas, I'm not really up on the whole situation of the SemWeb stuff, and I'll be heading over to the referenced post after this, but I'm going to take slight issue with your post on a theoretical basis anyway, cos I know you love that ;)

These 2 are slightly different, I argue:

1. "Tools trump formats and processes."

2. "A format without a tool is like a technology without a business model."

I agree with the second, but that doesn't necessarily lead to the first. Rather, tools and formats/processes are 2 sides of the same "individual-group" translation/standardisation coin that make an "idea" live or die. The opposite of (2) is also true - a tool without a format, or a purpose, is equally useless.

Formats and processes are "tools" too, in this way - tools to enhance and facilitate interoperability, links between actors. Tools provide the translation between the individual and this substrate. Hence you need both for a system to transcend the 2 levels of fractality.

Maybe you synweb/semweb tribes should learn to get along more ;)

Composing said...

Well, I'd say a tool without a purpose isn't a tool at all. It's part of the concept of "tool" that it has a purpose for which it is "fit".

And I'm not saying that formats are unncecessary, they're vital. In this sense, there are two sides or parts of the same whole. What I am saying is that their "goodness" (however defined) is of secondary importance to the goodness of the tool.

You really do see this throughout the history of computer science. Think how many superior programming languages lost out because the libraries weren't there, or superior operating systems lost out because the applications weren't there etc. Theoretical beauty, especially defined by the creators in advance, never trumps practical application.

Composing said...

"Maybe you synweb/semweb tribes should learn to get along more ;)"

In other places, maybe. But this is Platform Wars. Think of it like "Robot Wars", commented by a breathlessly excited Craig Charles. :-)