Re: A bit of bomb throwing....

From: Michael McDougall <mcdougall_at_grammatech.com>
Date: Mon, 16 Jan 2006 23:32:51 -0500

Zack Rosen wrote:

>
> This is exactly what I would like to see SIMILE become, but it isn't
> there yet and without a restructuring of approach I don't see it
> becoming that any time soon. As someone representing the development
> budgets of organizations looking to employ semantic web technologies
> I assure you there is no realistic way to employ them yet. With all
> the effort and passion going into this research I don't see why this
> has to be the case.
>

I've read your blog post and I your emails here but I don't really get
it. You write in your blog that

"Researchers need to stop thinking of themselves as researchers and
start thinking of themselves as implementors."

But researchers *aren't* implementors, they *are* researchers. They are
researchers because they want to do research. Their main aim is to
advance the state of knowledge, to get published, and to keep their
funding agencies happy. Extending a description logic or creating a new
social network analyis algorithm is research -- if you do it right you
solved a problem that humanity has never solved before. Getting Drupal
to work with RDF is not usually research--it's not going to get you
published, or a PhD, or tenure.

Researchers are not some idle pool of labor looking for work. They have
specific aims of their own and they have to please the people who fund
them. David Huynh is working on a PhD. I'm sure he wants to make Solvent
a rock solid and wildly popular tool, but if it comes down to pleasing
users or finishing the PhD, he's probably gonna pick the the PhD--as he
should.

You ask "Why aren't these tools being employed in real world use cases?
Why are they being built in silo'd development environments? Why
aren't researches collaborating effectively with organizations,
businesses, and open-source developers actively pursuing semantic web
concepts? "

Because real world use cases are real real hard. Inter-organization
communication and collaboration is really hard. Software development is
painful enough when everyone involved works for the same company; when
various organizations have different agendas and needs, cutting through
the communication overhead and zeroing in on the concrete benefit and
still having enough time to do your actual research is just too hard. A
silo'd environment is much easier to cope with--you run much less risk
of wasting months of effort and having nothing to show for it. It's much
easier (and much less risky) to pick a small niche of knowledge and make
advances in that niche.

There are plenty of researchers who like to create cool tools, but most
of the time they are doing it as something of a hobby between their
paper writing and grant getting.

If you really want of team of implementors, you're going to have to hire
some. Or you can get a bunch of bobbyists interested in your use case,
but it seems like the best way to do that is to put in a lot of the
effort yourself.

As a few here have mentioned, I find it a little bizarre that you single
out one of the research groups that seems most interested in generating
running software that actually does something useful.

Maybe it would help if you were a little more concrete about what you
want to see and what tools you think are lacking. There are tools out
there--I got a decent little semantic web application going with a bit
of python, bash scripts and XSL. You can't just throw an application
together, but this *is* software--like a lot of software, it often takes
some work to get the pieces working together.

Michael
Received on Tue Jan 17 2006 - 04:32:30 EST