Zack Rosen wrote:
> I think the sweet spot for real-world RDF store applications at the
> moment are small to mid sized business like yours who have the
> resources and wherewith-all to do the deep dive required to implement
> the new technology. Companies such as yours who very much in control
> your technology destiny are in a good position to experiment as long as
> the developers who work for you are comfortable with your platform.
That could very well be true, and is probably precisely why I am working
in a small business and not an "enterprise" setting. :-)
> The problem is this work will not translate well to 'enterprise'
> providers or widely adopted open-source projects. Large companies are
> going to stick with proven technologies and open-source projects
> (successful ones anyways) are going to stick with simpler
> implementations. Upstart open-source projects will inevitably come
> about to challenge the currently successful CMS's but they will be
> fighting an uphill battle.
That could very well be true, and I don't see that as necessarily evil.
I mean, just for myself I am having a hard time deciding which RDF store
to use (Jena, Sesame 1.2.3 or Sesame 2.0?), not to mention what query
language to settle on (not even gonna try to enumerate these).
"Enterprise" settings have a tendency to create lots and lots of code,
so doing the "right decisions" is indeed critical, because they are hard
to change once the ball starts rolling. I would never ever encourage an
"enterprise" setting to use the same tactics and techniques as I do,
because we have fundamentally different situations, from an
"evolutionary perspective".
> I think the scalability issues are not the biggest obstacle to RDF-
> Store adoption although they are going to be a serious pain in the ass
> to work through. The problem is your going to have a heck of a time
> convincing open-source developers to ditch their simple, proven, and
> well understood methods of data retrieval and internal representation
> such as straight SQL, stored procs, etc. RDF stores are a new
> paradigm.
Well, that's easy then, because I'm not in the "convincing" business :-)
My perspective on this is simple. I do what I do, and others either
follow or don't. Historically speaking those who follow seem to be doing
quite well though, judging from the successes of JBoss, WebWork and
XDoclet. But I do some stupid shit as well, and this may be one of them.
In other words, the results are always what matters. Not words.
That being said, as already mentioned it didn't take much "convincing"
to get "Semantic Web" and "RDF" to be the "hot topic" of this years
developer conferences here in Sweden, so I do expect there to be more
talk about it soon here.
> For open source developers the question of adoption will
> hinge on a cost-benefit analysis of: "How much will RDF stores cost me
> to 1) understand community wide 2) implement 3) support long term VS
> how much new functionality can I leverage with RDF object stores?".
> The history of open-source web application projects is littered with
> the carcasses of "very powerful" but overly complex application
> environments. Just as Plone is being over-taken the lighter-meaner
> Drupal (http://tinyurl.com/8lf3s) Django and ROR are starting to rip
> through the overly dense and verbose world of JSP web-apps.
> Open-Source web app project leads are taking note and the conclusions
> are going to make SW technology an even harder sell
Perhaps, but there's also a trend of people wanting more and more and
more features, and there's always a limit to how far you can go with
"simple tools". If RDF tools can become both easy AND powerful, then I
think the situation will become much nicer.
But time will tell.
/Rickard
--
Rickard Öberg
rickard.oberg_at_senselogic.se
_at_work +46-(0)19-173036
Received on Wed Jan 18 2006 - 07:43:46 EST