I haven't used this extension yet but I have plans to use it.

I think you should consider SMW to be a "cache". If you want to query data in the cache you'll first have to cache it, e.g., load the pages that reference external data and then it will be available for queries.

SMW can be considered a convenient, high level interface to your data. It comes with a lot of overhead and won't scale forever. If you plan to use it extensively you'll need to provide just as much, if not more, compute resources to it than to your external data store. Your external store needs memory, SMW needs cpu.

On Wed, May 27, 2009 at 1:26 PM, don undeen <donundeen@yahoo.com> wrote:

Mostly a question for Yaron, but the broader issue may have more interest.

I was checking out the External Data extension, as it seems it could be very useful to us. We've got data in other stores, that we want to be able to reference from within our wiki.

Let's say, for example, I've got a page, Category:ArtObject, named 1992.45.34. And we've got an relational database (called MST), that tells me that art object was made by "Fabio Ficasso",

Using the External Data extension, I see how I can load the data from our art database, using an API like
http://ourmuseum.org/api/getObjectData/1992.45.34 , and then parse it onto the page accordingly.

What we're really after, however, is to go one step beyond that. I'd like to be able to write an ASK query on some other page, that queries, say, for pages of category:ArtObject,  which were made by "Fabio Ficasso". So what would need to happen is for that external data to be queried as well.

Put another way, we want to be able to use the wiki itself as our data integration point. Insane? You tell me.

I'm also looking for any halfway measures that anyone can propose. The External Data extension is itself a huge step towards that halfway point; I'm wondering how much farther we can push it.

If I imagined the SMW backend as a literal triple store (instead of a mysql database) supporting remote endpoint queries, I could sort of imagine how the backend to all this could look.
Judging from this thread:
http://sourceforge.net/mailarchive/message.php?msg_id=9984a7a70709231041s16e90db9r7a5cfd2848347e85%40mail.gmail.com

It seems people don't think much of the idea of replacing the MW backend with a full-fledged triplestore. Am I right? Performance issues (temporarily) aside, doesn't it seem like a better fit for SMW to sit on a triplestore? Performance issues NOT aside, perhaps not so much?

Are there some other approaches people have taken to encorporating non-wiki data into their wikis?


thanks again, for all your time and hard work.






------------------------------------------------------------------------------
Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT
is a gathering of tech-side developers & brand creativity professionals. Meet
the minds behind Google Creative Lab, Visual Complexity, Processing, &
iPhoneDevCamp as they present alongside digital heavyweights like Barbarian
Group, R/GA, & Big Spaceship. http://p.sf.net/sfu/creativitycat-com
_______________________________________________
Semediawiki-user mailing list
Semediawiki-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/semediawiki-user