Comments on: Scalability of the Semantic Web http://www.mkbergman.com/227/scalability-of-the-semantic-web/ Mike Bergman on the semantic Web and structured Web Mon, 06 Feb 2012 00:15:33 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: AI3:::Adaptive Information » Blog Archive » Methods for Semantic Discovery, Annotation and Mediation http://www.mkbergman.com/227/scalability-of-the-semantic-web/comment-page-1/#comment-34920 AI3:::Adaptive Information » Blog Archive » Methods for Semantic Discovery, Annotation and Mediation Tue, 06 Feb 2007 17:01:29 +0000 http://www.mkbergman.com/?p=227#comment-34920 [...] Once all of these reconciliations take place there is the (often undiscussed) need to index, store and retrieve these semantics and their relationships at scale, particularly for enterprise deployments. This is a topic I have addressed many times from the standpoint of scalability, more scalability, and comparisons of database and relational technologies, but it is also not a new topic in the general community. [...] [...] Once all of these reconciliations take place there is the (often undiscussed) need to index, store and retrieve these semantics and their relationships at scale, particularly for enterprise deployments. This is a topic I have addressed many times from the standpoint of scalability, more scalability, and comparisons of database and relational technologies, but it is also not a new topic in the general community. [...]

]]> By: Mike http://www.mkbergman.com/227/scalability-of-the-semantic-web/comment-page-1/#comment-8634 Mike Sun, 06 Aug 2006 23:00:44 +0000 http://www.mkbergman.com/?p=227#comment-8634 Danny's comment is interesting, as our most all of his observations. I think the difference of opinion is one of use and perspective. No one is saying that meaningful things can not be done (or can not be "useful") with triples in a semantic Web sense using conventional DBs. But from the perspectives of both: 1) enterprise applications (which is where I mostly reside) with mucho metadata and entities, scalability is an imperative; and 2) over the long haul metadata and attribute IE and richness will (I believe) become ubiquitous for the characterization and management of document-centric information. I like baby steps; baby steps are fine. But I also like to think what the strapping child might look like when it heads off to college. Danny’s comment is interesting, as our most all of his observations.

I think the difference of opinion is one of use and perspective. No one is saying that meaningful things can not be done (or can not be “useful”) with triples in a semantic Web sense using conventional DBs. But from the perspectives of both: 1) enterprise applications (which is where I mostly reside) with mucho metadata and entities, scalability is an imperative; and 2) over the long haul metadata and attribute IE and richness will (I believe) become ubiquitous for the characterization and management of document-centric information.

I like baby steps; baby steps are fine. But I also like to think what the strapping child might look like when it heads off to college.

]]> By: Danny http://www.mkbergman.com/227/scalability-of-the-semantic-web/comment-page-1/#comment-8627 Danny Sun, 06 Aug 2006 20:42:26 +0000 http://www.mkbergman.com/?p=227#comment-8627 What Henry said ;-) Triplestores are only conceptually only really a cache of a particular part of the Semantic Web. The Web doesn't depend on hugely scalable DBs to function (because it's distributed) and the same applies to the Semantic Web. Check timbl's Tabulator, it only remembers maybe a couple of hundred statements every session, yet is a useful browser. Additionally existing databases (of any size) can be exposed on the SW, either providing data in RDF formats or (better still) via the SPARQL query language. (This point was raised in relation to your post <a href="http://chatlogs.planetrdf.com/swig/2006-08-06.html#T16-39-54" rel="nofollow">over here</a>). That's not to say triplestore scalability isn't nice to have, it's just not a prerequisite for the Semantic Web, your main argument here is something of a strawman. What Henry said ;-)

Triplestores are only conceptually only really a cache of a particular part of the Semantic Web. The Web doesn’t depend on hugely scalable DBs to function (because it’s distributed) and the same applies to the Semantic Web. Check timbl’s Tabulator, it only remembers maybe a couple of hundred statements every session, yet is a useful browser.

Additionally existing databases (of any size) can be exposed on the SW, either providing data in RDF formats or (better still) via the SPARQL query language. (This point was raised in relation to your post over here).

That’s not to say triplestore scalability isn’t nice to have, it’s just not a prerequisite for the Semantic Web, your main argument here is something of a strawman.

]]> By: Atanas Kiryakov http://www.mkbergman.com/227/scalability-of-the-semantic-web/comment-page-1/#comment-4900 Atanas Kiryakov Thu, 22 Jun 2006 00:02:21 +0000 http://www.mkbergman.com/?p=227#comment-4900 An update regarding <a href="http://www.ontotex.com/owlim/" rel="nofollow">OWLIM</a>, which seems relevant to this discussion: recent evaluation demonstrated that its BigOWLIM version can handle a billion of triples, namely to pass the LUBM(8000,0) benchmark, including some OWL reasoning. Please, check the <a href="http://www.ontotext.com/news.html#bigowlim06" rel="nofollow">press release</a> An update regarding OWLIM, which seems relevant to this discussion: recent evaluation demonstrated that its BigOWLIM version can handle a billion of triples, namely to pass the LUBM(8000,0) benchmark, including some OWL reasoning. Please, check the press release

]]>