[Top] [All Lists]

Re: [ontolog-forum] End of the line for triple stores

To: sowa@xxxxxxxxxxx
Cc: Mills Davis <mdavis@xxxxxxxxxxxxxx>, "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Kingsley Idehen <kidehen@xxxxxxxxxxxxxx>
Date: Thu, 23 Jul 2009 15:42:16 +0100
Message-id: <4A6876C8.2000109@xxxxxxxxxxxxxx>
John F. Sowa wrote:
> Kingsley,
> I certainly agree with that point:
> KI> You **shouldn't**  need a data center the size of an
>> airport to play ball on the Web :-)
> There are many applications for which a triple store is useful,
> and there is no reason why applications that don't have huge
> data requirements can't use them.
True. And if you need to deal with large data sets, just pick a store 
that can handle large data sets :-)
> But the main point I was making is that the ANSI/SPARC approach
> to the conceptual schema from 1978 was correct:  the logical
> view of data should *not* be tied to a particular technology.
> The idea of triples for RDF and OWL was justified by the
> claim that triple stores are efficient.  But that is a
> short-sighted view that ignores the following point, which
> I repeat from my previous note:
> JFS> There is always somebody with a bright idea just around
>> the corner who will find an even better algorithm.  The
>> fundamental principle is that you should never distort your
>> logical view to fit a specific physical representation.
> KI> Virtuoso is a multi-model DBMS engine.  And when it comes
>> to MapReduce we believe and can prove that the hybrid
>> approach is better.  
>> MapReduce disregards DBMS technology innovations
>> inappropriately, and the Linked Data Web is how some of this
>> will ultimately crystallize.
> I'm happy to concede the point that there are better
> technologies than MapReduce.  Stonebreaker also believes that
> the Googlers ignored the latest DBMS technologies that make
> relational DBs run circles around MapReduce.
Correct!     (01)

And we are totally in agreement with the dear Prof. :-)
> But my main point is that there will always be better
> technologies of all varieties, and the idea of tailoring
> the logical view to any particular physical implementation
> is a bad idea.
> However, I recognize that there are many people who have
> invested a lot of time, effort, and money in the current
> versions of the SemWeb.  Therefore, any new version should be
> upward compatible, so that the current implementations can
> interoperate indefinitely with whatever new versions come out.
> But I strongly believe that the new versions should finally
> take to heart the wisdom of Knuth and others from the '60s:
Amen! (for this one I wish the was something after Amen :-) ).    (02)

> 1. Premature optimization is the root of all evil.
> 2. There will always be a better algorithm that will outperform
>    anything that anybody is doing today.
> John Sowa
We are in agreement, my ire is really directed at the over hype coming 
from the Mapreduce and so called "No SQL" brigade :-)    (03)

BTW - I wrote about the demise of RDBMS primacy a while back [1].    (04)

Links:    (05)

1.  http://tr.im/tGEx - Why the RDBMS value pyramid downgrade is nigh!    (06)

--     (07)

Regards,    (08)

Kingsley Idehen       Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO 
OpenLink Software     Web: http://www.openlinksw.com    (09)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (010)

<Prev in Thread] Current Thread [Next in Thread>