Cory J. Geesaman » Distributed Computing http://cory.geesaman.com Personal Blog Sun, 08 Nov 2009 19:57:10 +0000 http://wordpress.org/?v=2.8.5 en hourly 1 NovaDB http://cory.geesaman.com/2009/10/novadb/ http://cory.geesaman.com/2009/10/novadb/#comments Sun, 04 Oct 2009 08:45:06 +0000 CoryG http://cory.geesaman.com/?p=172 I’m releasing the Nova Database project that is still underway under the Apache License Version 2.0.  The code is c and c++ wrapped in a KDevelop project.  The project is taking a bit longer than initially planned, and I may be taking on a position that would require significantly less time be put into this so it seems worthwhile to make open source and see if anything comes of it with slower development otherwise.

The project home can be found on http://www.novadb.org/

-or alternative on Google Code at http://code.google.com/p/novadb/

-or SourceForge at http://www.sourceforge.net/projects/novadb/

]]> http://cory.geesaman.com/2009/10/novadb/feed/ 0 Unified Clouds and Semantic Search http://cory.geesaman.com/2009/02/unified-clouds-and-semantic-search/ http://cory.geesaman.com/2009/02/unified-clouds-and-semantic-search/#comments Sat, 28 Feb 2009 08:40:06 +0000 CoryG http://cory.geesaman.com/?p=19 I read an article on a Google project site at http://code.google.com/p/unifiedcloud/wiki/UCI_Requirements just now and have to say it seems a bit off in it’s assumed implementation.  Following a cryptographic paradigm (once you think cryptographically you won’t think any other way) every entity has to be considered a potential threat to your software.  The initial extrapolation for this is simple, you don’t code to standard (specifically not for something defined as a semantic search mechanism), you assume noone else will and account for it.  Semantic search is especially good at this, being capable of cataloging information without regard to context by simply not holding the meaning within the search mechanism itself, rather pivoting on possible matches in a depth based mechanism.  The semantic search I have worked with is a bit in the past, having been an early project of Novorum and dying with the restructuring of the company, however I am presently working on distributed functionality for the Nova Database and the most striking component of the aformentioned article is that conforming to any such standard on an implementation level is not only wasteful operationally, but limiting in potential future expansions.  A much more realistic and limitless implementaion (since the article mentions semantic search) would be to extrapolate combinations of data based on correlary combinations of data between the two datasets.  A training time would of course be required when building interfaces between any two systems, but you wouldn’t waste time on and stiffle future development by conforming to a standard, because if standards can teach us anything its that they are never complete for 100% of implementations.  Creating a semantic data import/export daemon with a generic interface API that handles all the cross referencing itself seems much more worthwhile, and much closer to the UNIX philosophy (though of course not always the route to go, it’s great when you don’t need a GUI).

]]> http://cory.geesaman.com/2009/02/unified-clouds-and-semantic-search/feed/ 0