Am 30.04.2013 um 15:33 schrieb Georg Fette <[email protected]>:

> Hello,
> I have written a library which can save and load CASes and their type systems 
> in a relational sql database:
> http://code.google.com/p/uima-sql/
> I added a documentation and documented the main interface. If somebody likes 
> it I would love to hear some comments. Its the first project I like to share 
> with a community so don't hesitate with criticism but be not too harsh, too.

Looks interesting. Apparently even feature structures referencing other feature 
structures are supported. 

Was there any particular reason to add the covered text as a column in the 
annotation table? That seems to introduce quite some redundancy.

Adding a reader and consumer component may provide newbies with a good entry 
point to see how the stuff works.

A framework like Hibernate could help avoiding the if/else statements currently 
sprinkled around to discern between MySQL and MSSQL statements ;) Maven maybe 
by any chance?

Cheers,

-- Richard

Reply via email to