tinytim-discuss

[tinyTiM-discuss] Ideas for a persistent backend for tinyTIM

[Stefan L. <li...@ap...>]

Hi Lars,

as i went through tinyTIM 1.5 i saw that you are using a pluggable Java Collection Framework
(java/trove for now). I remembered some project that was building a persisten JCF with a JDBC
backend. I did not found it directly but some similar stuff e.g. the berkeleyDB JE [1]

so maybe think about this easy way to persist tinyTIM in the future or lets extend tinyTIM using
JPOX or hibernate

what do you think about a persistent layer to tinytim 1.5/2.0

Stefan


[1]http://www.oracle.com/technology/documentation/berkeley-db/je/collections/tutorial/intro.html

Re: [tinyTiM-discuss] Ideas for a persistent backend for tinyTIM

[Lars H. <he...@se...>]

Hi Stefan,

[...]
> what do you think about a persistent layer to tinytim 1.5/2.0

This is an interesting idea. But I am afraid that it would complicate
tinyTiM since we'll have to provide transactions and so on. tinyTiM
wouldn't be sooo tiny then. ;) What I personally like about tinyTiM is,
that it sooo easy to deploy. Just add a few libs and you have a TMAPI
compatible Topic Maps engine. It would be great if we could make the
backend plugable. I am not sure if that is possible. On the other
hand, we have TM4J as Open Source Topic Maps engine which supports
persistent topic maps. And it seems that TM4J will be developed
further in the very near future. Xuân has already developed a TMDM
compatible version (it seems to support JDO, not sure).

Having said that, a persistent backend has currently no high priority
for me, but we should consider it.


Best regards,
Lars
-- 
http://www.semagia.com