[clean-list] Transparent Persistence using a "natural" store

Mark Koenig questionmarkfarm at gmail.com
Sun Mar 14 15:51:00 MET 2010


Greetings

I am so very new to FP and just signed up to Clean.  I hung out around
Haskelites forever wondering if they ever got around to the real world, then
discovered Clean and the I/O library!  Ive died and gone to software-heaven.

But that is tempered by what I still want (we're never satisfied, right?).

I want a development environment in which any type, function or value can be
persisted without any fuss.

OK operations would include: OPEN, and CLOSE: just housekeeping.  ROOT might
be needed to link code to DB.

START TRANSACTION, and END TRANSACTION because I can't see a system
understanding what should go inside a transaction and what should not.

But other than these exceptions the compiler and GC would generate any
additional code required to "make it so".

Many current projects implement such concepts with an underlying RDBMS.  God
forbid CLEAN would stoop to such.

It seems so logical to use a graph database for the store.  Ive been
researching many emerging commercial and experimental products and believe
the Hypergraph project using Berkeley DB may be the best fit.
http://www.kobrix.com/hgdb.jsp


So is there anyone with a Hypergraph interface library?  And what are the
limitations of using a Java interface?

Anyone implementing real world Transparent Persistence for CLEAN written in
CLEAN?

QuestionMark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/clean-list/attachments/20100314/2377700f/attachment.html>


More information about the clean-list mailing list