I have no objection to the 'Data Store API'.  If you plan to really
build a Data Store API, and along side you happen to build an
experimental (or reference) implementation that uses the API and it
happens to use a 'transaction-per-directory' model, that's fine with
me.  It shows that the API works.  It also means that as an
open-source person, I can re-implement the data-store to a more
scalable solution ;)

The important thing, IMHO, is that API -- it needs to be designed
appropriately.  Even though I know nothing about SQL databases, I'm
willing to teach myself in order to provide (hopefully) valuable input
to the API design.  OTOH, I think the API is probably fairly obvious
;)

-derek

[EMAIL PROTECTED] writes:

> 
> It's been rumoured that Christopher Browne said:
> > c) By virtue of being different, it encourages the decoupling of
> >    front end from back end, which is Highly Desirable.
> 
> OK, you convinced me.  Its a good idea.
> Note, though, the 'engine' is already mostly (I want to say 'completely') 
> decoupled from the 'persistance model' (aka file format or sql back end).
> 
> The engine should be able to trivially support the
> transaction-per-directory data persistance model.
> 
> The engine may need extensions to handle random access to stored
> data (e.g. to pull out onlyu a small subset of sql records). Although, 
> come to think of it, if the 'Query' object is used consistently by the 
> GUI, (and I think it is) then tweaking shouldn't be necessary. 
> Although, reqalisitcally, tweaks will probably be needed.
> 
> 
> --linas
> 
> --
> Gnucash Developer's List 
> To unsubscribe send empty email to: [EMAIL PROTECTED]
> 
> 

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/      PP-ASEL      N1NWH
       [EMAIL PROTECTED]                        PGP key available

--
Gnucash Developer's List 
To unsubscribe send empty email to: [EMAIL PROTECTED]

Reply via email to