On Tue, Jul 21, 2009 at 1:51 PM, Bottiger <bottig...@gmail.com> wrote:

>
> Do you mean that we will support couchdb or that the DAL will allow
> support?


This discussion should probably be moved to the web2py developers group:

I suggested that a clean, column-oriented DAL would be an interesting
experiment / excercise.
This was mostly because of discussions around GAE, how big tables api offers
things that didn't fit well into DAL abstraction, etc.

Rather than make something GAE specific, I suggested - in order to force
discovering a "useful" abstraction - to select several column stores which
were most likely to have use (e.g. testors, interested, knowledgable
parties).

Support seems highly premature in the discussion; the experiment should show
something first, and then - to me an interesting - a review of the overlap,
lack or, possibility for better overlap between (what I dubbed, for
'column') C-DAL and current, or R-DAL would make for an interesting meeting
/ session.

>From this motivation, I hesitated to throw couchDB in the mix, as I think
its different enough that it would potentially muddy the discovery of a
crisp abstraction (I would be more for investigating the abstraction by
adding something like couchDB to a second stage of the experiment -
extending C-DAL, and testing how the abstraction holds up).

Anyway, those are my thoughts and motivations.


> I've talked to the CouchDB developers and they have said that its
> currently in Alpha and they will probably be making changes that break
> backwards compatibility.


the Alpha state (which I also heard @ ChiPy) was another reason my gut is
not for throwing couchDB in the mix for first phase of the experiment... but
those who do it should decide.  I think the more important thing is creating
/ finding / working out a "correct" abstraction for column-stores, and then
comparing w/ DAL and seeing what new insights that may give us.




> MongoDB on the other hand, is nearing 1.0
> release by the summer and is now being used by sourceforge.


This, like couchDB, is classified as another "document oriented" db...  just
on that, I would hold those off to see how they play w/ column-oriented
abstraction;

Remember, the thing I would like to see is development of a clean starting
abstraction.  Mixing storage types doesn't seem like a good idea (at the
start) for that.
Maybe a separate dal for Mongo/couch ... and later merge w/ column... but
hat feels too much like competition, and not enough like investigation.... I
think investigation is what will yield interesting results.

- Yarko

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"web2py-users" group.
To post to this group, send email to web2py@googlegroups.com
To unsubscribe from this group, send email to 
web2py+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/web2py?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to