Re: unified db?

"Stig S. Bakken" wrote: > > Lukas Smith wrote: > > > > Well we should hold off until the next release for benchmarks because > > the next release will include bulk fetching as was mentioned in the > > release information of the last metabase release (actually one function > > call will be able to send the query, fetch the results, and convert any > > of the returned data if needed - of course this can also be done step by > > step). > > > > There will be other stuff in there as well but Manuel is the best person > > to ask for this. But let's hold off with discussing these new features > > until the next release because I know otherwise this thread will > > degenerate into a flame war again. I think with the next release > > Metabase should have overcome pretty much all of the (major) criticism. > > > > I was mainly looking for comments on the new OO interface. I want to > > hear if people think that this is the direction where they can deal with > > Metabase forming the basis for the new DB abstraction layer. > > The API is what's important to me and to people using DB. If we can > keep the API but make it "powered by Metabase", that's fine with me. > > > If yes I will try and make some time available to pearify Metabase after > > the next release. I will definitively not want to work on something that > > will start off with fewer features than Metabase. We have actually > > already started our port to Metabase in December in my company so right > > now I have 2 abstraction layers in my system (Metabase and our old > > proprietary db wrapper) and I want to clean this situation up ASAP (with > > the new Metabase or a pearified Metabase). > > > > If not then I will focus any future development on Metabase. Of course > > this will not prevent of taking up whatever merging whatever opensource > > DB abstraction layer they see fit to form the basis of a full featured > > DB abstraction layer > > I'm positive to this, let's hear it from other DB contributors. Tomas? > Sorry to not answer before, this week I was terrible busy. About pearifing Metabase, personaly I won't take part directly on that. I like PEAR DB because is a modern library with clean code, a unified cool error handling system, the eficient way it resolve abstraction problems and also because the kind of people that form the core team, contributors and users. So I prefer to continue supporting DB. What I could indirectly help on that is giving ideas on the new methods or API modifications that this Metabase adaptation could need. I will try to adapt the posible changes to PEAR DB if the people find them good. I surely give my best wishes to that project and I will be very close for tracking the progress. Tomas V.V.Cox