2014-10-03 17:35 GMT-03:00 stepharo <stepharo@free.fr>:
so a question: could we have a clear evaluation of the situation? What should be done? Who could write a clear roadmap on DBMS so that we can discuss internally and see what we can do in the next 6 months - Glorp? should be updated?
It certainly should be updated. But we first should know when the departure from the main trunk took place. We don't even know how much outdated are we (both in terms of code and releases). GLORP is developed in VW and AFAIK only actively maintained in VAST. If the maintenance of a Pharo version is too difficult in terms of coordination, the other option is to depart (fork) forever from the main trunk, as Pharo did from Squeak. There is a lot of valuable code in GLORP.
- posgres drivers?
The newer, faster, featureful, the better. I have no particular needs for the latest version, but after seeing Stephan Eggermont's description of V3 driver, it comes loaded with a lot of beatiful, useful, features.
- DBXTalk (is is useful or should it not be better to focus on native drivers.
IMO OpenDBX is a failed project. It had good intentions of becoming an ODBC-like layer for Linux. But it's adoption was really small. The issue tracker is not active at all. The mailing list had 5 emails in the last 6 months, 2 of which were written by our fellow Peck. All of the mails were related to connecting to MSSQL databases. I would go native. For which there are already some drivers (PgSQL, ODBC). Having an abstraction layer at the image level would be nice, but it would be costly, both in terms of development and performance. Regards, -- Esteban.