Forum OpenACS Q&A: Response to PostgreSQL Large Object functionality
And in PostgreSQL 8, there will be a feature....
I truly appreciate the work being done on PostgreSQL, and on the various future extensions that it will support, but it's difficult for me to tell my employers that they have to wait 6 months or 2 years or however long it will take PostgreSQL to support good large objects, before they can have their Web site.
And, even if I was capable of helping with the PG stuff, which I'm not, I'm supposed to be working on the Web site, not the database ;).
I would argue that the development of the driver should partly focus on supporting current features in the database so that non-C-hacker-type-pople can use the database to the fullest. Tying the addition of new features in the _driver_ to the _future_ _reimplementation_ of the feature in the database (however good and necessary that reimplementation will be) isn't going to attract developers, because they need the functionality *now* and will accept the shortcomings.
(Brings to mind that poor guy on the PG mailing list who was using PHP to split his text fields into 6k chunks to save them ;).
I eagerly await the new implementation of large objects -- and my employers are very happy with my current implementation of their Web site.
P.S. I now completely agree that these extensions have no place in future pg-as distributions! But I don't see any harm in my making them available ;).