Thanks for the responses! (combined reply)
I'm just now hacking together yet another Ruby CouchDB library. It's called CouchRest and is based on Heroku's RestClient as recommended by Noah Slater.
I just took a look at that and it's a lot simpler than I thought! I'll definitely check out rest/client, maybe that's what I wanted all along.
You might try looking at CouchObject or CouchRest to see how the thinner (less code) libraries handle CouchDB. As Johan said, the CouchDB HTTP API is so simple and elegant, that you may not need as much indirection as you're used to with Postgres etc.
I think you might be right about that. I was assuming it was going to be a nightmare (like, as you said, writing an interface for Postgres) but looking at how simple your initial implementation is makes me think I should just roll my own. Almost all the low-level stuff I was not looking forward to doing is handled by rest-client.
That said, I'd say it really depends on what you plan on doing, since you mention replication across three continents I'm guessing you're comfortable with getting down and nitty-gritty with some real code (and not just through a "framework").
Well, I don't mind getting down and dirty - it's just a matter of how much time it takes : ) But I'm beginning to think it might take so little time it's not especially faster to learn a pre-existing "framework" after all.
Thanks for the responses, guys! And thanks for the code - it's really helpful.
Sho
smime.p7s
Description: S/MIME cryptographic signature
