On Mon, Mar 9, 2009 at 2:39 AM, Jason Smith <[email protected]> wrote: > Noah Slater wrote: >> >> On Sun, Mar 08, 2009 at 08:42:10PM -0800, Chris Anderson wrote: >>> >>> The overall picture here is not getting it right but rather having >>> running code. >> >> I absolutely disagree, but that's just how I roll.
My point is that I'd have no idea what should go in such a draft until I'd actually squashed bugs in real implementations. I've got nothing against standardization but to me the implementation doesn't seem like a huge challenge. Build it first, fix it for a while, battle test it, then think about writing it up in non-code. But we're just talking about different rolling patterns... >> >> It seems the goal here was to collaboratively develop a standard for >> interop >> between developers in the community. Opening that up to the CouchDB >> community is >> great, but it's not enough. If this is something that I'm going to care >> about, I >> need to see it opened up to the entire engineering community, and focusing >> our >> efforts around an RFC is exactly the way to do that. > > Hell, why stop there? Send it to the United Nations. Or the United > Federation of Planets. > > I agree with Noah in principle, but the low chance of receiving an IETF RFC > blessing discourages me from investing my own time into pursuing that. > > Earlier, someone called this middle-ware and I like that idea. It seems > improper to be in CouchDB itself. Perhaps it could be an external process, > or as another HTTP service (either an intelligent CouchDB proxy or just a > separate tool that depends on couch). Since developers of all languages > would likely want this, it doesn't make sense to do it as e.g. a Python > library. This could be implemented in JavaScript as a replicateable reusable CouchApp. I'd venture to guess that I could get it to handle the easy cases in a short afternoon. No timeframe for getting started, I got other fish to fry. Cheers, Chris -- Chris Anderson http://jchris.mfdz.com
