[Nitro] RFC: Basic Points for a Better Nitro/Og
transfire at gmail.com
Fri Sep 21 17:13:44 EDT 2007
On Sep 21, 11:59 am, Aidan Rogers <ai... at yoyo.org> wrote:
> I think there are three things that stop Nitro/Og from greatness:
> 1) Getting started is hard
> 2) Features are more highly valued than stability
> 3) Rails
> Arguably, we're all here because we found things we didn't like with
> #3. At the same time, I doubt there's anyone on this list who has
> not written at least one Rails app because it's easy to do.
> #2 - George loves new features, and we have to be grateful for that -
> otherwise there'd be no Nitro! Many of us just want it to work
> reliably and testably.
> But #1 is what kills Nitro. Last year I spent six months during
> which I was writing an app using Og and little bits of Nitro. I
> contributed a few patches to the Og repo and I got to know the
> codebase intimately. Now I come back and find that I have to learn
> everything again because so many things have changed: how to start an
> app, how to configure Og, how to build classes so that they are Og
> friendly (luckily that wasn't too different), etc.
> Right now, Nitro has only one developer and that's George. There are
> a few others out there whose contributions are patchy (pun
> intended). We cannot expect George to do everything: maintain a
> website, mailing list, gems, repository AND write new features for 4
> codebases AND make sure it's all well tested AND write documentation.
> Personally, I feel that a feature freeze is necessary - George, I
> think you are your own worst enemy in that regard :-) However, given
> that none of us can prevent you from doing this and that it's
> probably what you enjoy doing most, perhaps we should branch the
> Nitro repository?
> If we create a "stable" branch and a "development" branch and get
> stable to 1.0 as quickly as possible - a full suite of specs,
> documentation, website etc. Meanwhile, George and whoever else can
> continue to develop the bleeding edge but nothing gets checked into
> "stable" until it has specs, docs and good examples.
> Thoughts? Volunteers?
Good assessment. A stable and development branch sound like a
reasonable idea. Even so, I think API stability from here out is going
to make or break us, so it might not matter as much as it would have.
Still, couldn't hurt and I could be wrong --George might be planning
to rewrite the whole system again ;)
> Whether you split Og off or not, I think the same applies for Og.
> I'd be happy to volunteer to get Og ready for prime-time (I know very
> little about Nitro).
Wow. Now there's a great offer. George?
More information about the Nitro-general