Friday, June 05, 2009

More of the Ruby Mendicant: Gregory Brown's Back with Another Great Project

For the second interview in this morning's double-header, I've got a quick one with Gregory Brown (@seacreature) about his recently announced second run as the Ruby Mendicant. This is a big opportunity for the ruby community. I hope you'll pitch in a few bucks.


RubySpec has been a huge success in the Ruby Community, but it's just one part of a vision that Charlie Nutter put forward at the 2006 Ruby Implementors Summit. Another leg was an live look at how any given Ruby implementation performs against the spec . You're talking about making this happen as part of a second instance of Ruby Mendicant. Can you tell us a little more about your plan?

Greg I will essentially be trying to build and deploy a web based reporting system that aggregates RubySpec definitions and results across every Ruby implementation and language version we can get our hands on. This will help people know what the state of the specs actually are, and how well various implementations and Ruby versions run against them. I made a rough mockup of what one of these reports might look like. We already have some changes to MSpec and a skeletal database model in place that could facilitate this sort of report. The full scope of the project will emerge once I have a chance to talk with implementers and Ruby community members about what their needs really are, and I expect to have those details collected together by June 15th. But basically, I hope to support in one way or another most or all goals that Brian Ford mentioned in his idealized case of the needs RubySpec would serve.

Whatever the final roadmap is, I hope to implement something functionally useful to the community over the course of just 8 weeks. Based on the ideas I have in my head right now, if I just hacked on this in my spare time, it might be something I could put together over the course of 6 months or a year. But with me taking hard math classes in the fall, I'd rather fill my summer lows with some open source Ruby awesomeness. That's where a sustained, community funded effort, will come in handy. The cool thing is that it'll be easier to do it this time around than it was with Prawn....

What did you learn from being the Ruby Mendicant last time that will change your approach this time around?

Greg Okay, so there are two things that'll change pretty radically:
Funding and Organization.

When it comes to organization, I plan to have a fairly clear road map set up before I begin on this project. Ruby Mendicant collected 100% of funds before a *project* was even selected, which made it a very different proposal. The upshot here is that since we're mainly just building a web application to interface with an existing technology, there are less unknowns in the core needs for this project. While there are a ton of directions we could take it that may lead into dark spots, I think I can define with some accuracy a useful set of features from the get go. I don't want to get all ponies and rainbows about what you can expect from this project, things will change and I have no way of promising to deliver a certain set of features within a rigid time-line. But generally speaking, there are less variables here so long as we keep the project scoped tightly enough. This new way of organizing things might appeal more to those who were afraid of the ambiguity of purpose that the Ruby Mendicant project had.

Then there is funding, a topic I absolutely hate, but over the course of the years, I have realized is pretty important to successful free software development on all but the most sexy projects.

Since the time I finished up with Ruby Mendicant, I've been doing most of my work through a devshop called Madriska Media Group run by my friend Brad Ediger. Brad has been pretty cool about letting me spend some company time on open source here and there, which has been great. When I mentioned my ideas for what I wanted to do with RubySpec, it went over well. Madriska will be dedicating some time, money, and resources to helping me get this new project up and running. This will help out a lot, but due to the aggressive timeline and scale of the project, won't by any means cover the full costs.

So while I'm most of the way covered here in terms of funding and available resources, I'd like to help make sure I don't run into financial troubles again, like I did with Ruby Mendicant. That means I'll need to open up a donation drive even though I have the effort partially funded.

How can people contribute?

Greg Code help will definitely be desired a little later on, but right now, the community at large can most easily help out by throwing me some money. I am asking for $500 a week, for which I will guarantee at least 10 hours of dedicated work on this project. This will of course be in addition to any work that Madriska pays for, and not just a weekly minimum.

I plan to do the funding in two drives, one that's already open, and another that will open July 15th. Each of these cover 4 weeks of funding (a total of $2000). This time around, I won't need to raise all the money up front before I start, so those who are skeptical can wait to pitch in if they'd like. While work will begin officially on June 15th, I'll be preparing a roadmap and doing some preliminary planning over the coming weeks. Unlike Ruby Mendicant, you'll have a pretty clear sense of at least what my goals are before this project kicks into full gear.

Of course, those who trust me to do good work, and enjoy what I did with Prawn can go ahead and donate now if they'd like. I'm not going to complain :)

Have any organizations stepped up to help?

Greg When it comes to funding, MountainWest Ruby will match dollar for dollar on the first $500, so thanks for that Mike :) Obviously, if other organizations want to sponsor this effort monetarily, I'd be happy to take their money, so do get in touch if you're interested. What I do need to let folks know is that this is not a bounty, or a hired job, so consider your donation a gift if you make one at all. I can do my best to make sure you won't regret it, but be sure to ask questions if you don't know what to expect.

When it comes to slinging code, I've been having good conversations with the JRuby and Rubinius guys, and will reach out to the other implementations soon as well. I mostly just need their guidance on certain aspects of things because I'm fairly new to the happenings of Ruby implementations. While there is no formal agreement, I think that these guys will help out when they can, and that alone will be huge.

I think the best bet is to spend a few weeks getting together a function alpha for the community to poke at, test, and provide feedback for. After that happens, regular contributions via github would be most welcome. BTW, the working title of this project is Unity, which hopefully describes its intentions pretty well :)

Click here to Tweet this

No comments: