Open Source Infrastructure (editorial)

Jeff Breidenbach has written an editorial on something he feels would significantly assist the Open Source community- a single place to get the infrastructure necessary to develop open applications. All the tools are available, but they need to be provided in an easy way so that companies, as well as anyone else on the net can get started without to much duplication of effort. Read what he has to say about it by hitting that link right there.

The following is an Editorial by Slashdot Reader Jeff Breidenbach

I read with interest the editorial on the Linux Labor Force. Buried in the discussion was the suggestion of putting together a suite of development applications for working on an Open Source Projects. Such infrastructure would include things like a bug tracking database, a version control system for code, an archived mailing list, and other tools.

Good development infrastructure is no substitute for good design and good people; but there are occasions when it can help. Today, each Open Source project puts together this infrastructure themselves, sometimes with a bit of blood, sweat and tears. I work for a startup commercial software company; we went through the exact same process in-house, often using the exact same tools you'd expect to find in an Open Source project.

This could be easier. I'd like to go to a website and say, "Give me all the infrastructure I need for an open source, distributed development project." And I want to walk away with an infrastructure at least as good as that used by mozilla.org, with no fuss.

My proposal for bringing this about is to think services. The infrastructure tools required are really services, rather than software. We could set up such services in a distributed fashion, across both the network and across organizations.

For example, there could be an organization whose sole purpose in life is to provide a bug tracking databases for assorted projects. The databases would be online and web based. Another organization might specialize in providing mailing lists. Another organization might specialize in providing CVS repositories for projects. Finally, there might be a completely separate service that specializes in tying these infrastructure components together.

Specifically, here are the things I would specifically like to see:

• A service that provides version controlled code repositories
• A service that provides mailing lists
• A service that provides automated documentation from source code
• A service that provides code browsing
• A service that provides a bug tracking database
• A service that provides a feature request database
• A service that provides real time communications
• A service that provides testing frameworks
• A service that bundles the other services for one stop shopping.

Some of the pieces are already there. People who need a project mailing list archived could take advantage of www.mail-archive.com (not a random example - that's a service I set up.) There are probably other pieces out there, and some will have to be built. Together, these specialized services would provide instant infrastructure for any distributed project on the net that wants it.

Don't get me wrong; not every development project need use the same prefabricated tools. Not every project will want or desire autogenerated hypertext code documentation of their source code. But those things should be available, if desired, in an easily accessible way.

I suspect that such tools and services would appeal to many development groups (including Open Source folks, including hobbyists, including professionals) and I imagine that some of these services could be commercial.

I would be happy paying money in return for instant project infrastructure. I can imagine that world class infrastructure could be made available extremely cheaply (or freely) to distributed development groups. I think things would probably work best in a free and open atmosphere; hopefully there will be an opportunity to prove it. Maybe the service providers will contribute to open source development and make money; who knows?

Are people are interested in pursuing such infrastructure? Take your favorite development tools and build them into a service. Make that service the best darn service you can. Be a piece of a larger development infrastructure. Questions? Thoughts? Discussion? If there is enough momentum I will do what I can to help coordinate, if that's what people want. I have set up a mailing list [to subscribe send the word subscribe in the body of an email to [e-mail]] for people who want to participate or at least talk about the matter further. Or perhaps this is the sort of thing that groups can go strike out on their own with.

If only someone could think of a catchy name....