Wednesday, February 17, 2010

Open source: dangerous to computing education?

This is a response to a recent blog post by Mark Guzdial, chairman of the ACM education board.

Open Source is an integral part of a well rounded computing education.

While I will praise efforts to raise awareness of demographic barriers in open source development, I would challenge Mark to do a better job lining up his arguments and checking his references. Especially when many of his links have little or nothing to do with the statements attributed to them.

His link "more closed and less diverse than commercial software" links to a National Center for Women in Technology recruitment flyer. -Which sounds like a wonderful organization, but having read it now 3 times through says nothing about open source being more closed or less diverse.

The next link "overwhelmingly White or Asian and male" links to a OSCON presentation which says a lot about the male/female ratio... but nothing about white/asian demographics. Furthermore, it is a presentation which shows that the open source community is making efforts to draw awareness to the issue and improve the situation.

Mark attribute a quote to Linus Torvalds: "Talk is cheap. Show me the code." -But that quote is notably absent on the page he links to to reference it.

I'll grant, that that quote certainly sounds like something Linus would say. I have heard it many times in many open source projects. Taken out of context, it is used to imply that talk and conversation are undervalued. However, that saying is most frequently asserted during a conversation about:

1) A potential bug
2) Someone asserting that one implementation is preferable or superior to another

In the first case it is notoriously difficult ( to isolate a problem without access to the code.

In the second, the noise to signal ratio of conversations quickly become quite poor unless the conversation is rooted in actual code.

Legitimate Peripheral Participation exists in open source communities. Non-developers provide some of the best documentation. For the usability issues mentioned, non-developers are often better at writing documentation than the actual developers. They often help triage bug and issue tracking systems. And they can gradually move into development by bug testing and submitting tests which reproduce those bugs. And even when they don't gradually move into a developer role, they can become useful and respected members of a project by helping other users on mailing lists and IRC channels.

In Mark's closing paragraph, he disparages the open source community by painting it broadly with the religious zealot/fanatic brush stroke. It is true that open source has its zealots and fanboys. Apple, Microsoft, Google, etc. all have some rather intense supporters whose support relies more heavily on a go team mentality than on solid engineering. However, successful open source projects like the aforementioned companies also tend to have a core of solid engineers and skilled developers.

In closing, I applaud Mark for raising awareness of gender disparity in open source communities. And I would criticize him for implying that Open Source is "dangerous". I would remind everyone that correlation does not imply causation. There are many causes of a gender gap. My current employer has roughly 33% women developers in my department. When I started it was 0%. Things are getting better. That isn't an excuse to be complacent or dismissive of problems where they exist. We can and should strive to do better.

No comments:

Post a Comment