OSCON keynote: Standing out in the crowd [LWN.net]

OSCON keynote: Standing out in the crowd

Posted Jul 28, 2009 12:35 UTC (Tue) by dskoll (subscriber, #1630)
Parent article: OSCON keynote: Standing out in the crowd

I do notice that the FOSS community is overwhelmingly male, more so even than commercial development demographics. And I agree with others that the at-times aggressive nature of FOSS discussions puts off a lot of men and probably even more women. This is really too bad.

We should find ways to include more women in software development, if only because we're ignoring a huge development resource.

to post comments

OSCON keynote: Standing out in the crowd

Posted Jul 28, 2009 13:40 UTC (Tue) by tialaramex (subscriber, #21167) [Link] (10 responses)

The purely resource-based argument might not go how you'd expect. Spolsky is right - one or two superstar programmers can get you a lot more work done than dozens of merely competent ones. And hundreds of clueless noobs? You could easily get less done with them than without them. So for it to make sense to alienate some existing contributors (e.g. by telling them their behaviour is unacceptable) under the logic of maximising resources, you need to be sure that these contributors are less useful than the people you'll attract by getting rid of them. Get rid of one superstar and attract five competent people and you may find your project is struggling even more than before...

But not all projects are _about_ getting stuff done. The projects used as illustration for this keynote seem to have a "getting stuff done" as only one of several objectives. If your priorities look like:

1. Enjoy ourselves
2. Learn stuff
3. Build the world's fastest regular expression engine

then you will need a different approach to a project whose priorities look like:

1. Build the world's fastest regular expression engine
...
17. Enjoy ourselves.

Anyway, all I wanted to say was that you should beware of arguments that its efficient to do something when what you probably _really_ care about is that it's Just or Fair, or Good.

I do wonder if the fact that we're mostly working via the Internet in the medium of written text (which is somewhat opaque to questions of gender) perversely makes this problem worse. I'd always rather live and work with a mix of people. If I was conscious of a particular Free Software project having a better mix, I might be more interested in working on it. But as it is I don't notice at all. I couldn't give you even a rough estimate of how many women contribute to e.g. Gnumeric. Maybe (this keynote seems to say it's depressingly likely) there are none at all.

OSCON keynote: Standing out in the crowd

Posted Jul 28, 2009 18:51 UTC (Tue) by dskoll (subscriber, #1630) [Link] (4 responses)

The purely resource-based argument might not go how you'd expect. Spolsky is right - one or two superstar programmers can get you a lot more work done than dozens of merely competent ones.

You are absolutely right. But I can only assume that somewhere in the large numbers of people (men and women) who decide not to participate in FOSS are some superstars. The more people you attract, the more likely you'll stumble across a superstar.

OSCON keynote: Standing out in the crowd

Posted Jul 29, 2009 2:24 UTC (Wed) by Baylink (guest, #755) [Link] (3 responses)

I would assert that *if they don't dive in and start doing, and -- if necessary ignore the idiots until they cannot be ignored*, *then they are by definition not superstars*.

Problem solved.

FOSS is a meritocracy, even if not everyone agrees on which merit is important.

OSCON keynote: Standing out in the crowd

Posted Jul 30, 2009 1:40 UTC (Thu) by jlokier (guest, #52227) [Link] (2 responses)

If a superstar is defined as a person who gets a great deal more done in a particular domain than most of their colleages, then by only considering people who get a great deal more done in environments which require diving in and ignoring a high level of idiotic confrontation, you are dismissing other people who, by their nature and abilities, would get a great deal more done than most of their colleages in a _different_ environment.

By that definition, the latter class of people are superstars in less confrontational environments, and when we fail to make those environments, we are losing access to their potential.

Do you mean that a true superstar would transcend their environment and get on with what needs doing in _any_ environment? Certainly there's a type of person who does that.

But I suspect that many of the programmers we call superstars because they are outstandingly productive in competitive, harsh environments would fare quite badly in less confrontational, more polite environments - at least, until they'd had time to adjust their personality to suit.

OSCON keynote: Standing out in the crowd

Posted Jul 30, 2009 1:51 UTC (Thu) by Baylink (guest, #755) [Link]

+1, Insightful

OSCON keynote: Standing out in the crowd

Posted Jul 30, 2009 13:50 UTC (Thu) by paulj (subscriber, #341) [Link]

I don't think super-star-coderness aligns with thriving-on-confrontation (which implies they have a somewhat confrontational personality themselves).

The super-star coder I've dealt with was, if anything, the reverse. They not only could get through tricky bugs quicker than others and get beautiful code written faster than others, but they also spent a lot of time on mentoring others. In short, they enhanced the productivity of the other coders around them.

With a grain of salt

Posted Jul 28, 2009 19:07 UTC (Tue) by man_ls (guest, #15091) [Link] (4 responses)

What you say has a kernel of truth, but I think you are stretching it too thin. It is true that a superstar developer can do more work than ten mediocre workers, but not "dozens". And that is only developing; creating documentation, helping others on mailing lists etc. does not scale the same way. Also, dedication varies wildly from the paid developer to the person that can dedicate a few odd nights; if your star developer is in the last position he or she will probably produce less than a dedicated but less gifted cooperator.

Finally, notice the "mediocre" bit in "ten mediocre workers". At work you have surely experienced what these "mediocre" people look like: social butterflies, people uninterested in technology but looking for high salaries, climbers up the corporate ladder... These are the people that lower the average. In most open source projects they are not a problem, so differences in skill and productivity are rather in the same ballpark. A really good guy might do the job of two or three others, but not much more. In this light, your sentence:

Get rid of one superstar and attract five competent people and you may find your project is struggling even more than before...

does not really hold water.

With a grain of salt

Posted Jul 28, 2009 23:21 UTC (Tue) by paulj (subscriber, #341) [Link] (1 responses)

It is true that a superstar developer can do more work than ten mediocre workers, but not "dozens".

Sure it's true. One developer might be able to achieve things that no amount of mediocre developers could.

With a grain of salt

Posted Jul 29, 2009 19:00 UTC (Wed) by man_ls (guest, #15091) [Link]

I stand corrected; for several categories of work, one developer can do things that hosts of others cannot. (For ordinary, every day grunt work not so much.)

With a grain of salt

Posted Jul 29, 2009 1:05 UTC (Wed) by dskoll (subscriber, #1630) [Link] (1 responses)

What you say has a kernel of truth, but I think you are stretching it too thin. It is true that a superstar developer can do more work than ten mediocre workers, but not "dozens".

I think the original poster was correct and possibly even understated things. One excellent developer can often make the difference between success and failure. One superstar can sometimes produce something an infinite number of average programmers could never produce. This has been my experience as a professional software developer for the last 20 years or so.

I want to include more people in FOSS development so we're more likely to discover superstars.

With a grain of salt

Posted Jul 29, 2009 6:40 UTC (Wed) by man_ls (guest, #15091) [Link]

That is true, we are assuming that the superstars are men (because they have traditionally been), but that does not need to be the case. It seems that Asperger's syndrome is more common in men, so the archetypal superstar developer that manages to annoy everyone should be more often a man than a woman. On the plus side, you might find a superstar developer that does not annoy people.