the tone of the discussion

Posted Dec 20, 2007 17:22 UTC (Thu) by jd (guest, #26381) [Link]

I believe both sides have confused specifics with generalities, and generalities with personalities, but that the OpenBSD developers have done so more than RMS. This is not a criticism of any individual - every single human being alive has fallen into that trap more than once and will do so many more times in their lifetime. Rather, it is a bug report of a potential exploit that exists within the mind. And, like any other bug report of an exploit, it merits analysis and patching. That it is a wetware bug, not a software bug, should make no difference. It is a bug and developers fix bugs. That should be the end of story.

In practice, it's not that easy. The mindset that makes for a good developer appears to be exactly the opposite of that which makes for a social, tolerent and non-hostile individual. It is unclear if fixing the mind bug will break the developer capability. It's very possible. This leads to a difficult conundrum - would fixing the mental bug create a regression far worse than the bug itself?

If the answer is yes, then nothing can be done on the side of the verbally abusive developer without making things worse. Rather, the solution lies in a beter protocol handler on the part of those communicating with such developers. It's an ugly hack, but nothing better has appeared on the horizon. I would argue that Theo is probably in this category, that his aggressive, anti-social, monomaniac attitude is actually critical to his phenomenal talent as a programmer, that attempting to change him would be a disaster in the making. It's not behaviour I like or would normally encourage, but in his case it is an inevitable consequence of his particular type of genius and so I actually encourage people to make an effort to assist him in being as true to himself as possible.

In some cases, the answer will be no, but the developer won't apply the necessary fixes. In other words, fixing their behaviour won't harm anything, but they're not prepared to make those changes. A developer who won't fix a known bug in the system they can change the easiest (themselves) probably shouldn't be trusted to fix anything else and should therefore not be given so much power on mailing lists.

Finally, there will be developers who do patch their own attitudes. They may still not be who we'd like, but credit should be given for the development done and the bug fixed, no different from any other bug fix of high importance.