The Atom syntax mail list has been busy recently. Issues that were discussed at length (long before I got involved) have been reopened, and if previous decisions are overturned, big changes may be in store. I heard that the size of the mail list has ballooned significantly since the push to join either IETF or W3C began in earnest, and it appears that some, though not all, of the stirring of the waters is a result of all the new blood. While many of the comments coming across the list are insightful and based on years of experience, I sometimes wonder whether a small dose of ignorance may have resulted in greater bliss.
The upheaval that can result from large influxes of interested persons exposes one of the weaknesses of keeping the community so open. Don't get me wrong--I'm all for an open community. Without it, I couldn't have gotten involved. I'm just comparing reality to a hypothetical world where the group started out completely open, everyone who was ever going to get involved did so at the beginning, and then the doors were closed until one day they opened up and out popped the specification. Such a world would benefit from not requiring periodic reopening of issues. But of course, such a world is impossible, not only because not everyone who would eventually get interested knew about Atom in the beginning, but also because some people only decided to participate after Atom had enough momentum to influence their financial or emotional bottom lines.
Leaving the purely hypothetical, I'm reminded of a quote once related to me by my brother (a lawyer who manages to juggle an incredible number of things all at once): "Perfect is the worst enemy of good enough". In other words, once you've made something good enough, call it finished and move on to something that's not good enough yet. Most things can never be made perfect, and even fewer need to be.
I won't go so far as to say that the issues being reopened have been decided in a way that's good enough, and that we should refuse to reconsider them. That may be true and it may not. But I do think that we need to take care not to be too open to the reopening of issues, or we'll spend too much time trying to perfect those areas, and never get around to making other areas of the specification good enough. As long as we can come up with a 1.0 specification that isn't so full of bad decisions that it can't be refined later, we would probably do well to aim for good enough, and get there as soon as practically possible.
Tim Bray just sent a message to the mail list expressing the same sentiment. An excerpt reads: "At some point, to get this sucker out the door, we're going to have to drop a sharp pointy blade that will cut off a few ideas that might have been real g...
June 7th, 2004 at 10:02 am
Tim Bray just sent a message to the mail list expressing the same sentiment. An excerpt reads: "At some point, to get this sucker out the door, we're going to have to drop a sharp pointy blade that will cut off a few ideas that might have been real good." It'a always interesting to post something here and then see the same idea come up on the mail list or someone else's blog a day or two later. I suspect that most often it happens because I'm blogging about the same things others are seeing, but I'm always curious whether the people who write these things are reading my blog.