I think what you’re saying about that Joel article is “Don’t take lessons from it that Joel doesn’t really intend for you to take.” Which is good advice for any text, software or otherwise.

Yes, there is valuable knowlege to be gained but I find many articles by people who are experienced basically boil down to: Don’t do this dumb thing or, more specifically, don’t get too wrapped-up/excited about X becuase then you’ll do this dumb thing, or I/my team/my coworkers/my boss did this dumb thing, please don’t do this dumb thing. This can be important, but you have to be very careful because one person’s interpretation of what makes a thing “dumb” can be quite different. That means, you have to be very careful about extracting meaning from stuff like this.

It’s very easy to over-generalize or miss the point. It reminds me of a Buddhist saying that goes something like “…thus we should be more like the baby, but if you say to people ‘The baby is the Way’ they will misunderstand you.” So for this article, it’s very easy to jump from “architecture astronauts” to “OMFG STARDARDS IZ 4 DA RETARDS!!!111 MARKETZ R00L!!”. Again, the kernel of truth here is more like “If you overdesign things in a vacuum without understanding the requirements, you will create a solution in search of a problem. Moral of the story: Don’t get over-excited about your own interior interpretation of things or you will do a dumb thing.”

Again, I’m not so down on articles from Joel on Software, but keep that grain of salt handy.

Generally, you have to let folks thrash around for a while to figure out where things are going because you can’t necessarily architect stuff up front. That said, sometimes, it’s OK to take a look at the field and say “Man, these all suck. Let’s do something smarter.” and make that the standard. That doesn’t always work, but sometimes it does. Am I wrong in thinking CSS was like this? That is, everybody took a look at all the crap going on with HTML tables and what-not and said “Man, this all sucks. Let’s figure out the right way to do this.” and lo and behold we get CSS. Gratned, it’s not 100% adopted or whatever, and it needed some follow-on work, but it’s getting wider and wider adoption because folks see it as going more in the right direction than making everything flash or continuing to screw around with tables or what-have-you.

In practice, I wonder how the politics here work out. And would it be better just to let the market work out the confusion on its own? That’s how Beta v. VHS worked out, right?

To really answer the question involves, I’m sure, digging into a lot of particulars. Some companies probably stack standards-making bodies with members that are friendly to them; some use the market to destroy the competition; etc. Shapiro and Varian have a chapter devoted to this.

The problem comes at convergence-time, when there is a lot to be gained by having my implementation become the standard (less code I have to change, more change for my competition). The politics of this process can lead to drawn-out battles that hurt the consumer (HD-DVD v. Blu-Ray, anyone?) and delay further innovation. Any ideas on that problem?