Good IASA Meeting

Last night’s IASA meeting went very well. I always take it as a good sign when community get-togethers end because building security kicks you out 45 minutes after you were supposed to be done. About 25 people showed up and were treated to food, free materials and factories. Steven did a good job presenting Software Factories, and he was even nice enough to let me chime in on a regular basis with my $0.02. BTW, the next meeting is on Oct 27th (last Wed of the month) and I’m going to miss it because of OOPSLA. Watch the IASA website for more details.

One of the key questions asked was if MS was going to “break tradition” with factories and embrace open systems. I guess we still have a long way to go to change that perception, even though we standardized CLI & C# and continue to be a major driver of the ongoing web services standardization effort. Regardless of perception, I agreed – I think it’s important to enable Software Factories to be embraced by the industry at large. The Architecture Strategy team has several members who are heavily involved in standards bodies, and they had similar feedback. Of course, we have nothing to embrace beyond the book so far, but we have hinted at tooling to come. As promised, I sent last nights feedback to Keith, Jack and the other folks involved and I hope we’ll see that incorporated into their long term plans.

I asked the hard-core folks who stayed late last night what they thought was the best approach for MS to take for getting Software Factories embraced by the industry. Obviously, any tooling around Software Factories would be built on a set of specifications. We could standardize these specifications as we have for CLI, C# and web services. Alternatively, we could simply publish the specifications with an open and royalty-free license (similar to the Office 2003 XML Reference Schemas). Given recent comments by Grady Booch and Simon Johnston (both from IBM) disparaging domain-specific languages and software factories, I’m not sure how well a standardization process around DSLs would proceed. On the other hand, I’m not sure how much industry buy-in there would be for specifications that weren’t collaboratively developed. Leave your opinion, and I’ll make sure it gets to Jack, Keith, et. al. just like the feedback from last night did.


I'd like to see an open and royalty-free license for the entire factory markup. My recent experiences with the application of a closed system were very uncomfortable. I'd love to see industry adoption of the idea, but I'd hate to see any projects lose steam because all of the people who hold the vision for it are too busy convincing the non-believers. If it's a truly powerful tool (and I believe it will be), I think the benefits will make themselves clear and they'll eventually switch to what works. Meanwhile, the owners and users of the tool can construct far better solutions than they ever could before.