I'm clearing out some of the links on my to-read list from Martin Fowler's weblog:
Some thoughts on UML 2: part of the problem with UML is that it's not unified. People use it as a sketching language, as a blueprinting language, and as a programming language. You need to sync up your expectations before you can effectively use or communicate with UML.
Software architect's aren't. Why? Because architects are something like a meld of user interface designers and project managers. Architects are the primary interface to the customer, and they sketch out an idea for the building. Software architects, in my experience, are not like the architects that Fowler describes, but more like "lead designers" or "first citizens" of a team of programmers. The allusion to architecture and architects is simply pretentious.
More thoughts on software productivity: the reason we can't measure productivity is because we can't measure output.
Patterns aren't anything new, and they shouldn't be. They should be concise descriptions of things we are already doing.
And finally, a discussion of some criticisms of refactoring. The point of refactoring isn't to expend effort to rewrite something that's complete and working, but to improve code you are actively developing. No one is advocating the Refactoring Squad visit your project and leave a body of unintelligable refactored code in their wake.
