I'm a new software architect/lead, coming up with software design for a team of software developers. I'm coming up with the requirement spec, interface header files, and visio software design docs, and build plan, etc.
My question is: what do the rest of the team do during this period? I'm certainly engaging them in the design, but we dont need the whole team actively working on what I'm doing all the time.
Are there any good books for new software architect?
As other have said, you typically want a ramp-up period during the first part of the project, and through the first iteration. You're planning on building this iteratively, aren't you? Start with a core team (nor more than 3-4 people, since you're going to need to communicate heavily with each other) to help you explore the requirements, get a basic data model in place, identify and setup any frameworks, identify and setup build and test tools. Some coding activities typically take place in the design phase: for UI mockups, run-ahead prototypes of technically sensitive areas (whatever risks you have should be mitigated by explirative coding: be they new technologies, undocumented interfaces to integrated systems, or unstable requirements).
But coders in the design phase should help with the design, in order to get their buy-in, and to help train up the rest of the team during the first iterations. Your role during this is to ensure that the major nonfunctional requirements (e.g. are known, prioritized, are met by the design, and can be tested). You should also collaborate with the project lead or whoever else is responsible for staffing and financing in order to sketch out the iterations and the staffing levels needed. Ensure the solution can be built iteratively, and aim at implementing only a basic structure during the first iteration, both to build confidence, and to eliminate risks. (Sometimes, you can push major risks to the second iteration, and focus the first towards confidence and team building.)
And of course, be sure you are not designing every detail. You should be able to use every design artifact in the next iteration (and elaborate them later as needed). Since design decisions are expensive to change, try to postpone them. However, some influence the entire solution (for instance, the data model, or your approach to security) and absolutely must be at least outlined up front. This isn't waterfall. This is just not closing your eyes and hoping a viable architecture will emerge by magic.
But design proceeds throughout the iterations. It's just that you do less of it as you go along, and with lesser impact on the solution (unless you're unlucky... and then things get expensive).