Re: [GD-General] Problem Breakdown
Brought to you by:
vexxed72
From: Brett B. <res...@ga...> - 2004-07-15 03:10:17
|
> One technique we use is what's called a sticky note meeting. We have used this same technique and found it really handy too. The only problem with it I have found is that it presumes everyone has the same level of understanding of the problem (as you mentioned). I'm leading a foreign team and although the team's English is good, the combination of English as a second language and different terminology for certain programming algos, constructs, etc. sometimes conspires against us. > Sticky note meetings are not meant to replace the low-level design. You > say "Rational Rose stuff" and UML are too high-level, but neither of > these assumes a levelness. UML is perfectly suitable to designing "Hello > World", a linked list, or a resource manager; just like a nuke is good > at killing things of all sizes, these tools work for all problem sizes. I guess what I meant to say is that when considering a problem with a clean sheet of paper these tools make some assumptions you know you want to create a "hello world" application. Maybe what I meant was actually "inappropriate"? > 1) high level task analysis done via group effort sticky note meeting. A > basic feature document must be pre-existing with a prioritized feature > set on a scale of 1-4 (1 won't ship without, 2 very important, 3 would > be nice, 4 can live without). > 2) mid and low level task analysis done by a single engineer (research, > prototypes, pre-existing solutions). Result is the design, a design > worthy of being estimated and implemented > 3) estimation > 4) group review and sanity check > 5) schedule created including milestones > Can this process simultaneously handle things like AI, geometry handling, shader management, batching of stuff, cache managment, networking, time, etc., etc. always seem to creep into any discussion we have about anything in the game. How do people breakdown these complex problems and assess the tradeoffs and impact in implementation? I'm finding it tough to do in a team environment. |