[Sigia-l] We could just use whiteboards instead.

[Adrian Howard]

On Monday, August 18, 2003, at 03:35  am, Matthew Rehkopf wrote:
[snip]
> How does one manage revisions and reasons for
> revisions when all of the information on the system is
> "loose" on the wall? It seems that with so many people
> able to make changes, that decisions made prior would
> eventually get overwritten because the "changer" was
> not aware of the decision, or no one remembers why
> something was placed in a particular spot. I guess
> this is a question for all of those that work with
> stickies until the product is nearly fully designed...
[snip]

Things that I've found useful in no particular order:

-	Have an explicit process for recording changes. Access database, 
post-it notes or, my personal preference, a stack of index cards.

-	Wait to see if there is a problem first. With a good team working 
together the problem can be rare enough that its a non-issue.

-	Have a policy of telling everybody why changes are made (easy if 
you're all in the same room). Then its more likely that a potential 
regression problem will be in somebody's head and you won't regress.

-	Have a policy of always making changes with at least one other person 
present. It's more likely two people will spot a regression problem 
than one.

-	Take a photograph of the board twice a day, or once a day, or 
whenever seems appropriate. You can then easily go back to earlier 
versions.

-	and, of course, move to a digital model with a change control system.

Adrian