tim Rowledge  (and others) wrote:

>
> On 8-Dec-05, at 1:48 PM, Yanni Chiu wrote:
>
> Lots of good stuff.

Thank you for these replies, they have been very helpful indeed.  I 
think the insight I'm gaining is that it's fairly common to hand craft 
or furnish a "prototype" image for one's development.  That custom 
development image is then "instanced" periodically as the need arises -- 
when starting a new projects or other reasons.  Also, each of those 
images might be check pointed periodically.  This is essentially the 
life cycle of the development environment.

I now need to spend some time looking at Monticello since I gather from 
the reading that I've done that it provides some structure (via Class 
and Method categories) about just what an application "is".

I also gather that a measure of discipline is required to avoid creating 
code that only works because you happened to instance an object in a 
workspace.  This doesn't appear to be required -- you could conceivably 
hand assemble a running application in an imagine but later deployment 
might be tricky.  Correct?  I imagine Unit tests can be used to help 
verify this.

As an aside, how to people move their images around?   I have several 
different machines that I do development at.  By checking my work in 
when I'm done at a given machine, a simple "svn update" at a different 
machine and I'm ready to continue work.    It occurs to me that I could 
check in my images "changes" file to a central server (svn or other) and 
simply fetch that changes file when I move to a different seat.  Does 
that make sense?  How do other folks do this?

Again, thank you for the replies any further thoughts will be appreciated...

Keith