View Single Post
Old 01-29-2009, 10:18 AM   #81
Dale
Banned
 
Join Date: Jun 2005

Location: Western Missouri
Posts: 960
All three of the logos that I've done, work fine in MA3. But I understand the points you make, Jim.

Originally posted by Jim Sachs:
Dale - I'm not avoiding installer questions. I've never dealt with one, so I simply don't know. Someone at Prolific is doing the installer.  
Yes, I understood that Prolific is doing the installer. But, in my opinion, testing an installer is quite substantially different from incrementally testing features being added (and unexpected problems being subtracted) to a simulation like MA3. Especially where personal perception is a vital part of the testing.

My point is that when I ask someone to test something complex (like an installer), I ask them to find out if it does all of the things that it's supposed to do (in all specified environments), and does not do anything it's specified to not do, and also to tell me if they notice it doing anything they think is "strange".

Some things, also, are only capable of being tested once (per environment) - like flash bulbs, and installers. Running the installer clearly changes the environment.

At the moment, the complex installer is being tested by lots of folks using personal criteria, and the testing is based on "random" environments and initial conditions. The results are valuable, of course - but there's limited value in "works OK here as far as I can see".

And, of course, everybody who has already tested the installer will not be able to (easily) test future iterations of the installer under the same conditions.

I'm quite willing to test the installer as thoroughly as I can, in whatever starting environments I can honestly duplicate on six computers.

But: (1) I'm not anxious to CHANGE those starting environments by running the installer before it's fully baked; (2) I would really like to know what the installer is supposed to do, before I tell them if it does.

As you said, Jim - somebody at Prolific will have to let us know how to test the installer - otherwise, we're just doing random testing that will find a few of the problems.

=============

Footnote: we could also "evaluate" whether the installer DESIGN does the "correct" things. To do that, we would need to know whether MA3 is going to be marketed as an update/upgrade to MA2.6 - or as a new MA that does not "replace" MA2.6. The installer would do different things (to a MA2.6 installation) in those two cases.

Jim, perhaps you know the answer to that one?
Dale is offline   Reply With Quote