So, as typical I may be a little ahead of myself.... But...
I've submitted an abstract for something I cannot find a "Best Practice" for on anywhere in the CWG.
What's the "Best" method for assigning classes and characteristics to a configurable material?
Yeah, I know we state "Class Hierarchy is the way to go" but I cannot find anything that shows benefits\drawbacks in a given situation versus something "easier" or "less abstract"... So I thought I'd try to create it, very much like the "Restricting Characteristic Domains" done back in 2005.
I don't yet know for a fact that it will be accepted, but building the presentation (and maybe even some stuff in the sandbox) is going to take a while... So I'm already starting on it with an assumption I'll be giving it... And here's a little teaser...
How many different ways are there to group and assign classes and characteristics to a group of materials (so far, I'm up to 7 with an 8th in the works (and I know there are more)?
Should you stuff every characteristic you need into 1 class?
Or a compelx class hierarchy with multiple inheritance?
Or something in between...
Global or product line or maybe even model specific characteristics?
Are there benefits or drawbacks to one method that don't apply to another?
I don't really believe we will be able to say "Method X is always the thing to do", but I would like people to at least see how to think about it and WHY they may want to do one thing versus another.
I want to make this easy to understand, yet difficult enough to reflect reality at multiple companies.
I'm planning on setting up some paper demonstrations of 2 "basic" product lines, with 50 materials to a product line.
20 characteristics consisting of 10 "Specification relevant" characteristics, 8 helper characteirstics, and 5 reference characteristics.
5 of the 10 specification characteristics are shared across product lines.
Then drawing up the (at least) 7 methods and drawing comparisons against each of them.
We won't get into the value sets yet, assume they are the same so we have some objective method to "count the number of rules"...
So tell me, is this something you'd like to see??
Cause if it isn't, I'll stop working on it and withdraw the abstract..
But if it is something you'd like to see and participate in, then chime in and let me know... Why am I asking you this??
Selfishly speaking, this one is actually much more work than "Restricting Characteristic Domains" was (mostly because I'm doing it myself this time and there isn't a "white paper" to draw from).. So I need to know if you think this is valuable.
On top of that, I'd like to see more of these types of presentations so that people have some "Baselines" when they get into this thing we call "Product Modeling"...
Assuming you think this valuable, what types of things would you like me to try and work into this??
Lemme know your thoughts!
PS... I WILL give away beer if you participate AND interact with me during the presentation!!