pros and cons of top down design

[Den E]

My 2 cents:


TDD is useless when it comes to tooling. You always start with the customer's product and build the tooling around the product. The only guidelines, or "skeleton", are dictated by the customer. Such as the machine the tooling is mounted onor cell or whatever applies to the finished tooling. Sometimes there are height or width constraints. Sometimes there are weight issues. Ergonomics is nearly always a factor. All that has to be taken into consideration when designing the tool. Tooling is always unique to each project, so top down design is laughed at in our line of work. We've been using Pro/E exclusively for around 10 years, and it works wonderfully for tooling. But you have to ignore the TDD approach. Also, we neverconstrain any of the tooling to the customer's product. That's been tried, with disastrous results. It may take longer to sort things out when the product is revised later on, but in the long run it works out great for us. I know PTC frowns at our approach, but nobody likes PTC anyway.

 

[dgs]

TDD is useless when it comes to tooling.

That's a pretty broad statement that I don't buy. It may be useless in your environment and how you've been applying it. I don't know what parameters are driving that statement.


I did injection mold design with Pro|E for 5 years and ha I know about skeleton driven TDD, I suspect I could have leveraged it's power to make my job easier and my data more accurate. Not to the same effect as I can in product development, but there would be uses for it. Pro|Mold had some TDD techniques built in in the way that ejector pins were created and such, but I suspect I could have done much more. It's been years since I did that kind of work, so who knows.


I agree that referencing the customer's model would be a bad idea, particularly if it wasn't native Pro|E. Anything you're not in control of can change drastically. resulting in a lot of extra work to fix it.


'Tooling;covers a broad range of applications. I'm sure for some, TDD is of limited value while in others it would be quite helpful.

 

[Den E]

TDD is useless when it comes to tooling.

That's a pretty broad statement that I don't buy. It may be useless in your environment and how you've been applying it. I don't know what parameters are driving that statement.


I did injection mold design with Pro|E for 5 years and ha I know about skeleton driven TDD, I suspect I could have leveraged it's power to make my job easier and my data more accurate. Not to the same effect as I can in product development, but there would be uses for it. Pro|Mold had some TDD techniques built in in the way that ejector pins were created and such, but I suspect I could have done much more. It's been years since I did that kind of work, so who knows.


I agree that referencing the customer's model would be a bad idea, particularly if it wasn't native Pro|E. Anything you're not in control of can change drastically. resulting in a lot of extra work to fix it.


'Tooling;covers a broad range of applications. I'm sure for some, TDD is of limited value while in others it would be quite helpful.

Maybe I was making a broad statement there, but in our experience, we have never run across a project where TDD would be of any use. And we have designed, and built, everything from tiny gages on up to complete robot weld cells. Granted, we have never done any mold tooling, I think, but there are plenty of design shops out there that can handle that type of work.

 

[SRINIVASANIYER1]

TDD is useless when it comes to tooling. You always start with the customer's product and build the tooling around the product. The only guidelines, or "skeleton", are dictated by the customer. Such as the machine the tooling is mounted onor cell or whatever applies to the finished tooling.

Den,


What you describe is a mixtureof TDD and BUD approach. As I said before you have to use BOTH to achieve control. CONTROL is IMPORTANT not the procedure. When you drive a car, use of both ACCELERATOR and BRAKE gives you control and prevents disasters.