Some free VBA-stuff

Well, if Philippe's CadML is the storage mechanism, revision control
becomes easy. There are dozens of time-tested, powerful tools
performing revision control for source code text files already. CVS for
example. Imagine being able to create a new 'branch' in your model
where you explore an alternate way of modeling a dozen features, and
then being able to back out of all those changes, returing to the model
you started with.

Jim S.

 

Is it using one file to store all modelling information (huge and slow
file?) or one file for each rev.?

So...where's the free tool that works with SW then ;-)

I know that many users are already quite confused with configurations
(maybe not here but generally) and if there is going to be something quite
similar but for revisions is it all getting too complex for most of the
people?

 

Markku Lehtola wrote:

Well, for something like CVS, it is a single file. The CVS system
injects text to mark what has been added/removed, etc, between changes.
The end user never sees the injected text, as it is stripped out when
the user 'check out' the file.

CVS overall, is one of the less intelligent systems. IBM has a current
open source project called Stellation which has a finer grained aproach
to source code storage, in which individual program elements (i.e. a
function declaration) can be tracked. As a user, you can say, show me
the function foo() as it was last Tuesday at 11:30 AM, and *poof*, it
appears.

Once again, this is for source code, but my mind gets spinning when I
think about the possibilities for other areas. A solid model, when
expressed as text, isn't all that different from a computer program.
Interdependencies between smaller elements are used to create larger
elements and so on.

I suspect that configurations are overused. I avoid them for the most
part because of the problems they cause when the time comes to make
drawings and BOMs.

Revisions are another matter. There are two types of revisions: public
and private. Public revisions are the ones that go through official
channels to let mfg, sales, etc know that something changed. Private
revisions occur when I think of an alternate way to do something and
want to explore it. Currently, you can use configs for this, but it is
cumbersome. You have to pay close attention to suppression state, and
even closer attention to avoid changing dimensions globally that you
only intended to change in the current configuration. The other option
is to copy the file and then hack away on it, but then you may have
issues if the 'hacked on' version turns out to be better. You have to
track down the places the part may be used and so on.

I want the freedom to fiddle with things without having to worry about
getting back to where I started. If a nice revision control system for
public revisions can come from that as well, I'm all for it.

Jim S.

 

Little bit afraid to see a part that has let's say 10 configurations and
each has several revisions and all that information is in one file..there
should be a tool to strip off some revisions that you propably don't need
anymore (but you never really know what you need and what you don't)

Only "problem" for me seems to be that you can't have the same file name
for part and a drawing, but because..of course...I know that I'm using
configurations it's not a problem at all. Custom Property softwares can
handle configurations...and if they don't you should switch
to one that can. Never had any other problems..

I agree. Maybe some day..it's good to have something to dream about

 

Philippe,

I wholeheartedly agree that XML makes for a great structure to store SW
information, especially since its flexibility is well-matched to the
flexibility of SW models. I've been using XML on kind of an ad-hoc basis in
building configurator type programs, so I was interested to see that you are
proposing a standardized vocabulary.

I was curious about how far you envisioned going with this, and what sort of
benefits you see in standardization. Depending on how much time you want to
invest in this, oasis-open.org has a forum for establishing Technical
Committees to develop standardized XML vocabularies. This could, however,
complicate the issue beyond what you're envisioning. BTW, in searching for
CadML, I came across another CadML proposal that seemed to focus more on
functional aspects of the modeled objects:
http://bingo.crema.unimi.it/ontology/doc/ontology/ont.pdf

For myself, the ad hoc methodology is still a great tool. I'd be interested
in learning how to take it further, though, if standardization opens some
new potential.

Best Regards,
Brenda