Multiple users on a single model/assembly

If you have an office with say 2 to 3 operators, and in ProE, would they be
able to work on the same assembly / model file and would it only be part
files where they can not work simultaneously, ie open only in read only
mode.
I have heard that Solid Edge has a capability where more than one user can
work on a model, they call it concurrent design, is this the same for ProE?.

 

I think that this working simultaneously by several people on a model in SE is a
myth. Maybe you could do a little research, Bo, and confirm this alleged
functionality. I tried, on the SE website, checking this brochure, for example
(http://www.solidedge.com/overview/files/se.overview.brochure.34.pdf), but could
find nothing about concurrent design defined as several people working on the same
model/assembly simultaneously. The site does, however, offer lots of resources for
you to become acquainted with SE's actual capabilities, or at least their own
exaggerated claims, but claims which are not so exaggerated as to include this
simultaneous feature creation/editing by several people. Functionality which, BTW,
is not even offered on the high end big brother, UG. Let me take this a step
further and suggest that the idea is so radical, startling and amazing that were
it to become a reality, we'd be reading about it on the front page of the WSJ and,
quickly, everyone in the world would know about it. IOW, the first place we'll
hear about it won't be as a rumored functionality of a midrange modeller (that's
them) alleged on an obscure newsgroup (that's us).

There are collaboration claims made for "Insight" PDM. Check out the advertizing
in this brochure:
http://www.ugs.com/products/velocity/docs/br_se_insight.pdf
Mostly this seems like the PTC product, Division Product View, which lets diverse
groups of people view and markup models and drawings or this, combined with data
management, releasing, etc. PTC, on the other hand, does a little more advanced
form of collaboration with Pro/COLLABORATE, design teleconferencing software which
lets groups in different areas view and edit the same model which everyone sees on
their screen, text message on screen and pass control back and forth between
people, all in real time. Pro/e is capable of several forms of design
collaboration which you can check out here:
http://www.ptc.com/appserver/mkt/products/home.jsp?k=3449
This will all eventually be absorbed into Windchill ProjectLink and PDMLink or, at
least that's been the claim for a few years now.

 

Not at present, and I don't see, within the sequentially created
features which constitute a model, you could have two people working
simultaneously without them 'treading on each others toes'

If you want to have a bunch of people work on an assembly, that's
different. If possible, break the assembly down into subassemblies, and
assign ownership of each subassembly to one designer, and make someone
responsible for making the whole thing work, and you should be able to
make this fly. Most pdm systems have some tools intended to help with
this sort of collaboration. Make sure the various people working on the
assembly can talk to each other without getting out of their seats, or
they won't.

 

Hello to all

There is (limited) possibility for two or three operator to work on single
part simultaneously using Pro/E. I've used this practice in the past (with
Pro/E Ver18) and I'm still using it now (WF I). I'm (we're) using the fact,
that Pro/e is not overwriting the part file, it just saves it versions. The
procedure we're using is as follows:

1. Part (or/and) assembly is stored on central server, so each operator is
saving the files in same directory (this is not absolutely necessary, but it
simlifies the work)
2. With colleagues we decide, on which "sector" of part will someone
conduct his/hers work (zone functionality just might help to cower up not
needed "sectors" of part). It is wery useful to decide which features
(Datums are the best for this) should each of collegaues use for primary
refrencing. It's also wery useful that each of colleagues puts it's features
in one group (if possible).
3. Each of us is doing it's stuff on part.
4. After finishing the work (all colleagues have stored their versions of
part) I clear the memory an open last version of part again.
5. I'm copying the features from parts versions, saved by my collegaues
(using feature operations>copy feature>from different version) into opened
part (groups of features which were grouped up from colleagues could be of
great help during this step). If IRC, this was much easier in Pro/E Ver18,
where some kind of "merge different versions" command was inplemented. Now
this functionality has moved to Pro/Intralink :-(
6. After all features from diffrent versions are copied into my part, this
one is saved as last (and actual) version.

It's obvious, what kind of limitations this kind of work has (Operator
should seat near one to another, they have to thorougly discuss their wokr
in advance, merging different versions of parts can be cumbersome task....),
but It has greatly helped me during time bottlenecks and still does.....

And may Pro/e respond to Your commands...

Kind regards from Solvenija

Joze BARBARIC

 

We´re using Pro/E as a multiuser modeler exactly this way for some time.
With a lot of discipline (about references) this works like a charm!
But what if someone lacks it... nights of painful patchwork.

How often did I wish the program could do the patchwork all by itself?
Stupid repetition of work already done: that´s what computers should do.

And it would be nice to have a selectable visualization on screen
of ones colleagues work currently in progress - to avoid collision
and to increase synergy.

Btw., I know of an editor (nedit) that frequently alerts me to reload
when "another program has modified the file"... so it must be possible
to have multiple users work on the same file (in different sections?).

Pro/E´s opulent feature dependency tracking should be capable of that.
At least avoid the deletion or redefinition of features required by some
others currently under work of another user.

Didn´t I-DEAS have some check-in/check-out multiuser functionality?
I never quite got how that one was supposed to work, though.

Walther

P.S. : a quick and dirty solution for multiple users on the same part
(when deadlines are near) is to merge parts of parts into bigger ones.
If you have to export a file for tooling (IGES, STEP) then you´re done.
Otherwise it is a bad habit because of dead feature chunks in the model.

 

I started by responding to the exaggerated claims of a victim of SolidEdge
propaganda. That's obviously not the whole story. Thanks to Joze BARBARIC and
Walter Mathieu for relating some ingenious ways of addressing the problem of
several people working on the same model. In so doing, they highlighted the
difficulties, possibilites and limitations of this approach to modelling. A couple
conditions are clear, though: 1) constant communication between participants to
make up for the fact that the software isn't/doesn't; 2) using neutral references
from original model or creating references offset from original; 3) additive
approach to modelling (as opposed to subtractive [big block, make lots of cuts])
produces far fewer lost references; 4) the big issue for all of these kludgy
solutions is reconciling/synchronizing all the individual models: normally, in
typical OS synchronisity, the latter overwrites the former, NOTHING HAPPENS
SIMULTANEOUSLY. So, here is the beauty of GROOVE.... it fools the OS into thinking
that all those people participating are the SAME person. The limitation, it seems,
is not Pro/e but the OS (Windows, Linux, Unix)-- they share the same aversion to
multiple users accessing the same data set, unless it is a database (Oracle has no
similar problem with mulltiple, simultaneous user inputs). But GROOVE has the same
serial limitation... no simultaneous user input. Or is it just how quick serial
is, IOW, two things happening in quick succession APPEAR to be simultaneous. So,
is the difference between GROOVE serial and apparent simultaneous just computer
speed and some programming tricks!?!

If it were a matter of a shared network memory space, reserved for simultaneous
working on such models, your model would update with a feature created by another
user, as if she were yourself and you had magically created the feature in 'your'
model (as there would be no aritifical distinction between 'your' model and 'her'
model). You would be, in fact, working on the SAME model.

This might be in the realm of the PDM/PLM system which could provide a shared
multiuser memory space. But, it wouldn't amount to another kludgey trick for
reconciling parts with disparate geometry. They can't be reconciled in the PDM
realm; they must be reconciled in the part realm

Thanks for pointing out not only some of the possiblities but the inherent
weaknesses. I was thinking, while you guys were showing alternatives and
workarounds, that there was even more stuff, like maybe ^C ^V, copy features or
copy geom from other model or copy shrinkwrap froom other model or merge from
other model or inheritance. Also, thanks for pointing out that most things get you
a dumb, nonparametric, featureless, surface-based "solid". None of these things
answer the challenge presented by the SE user~~simultaneous feature creation on a
single model. So far, it is not posssible, while the kludges that have been
presented ought to be possible on ANY modelling sytem. While the ultimate isn't
possible, the GROOVE approach comes closer than anything previously conceived to
this goal of multiuser, simultaneous modelling.

David Janes

 

Maybe it has but maybe it was transformed, bastardized, gutted, turned into a
subset of top down modelling. Who knows. I always had the impression that CV was
mainly a wireframe/surface modeller. In which case, lacking the kind of
history/p-c/dependecy relations that "feature based, parametric, solid modelling"
was based on, CV might have had far fewer obstacles to implementing multi-user,
near-simultaneous access to models/assemblies. I know that surface modelling
(which remains surfaces/quilts until 'solidified') is less of a hassle
re:simultaneous modelling than are features which naturally form associations and
dependencies with each other (which avoids surfacing "merge" challenges).
Surfacing just involves far fewer history and dependency challenges than do solid
features, so working in CV with multiple users might have involved fewer
complications from the outset. I've only ever seen CV products and never used it
so I could be all wet on this. Remarkably, CV CADDS5 is still a viable product and
continues to sell.