[ox-en] Conversation with an engineer
- From: Stefan Merten <smerten oekonux.de>
- Date: Tue, 21 Sep 2010 20:11:09 +0200
Hi list!
Recently I had a nice conversation with one of my neighbours. He is an
mechanical engineer and works for an automotive supplier. Of course I
asked him a lot of questions which are relevant for Oekonux and in
this mail I'd like to share this conversation with you.
One of my main interests of course was a comparison between
information production and material production. Well, the answer is
twofold. As an *mechanical* engineer my neighbour works more or less
exclusively with *computers*. In fact his day to day work largely
consists of manipulating 3D CAD models - and the remainder seems
largely to consist of talking to other.
This work style reminded me in many ways in my (former) work: As a
software developer it is more or less the same that you largely work
with models in computers - though you express them in programming
languages instead of CAD models - and the rest is talking (as a
software architect as I am today I guess you are talking more than
anything else ;-) ...). Admittedly this didn't totally surprise me but
I was still amazed of *how* similar the work in practice is. It's only
the subject we are working on which is different.
When we talked about the CAD models he is working on he made one
interesting remark: Today everything is in the CAD model. This means
that in the past of course you had plans and engineering drawings, but
those did not contain all the details a modern CAD model contains. In
the past you had to rely much more on mechanical models than today. I
understood this as another indication of how much the mechanical
engineering process moved into the computer meanwhile and is more and
more dominated by this move.
One of the consequences of these details in the CAD models is that you
can now compute results which formerly you had to build an mechanical
model for, crash it against a wall and during this have high-speed
cameras filming what happens during the crash. Today you can feed the
CAD model into a finite element analysis software and compute what the
result of a crash would be. Of course this is much cheaper than the
whole process involving mechanical models.
Still they are building mechanical models from time to time. Not very
surprising they often use 3D printers (aka fabbers) for this. Of
course I asked him whether he could imagine that machines like 3D
printers could do the whole "materialization" prozess. For him this
seems unlikely. Building the products involves processing steel and
the resulting products need to have very good material qualities which
can not be achieved by the technologies available for 3D printers at
the moment. Especially during the last years the trend to lighter
components in the automobile industry is strong and one result is that
meanwhile they are using high-tensile steel types which are probably
even more difficult to process than more classical sorts of steel.
Still the building process is highly automated. He told me that
meanwhile there are some plants in Poland which for Germany is
something like the low cost engineering site. Nonetheless even those
plants are highly automated and consequently do not need much labor.
Of course plants like this are expensive to build but he agreed that
the ongoing automatization process leads to production machines which
are more and more universal and what they really build is defined by
software rather than by their construction. In a way such a plant then
would be a large scale 3D printer.
One point he made that in his company he has two types of
requirements. One requirement is the requirement from the customer for
a certain product. The other requirement is a company internal attempt
to standardize as much as possible. In a way this again is similar to
software development for a customer where you also want to reuse
things you invented before.
My neighbor also made clear that a development of a new product takes
many years. So even a company for mechanical engineering has the
problem that the costs of development need to be put into each piece
of product you sell - an effect we regularly see in commercial
software. I think this should not be underestimated: In former times
the main costs of one piece was the actual building process because
this involved much labor. Today the design process is in many cases
already the biggest cost driver and with highly automated and
universal plants you have yet another type of cost which is not
directly related to a certain product.
Of course I asked him whether he could imagine that he would do this
work because of Selbstentfaltung instead of being paid for it. Well,
since he talked very enthusiastically about his job my guess was
right: of course he could. This is an answer which many years ago
Christof Beaupoil also gave: Just don't think that mechanical
engineers are very differnt from software hackers - during their
weekends they are in their garages and build fancy things. Once more
it is the creative work, the challenge to produce useful things which
motivate people like him.
However, my neighbor saw one one big limitation: the software he is
working with in his company is (probably) proprietary and expensive.
He would need his software to do his work, of course. Then I replied
that there should be Free Software for him. I just googled for it and
saw that there are many (try "open source" and CAD). However, they may
not be as powerful as proprietary solutions yet.
But I think the message is positive: If you would like to support
Selbstentfaltung of mechanical engineers bringing us closer to Free
Hardware and if you are able to program then Selbstentfalt in such
projects!
Grüße
Stefan