The Design Repository
Bryan Bishop
http://repository.designengineeringlab.org/
Free registration required-- most likely it's a funding thing. I'm not
so sure I like the XML or the function structures and the other
matrix-oriented formats at the moment. But the data is at least there.
I currently have a perl script fetching some of the XML because the
server crashes when I try to get all 5000 products from their servers,
but this might change soon. After it finishes I'll throw this into a
git repository.
[[Interestingly enough, it turns out that somebody form the Design
Engineering Lab found *me* before I even knew of its existence, because
of my involvement with the IEEE Manufacturing and Automation Group.
Small world.]]
So what can you do with this information? Nothing. You can't plug it
into anything and have it build it. You could couple the products with
3D models and feed it into a prototyper, but these aren't all plastic
products. Some have microelectronics. Most have metals. And this
information isn't entirely what's being captured there. But it's a
start. Any ideas?
The insightful might notice that I'm really proposing reconfigurable
manufacturing here; what else are you going to do with databases of
designs and knowledge, like OSCOMAK or the metarepository projects?
Just as you generate make files, so should you be able to do with
factory equipment or bootstrapping tools you have nearby.
And yet everyone knows just how much of a custom job a functional
factory is. Highly domain specific and domainly constrained. It might
have some influence on the ability to automate all of this ... maybe.
The details that allow something to actually come together to go
online, however, is what *we*, engineers, have to deal with. End-users
that can't afford to have any mistakes out in the middle of nowhere
would have to use 'hardened' and tested system configurations, no
doubt.
It would be a good first project to automatically design a Rube Goldberg
Machine and then generate all of the instructions for building each of
the components. (Don't worry about a common language for instructions
to machines *and* humans - this could be another abstraction layer in
the instructions/steps for building things from primitives in
manufacturing processes. Needs to be studied more, however.)
- Bryan
________________________________________
http://heybryan.org/
Engineers: http://heybryan.org/exp.html
irc.freenode.net #hplusroadmap