[ExI] Fwd: [Open Manufacturing] Re: Who's Developing P2P-L2G Related Software?

Bryan Bishop kanzure at gmail.com
Sat Aug 8 20:54:57 UTC 2009

---------- Forwarded message ----------
From: Bryan Bishop <kanzure at gmail.com>
Date: Sat, Aug 8, 2009 at 3:30 PM
Subject: Re: [Open Manufacturing] Re: Who's Developing P2P-L2G Related Software?
To: openmanufacturing at googlegroups.com, kanzure at gmail.com, diybio
<diybio at googlegroups.com>, diytranshumanist at googlegroups.com

On Sat, Aug 8, 2009 at 3:04 PM, ecd wrote:
> Yo! First post in this group. Found my way here via 4chan, oddly
> enough. (/r9k/ -> anarchy discussion -> disclosed some projects of
> mine & solicited interest/asked about extant work -> someone linked
> Bryan Bishop's page -> "Yo! First post in this group"...)

Huh, I haven't hung out in /r9k/ on 4chan in a while- or in fact,
ever- so I'm glad to see that you somehow made your way here anyway.

> The project I described in the aforementioned thread was essentially
> an online (but downloadable) "tech tree". (No, nothing to show at this
> point. :(   ) Since I anticipate the related tools (mainly data

Yep, you're in the right place. This is what we're mainly working on
here. The project idea has been mentioned in the past (decades ago)
hundreds of times over, but nobody has sat down and done it. I found
fenn (one of the other committers) by searching around for people
interested in David Gingery, and we've been working on this project
ever since we met. Smari and Sam Rose and a few others also started
emailing around last year about implementing this tech tree.

You're welcome to view our progress.


In particular, you can see a taxonomy of manufacturing processes here:


You can see some of the details filled out here:


You can see an example "package" in skdb, a package for a generic screw:


Ultimately this is hard to explain to individuals who are unfamiliar
with the concept of a "tech tree", or "tech graph", or unfamiliar with
"apt-get" or other free and open source software package managers.
These days I just tend to tell the diybio folks and the
openmanufacturing folks that it's "apt-get but for hardware". If you
haven't seen the xkcd sandwich reference,--


And of course, someone put some work into that very project already:


So, skdb is "apt-get but for hardware"- the idea is to be able to say
"skdb-get install milling-machine", or "skdb-make robot-maid", and
then get all of the requisite tools either by getting instructions
(computational representations of instructables), or by "opting out"
of the tree and just buying some OEM or proprietary components at a
certain point-- which I only mention because it's common that people
just want a kit, or don't want to have to forge every metal component
that they require to carry out their herculean feats.

> visualization & interchange (formats, languages, etc.)) will end up
> being pretty versatile, I was going to integrate them into a general-
> purpose collaboration-oriented website (and other services, if I could
> afford it and there was interest), where I'd also host the major
> projects that I personally wanted to develop with the tools:

So, one front-end idea that I have recently been bouncing around is
called "djangit". It's a python + django + git + wiki frontend system.
The idea is that many people don't want to hear about the guts of
skdb, but at the same time there's no reason to ignore proper revision
control systems; simultaneously, implementing in django means that the
python modules for skdb can be hooked in easily for rendering of the
package data (like the screw package), etc., while still allowing
human input over the web if someone so desires. At the moment djangit
isn't quite functional because I've been neglecting it for the past
few weeks.

>  * the tech tree (multicontextual (historical (from all perspectives)/
> [bio/psycho/socio]logical/cosmic/metaphysical contexts describing the
> development and applications of the various techs), and ground-up

> (such that someone who could read and had infinite time could proceed
> from step one ("find stones that look like this, mash them together
> such that they spark, collect hot sparks in dry fibrous material...",
> etc.) to building space ships. Yes, I've heard of the Foundation
> books. :P  But no, I've not read them.)

In your spare time, you might want to read more about what we've
previously said about bootstrapping on this list--


>  * major subsets of the tech tree (significant enough to people I care
> about/my own interests to warrant their own project pages (while still
> referencing&building upon the main body of data)):

One of the problems that I continue to come across is that many of the
"open source hardware" sites on the web are just keeping photographs
of their information, which isn't useful because there's actual
engineering information involved that should be uploaded (like a CAD
file, etc.). But ultimately, yes, just like the alioth server on
debian, it would be useful to have a way to link to individual
projects and their presence on the web, yes.

>  -primitive technology (sticks & stones, live off the land, etc.)
>  -chemical (substances, uses, algorithms, models, etc)
>  -CS stuff (for the development of "fluid" operating systems, another
> project I've an interest in.)
>  -biological (...I think there's already something like this... MIT?)
>  -materials (processes and tools for production, measurement,
> evaluation, etc.)
>  -tools (measurement, manipulation, etc.)
>  -salvage (where to find what in presently available artifacts, why
> they're there, and how you can use them)
>  -reverse-engineering (processes, results, tools, etc.)

Yep, we seem to be in agreement.

> In building a site around this tree, I imagine people (at least
> myself) could do write-ups (eg "Using the data in the multimeter
> nodes, I've built my own. Here's how I did it..."), and I would feel
> immensely fulfilled if people actually devised and contributed novel
> techs using the data I hope to accumulate & organize.
> I have two ideas that, while probably unoriginal, I think might
> distinguish this project:

That's one part that the debian community (among others) has solved.
It is well known that engineering is not necessarily the most easy
task in the world because you can't just "engineer de novo"- it would
be pointless to engineer everything under the sun from scratch each
time you build it, right? So for this reason, there are already
"package maintainers" in the skdb community that accept projects from
others and help them "package them up" into the packaging format. This
way, everyone can just sit at their computer and say "sudo make me a
sandwich" and the computer handles all of the details like ordering
inventory, or printing out new lego-manual-style instructions for how
to assemble parts into a system that you wanted, etc. For those who
run shops or who have large machinery laying around, it would be ideal
to allow those machines to assemble the components for you, but
that'll lead this discussion off topic fairly quickly. ;-)

> 1) on top of all that data, you could overlay data structures that
> organize it however you like. You could then share those structures,
> edit them, make meta-structures, search algorithms (also sharable!),
> etc..
> 2) Up/downloading & persistence - take relevant data with you, keep
> your structures and the data they contain sync'd to whatever (and
> whoever[']s[']) version you like.

Yeah, that's called a version or revision control system.

> What can I help with? Perhaps it would be nice to have a master list
> (on someone's wiki or something) of who wants/needs what capabilities
> (and where they want them - what hardware/microarchitectures/devices/
> modes..), and what currently delivers (free/os or not). I guess I'm

We have some of this in the skdb/inventory/ folder but it's not
complete. Smari was working on a web interface to this, but he hasn't
showed up in the IRC channel (#hplusroadmap on freenode) in a few
weeks so I'm not sure what his status is.

> asking you specifically (Hi, Nathan! Pleased to meet you :)   ) to put
> some more detail on that page of resources you've started? I see a
> list of projects, but I don't have a list of the concrete wants/needs
> that they address. Perhaps the rest of you have established enough of
> a group-mind to know what the others speak of when mentioning some
> project, but I haven't had the initiation.

I have absolutely no idea what Nathan is doing with yet another
project. We have a lot of momentum here that he's neglecting, and I've
invited him to learn more on numerous opportunities, but maybe I'm
just getting grumpy and old and grumpy.

> What exactly do various people want to do? I know what I want to do,
> and I have /some/ technical ability, and if there are already projects
> towards similar ends I'd be happy to lend a hand in whatever way I
> can. (By the way, I'm glad to see folks here seem cool with each other
> working on essentially the same problems in their own way.)

I suggest starting off by joining the IRC channel (#hplusroadmap) and
saying hi, hanging around and seeing what's up. No doubt that you'll
slowly start to get the picture of what some of us are doing (or not

> 2) I too am working on a suite of information tools relevant to
> collaboration.

Great. I hope you know your toolchains well :-).

> 3) I am willing to contribute mind/coding power (humble as it may be)
> to other collaborative projects.

Fantastic, I will eat your brain.

> 5) My brain tends to shut down when conversations get too abstract.
> Please keep this in mind when communicating with me. I like hard
> facts, objectives, and a clear context. (...But when I've completed my
> Magnum Opus and I'm finally out from under the thumb of scarcity, I
> will thoroughly enjoy philosophizing and mind-masturbating all over
> the place - Now with greater [augmented?] intelligence and
> experience!)

Hah. Well. We have a rule in the channel: no philosophy. We commonly
violate it without knowing it, until one of us reminds us of this
fact, and we realize we've all been barking up the wrong tree, and get
back to more practical work.

Anyway, nice to meet you.

- Bryan
1 512 203 0507

- Bryan
1 512 203 0507

More information about the extropy-chat mailing list