<wpwrak>
coming next: the solvespace social network ! socialspace ? :)
<whitequark>
why?
<whitequark>
we already have two social networks: solvespace IRC and solvespace.com/forum.pl
<wpwrak>
i mean with solvespace as a means to access it, extrapolating from what ohsix does :)
<whitequark>
oh
<whitequark>
I actually thought about that, integrating libgit into solvespace itself, and adding community part repositories
<whitequark>
it's a lot of work and I don't think I'll ever get over to doing it
<wpwrak>
yeah, may be tricky to get right, also considering how updated may happen
<whitequark>
there are some ways to do merges. CRDTs...
<wpwrak>
hmm, haven't heard of that. but ... from what i gather from the wikipedia articke, they still can't know how to resolve a conflict. i.e., you'd need some additional formalism on top for such decisions. correct ?
<whitequark>
yeah
<whitequark>
they're also somewhat nightmarish to get right
<whitequark>
it's a nice research project :)
<whitequark>
maybe one or two PhDs worth
<whitequark>
I think the use of CRDTs in Etherpad is probably the most dramatic combination of something ridiculously abstract and ridiculously down-to-earth
<whitequark>
I mean, Haskell comes to mind, but Etherpad is way more useful, and category theory (to the extent it's used in Haskell) is not that hard
<wpwrak>
heh :) i guess once you have a good oracle for deciding what to do with arbitrary merges, you've successfully obsoleted human program development
<whitequark>
nah
<whitequark>
this is solved by restructuring your entire workflow around only having the kinds of merges that can be resolved automatically
<whitequark>
that probably makes it unbearable
<whitequark>
that is, if you can come up with one. hence, one or two PhDs
<wpwrak>
i see that you have big plans ;-)
L29Ah has joined #solvespace
Na has quit [Quit: WeeChat 1.6]
Na has joined #solvespace
f8l has quit [Ping timeout: 240 seconds]
f8l has joined #solvespace
d42 has quit [Excess Flood]
d42 has joined #solvespace
<ohsix>
whitequark: can you supply an offset somehow to a lathe operation, and set the number of turns? (basically, make threads)
<ohsix>
engineer friend thought i modeled something with solvespace, wondered how i did it; most thread/knurling stuff is made with openscad
<ohsix>
also searching bugs and stuff, feel bad just asking all the time
<Na>
ohsix: You can't do real threads unfortunately (unless something's changed recently)
<Na>
No offset/helical lathe, and no partial rotations. Apparently doing NURBS things with rotated surfaces is hard
<ohsix>
hm
<ohsix>
well, not a big deal atm since it's just going to a 3d printer
<ohsix>
i need to somehow come to know why that stuff is hard/impossible :D
<Na>
There are some less terse explanations about it on the forum (though not especially in depth). Your google keywords are "screw", "lathe", and "triangle mesh"
<ohsix>
probably why it is called lathe and not revolve huh
<ohsix>
or loft, or whatever 3d modelling packages call it :p