This was posted before... "you could also make most of the parts to make another 3D printer" is not self-replication.
We don't expect this machine to be able to make itself out of thin air (though that would be awesome). But in order to be self-replicating, it must be able to build and assemble itself automatically from a set of input materials with no human intervention other than providing the input materials.
As it is, it is not self-replicating, it is merely "capable of producing some of its own parts to help in the later assembly of a copy of itself" - by no means an unworthy achievement, but let's call a cat a cat.
If it was able to produce all its parts but needed human intervention to assemble them, it would still not be self-replicating, merely "capable of producing all of its parts so that copies can be manually assembled".
Unfortunately, this machine is incapable of self-replication by design, since it cannot interace with an output product larger than or the same size as itself. Given this, it is even more wrong to call it self-replicating.
This reminded me of Cory Doctorow's great story "Themepunks", which includes self-replicating 3-D printers. You can read it for free on Salon.com, but the links between the chapters are a bit broken, so here they are if you are interested:
By the same reasoning you could say that humans are not self replicating. Even humans require some basic inputs (essential amino acids) to function. The same goes for the rep rap, and in this case it needs a microcontroller or what have you.
True, I should have said the goal is self replication and each RepRap version intends to get one step closer to achieving process. All the parts that are within the limitations of the machine are replicated and a community has been established to trade such parts. I do apologize for coming off as trying to push a sensationalist headline.
Don't apologize. Usually those behind the effort really try to play this up. I personally think the approach is flawed.
I would start with a real factory that can make everything in the factory. It would probably be the size of a big building.
Just like the first computers.
Refine the components, and make it smaller and smaller.
The initial goal shouldn't be a tiny maker, but something light enough to send to Mars or the Moon. I've been told (by one of Hod Lipson's students actually), that NASA estimates such a factory today would weigh in at 100 tons. That seemed small to me. They should try to make this.
It's the first thing you'd need to make if we knew of a big asteroid that couldn't be stopped.
Man! these kinda news only arouse my interest to combine my programming knowledge with some yet-to-learn microcontroller stuff. Both seem like a deadly combo that would make me wanna stop sleeping for 7 hours and situp and start learning electronics.
Astonishingly, this RepRap is based on Arduino!!! The Arduino Decimelia board has 16KB of memory(!). So it seems like the arduino serves for temporarily storing variable data while the computer does the processing. Wouldn't the printer be more faster if it uses a BeagleBoard or something like that? It would enable faster calculations locally and would help in saving the processing power of the computer for some other purpose(maybe to be to do further calculations since printing 3D requires more calculations ofcourse).
The http://parts.reprap.org site is a great idea. It gives those who wanna build reprap a great idea about prices and parts suppliers and what to specifically buy. It even supports creating your shopping cart on Digikey and Amazon etc :)
A cool idea with a really cool and helpful resource site.
And IMO, it wouldn't be too late before we seen an IronMan. I remember a scene in the movie in which Stark selects his colors for the suit and the robot paints the suit while Tony waits. Now its more possible ('realistic') to even design the suit and wait for the printer to print it.
Don't be shocked if someone comes outta the dark shouts I'm IronMan in a few days :D
"Not counting nuts and bolts RepRap can make 60% of its parts; the other parts are designed to be cheaply available everywhere. This is an interesting coincidence: we can make 60% of our proteins; the other parts are evolved to be cheaply available everywhere..."
Early prototypes of things that will be better in a few generations are cool.
However, Fab@Home and RepRap are sad, and make me embarrassed for them.
I'm hugely interested in self replicating machinery, metalworking, and other forms of constructing items (I launched http://SmartFlix.com based on these interests).
...and I've done a fair bit of reading on RepRap (I was considering building one a year or so back).
I am dramatically underwhelmed.
The construction is shoddy, the technique (basically hot-glue extrusion of thermoplastics) is hackish and has poor tolerances, the concept of "self replication" is grossly dumbed down (a very very large number of parts are not constructed, even from high quality inputs, but just store bought), etc., etc.
In short, I was fairly embarrassed for the RepRap folks.
Their toy is somewhat cool (although nowhere near as cool as, say, the Gingery lathe that is cast from aluminum melted in a cast iron cookpot, and machines itself as construction progresses), but to claim that it represents any important step forward in self replication ... not so much.
They obviously aren't injection-mold quality, but very usable.
The main problem that the reprap has is dribbling when they are moving the extruder head from place to place. This has become less of a problem lately, as they've made some better valves. Also, getting the right settings for temperature, extrusion rate, and lateral speed have taken time.
At least the reprap community is vibrant, and interested in improving the tool, as compared with the stagnant fab@home.
It will probably never reach "self-replication." You can always come up with another level of "self-sufficiency" that it will not meet. It will however, become a useful tool capable of making some of it's own parts cheaply, and assisting in bootstrapping slightly better versions of itself.
My first thought is that encouraging people to make their own highly inexpensive plastic objects and to spread the ability to their neighbors could lead to a significant increase in disposable plastic usage. In general, plastic doesn't really go away. Anyone know if the plastic used here is recyclable, biodegradable or otherwise mitigates the long term (granted, I mean, really long term) concern of flooding landfills?
The main plastics used are ABS and HDPE. HDPE is commonly used in milk jugs, so you could recycle your own milk jugs and make stuff at the same time.
Obviously, since all they are doing is melting a thermoplastic and extruding it, it is trivial to recycle the product: just grind it up and make something else.
I wonder how to get one, though. I searched on Twitter but apparently nobody in my vicinity has a copy. The closest reference I have seen was on one of Guy Kawasaki's photographs of the Google Lab in Munich. Perhaps I really should try to contact them.
We don't expect this machine to be able to make itself out of thin air (though that would be awesome). But in order to be self-replicating, it must be able to build and assemble itself automatically from a set of input materials with no human intervention other than providing the input materials.
As it is, it is not self-replicating, it is merely "capable of producing some of its own parts to help in the later assembly of a copy of itself" - by no means an unworthy achievement, but let's call a cat a cat.
If it was able to produce all its parts but needed human intervention to assemble them, it would still not be self-replicating, merely "capable of producing all of its parts so that copies can be manually assembled".
Unfortunately, this machine is incapable of self-replication by design, since it cannot interace with an output product larger than or the same size as itself. Given this, it is even more wrong to call it self-replicating.