Some preceding information:
With the term "crystolecules" I'm going to refer to the diamondoid molecular machine elements of advanced prospective nanosystems.
Before I start I need to take a step back:
I suspect the nanorex website went down because of the drop in funding due to the NNI censorship incident. Is that correct?
I doubt it would be expensive to at least keep it as an read only archive.
But I guess there is another reason for why it was not put up again.
Namely: The animations of crystolecules
Check out this blogpost:
"Nanomachines: How the Videos Lie to Scientists"
"By now, many scientists have seen videos of molecular-scale mechanical devices like the one shown here, and I have no way to know how many have concluded that the devices are a lot of rubbish ... If only all these videos could be recalled and upgraded in the way shown below ... In short, the videos seem to show devices in which the drag of sliding friction would be enormous, and the rate of heating would be astronomical. ... What the video shows isn’t vibrational motion ... but instead a stroboscopic sample of atomic positions ... "
here is the misleading stroboscopic illusion:
here is a more realistic visualization:
I feel the current situation makes it increasingly difficult to find the original files. I don't think hiding those animations and waiting till there are better methods for rendering crystolecules is a good idea. Instead I think it should suffice to add the note "beware of the stroboscopic illusion" (ideally directly onto the pixel-data so it is safely preserved).
Do you agree / disagree / ... ?
There is a second trapdoor in those depictions that might lead to false assumptions and quick dismissal.
E. Drexler doesn't note it in the aforementioned blogpost but has AFAIK mentioned this elsewhere (in fragments?).
It is about the misjudgment of the stability of surfaces.
While hydrogen terminated diamond surfaces (that do not even touch each other) do not pose so much of a problem (there are even detailed papers analyzing diamond surface stability on small particles by now) many crystolecules feature surfaces with fancy passivations containing other elements that may introduce significant internal stresses.
Since scientists are currently forced to work in reactive environments (including water, air and even UHV) they tend to judge chemical stability more than thermal stability. But in the envisioned working environment (a practically perfect vacuum PPV) thermal stability alone suffices.
Thermal stability alone is a much much weaker requirement than chemical stability against aggressive media thus the range of usable structures gets a lot bigger in a PPV environment. 
Here are some examples:
Lesser subject to mislead criticism: purely hydrocarbon universal joint:
certainly stable 20°C Vacuum AND certainly stable in aggressive chemical media.
Stronger subject to mislead criticism: alternatively passivated universal joint:
certainly stable 20°C Vacuum BUT stability in chemical aggressive media (including water and air) not yet investigated.
Stronger subject to mislead criticism: Drexlers famous big bearing!
I made this 3D printable version
The data is here: http://www.thingiverse.com/thing:631715
If you want one too you can order it here at shapeways for net cost price.
Lesser subject to mislead criticism: "DMSE-Tetrapod ("crystolecule")" (of my own design)
There is someone planning to gamify something nanofactory related (crystolecule design?).
Column bottom left here: https://hgfedcba-box.appspot.com/
I doubt this is going anywhere though see: https://sites.google.com/site/roboticscoop/home and judge for yourself.
I'd like to think that gamifying crystolecule design is a good idea.
Do you think it is practicably possible ... ?
Since this is "future backward" work (analog to top down & bottom up) I think one needs to be especially careful considering you let people loose on those problems who have no prior knowledge at all.
Here's one example for a possible pitfall:
Putting too much oxygen near carbon.
If you replace all the silicon atoms in normal Quartz (SiO2/silicon dioxide) with carbon you'd end up with something like "carbon quartz". While this material is probably thermally stable at room temperature (and super hard) it is very likely to be a highly potent explosive almost like sp3-nitrogen (solid at 20°C 1bar).
Also the last published nanoengineer-1 version does not correctly simulate electron deficiency bonds! You can check that by trying to enclosing a BN pair into a tiny diamond crystal. The result is waay to much repulsion between B and N. It's almost as if you had included two nitrogens atoms resulting in massive overlap repulsion between the two electron lone pairs. To prevent flawed designs emerging from this either boron and aluminium need to be made unavailable or the software needs to be extended to be able to handle these correctly (Is this possible with mere mass spring models?)
If gamifying crystolecules (among other things) works out we might really build up quite a bit of an
"left foot theoretical scientific overhang" like J. Storrs Hall mentions in this video:
I think this video needs more views. Please share it.
By now though I personally don't think we have so great of an overhang yet but that may change.
ps: (I'll edit in the images once I've dug them out - done)
 Does anyone here know whether the nanorex website with its nanoengineer-1 development wiki still exist?
Here's the old address - almost dead link (intended):
As far as I can remember there was some interesting stuff in there. So interesting that I even made printouts. Maybe I'll find them ...
 The only restriction is that everything on diamondoid atomically precise products that faces outward IS chemically stable so there are a lot less options available.
Related note regarding recycling: (a speculative idea)
Diamond does not dissolve in water or rot away so it has good potential to pollute the environment if it somehow manages to leave macroscopic chunks of machine phase.
Here is an idea how to make APM products biodegradable:
Instead of a thin diamond shell that is almost indestructible by all means of weathering and all of natures bio-attacks one adds a rather thick sacrificial layer that slowly dissolves or rubs off. This makes the hull. If the product is intended for longer use it may be replenishable from the inside. (Microcomponent dis-assembly will become harder with undefined outermost surfaces though)
One diamondoid material (of several that use abundant elements exclusively) that may be usable for this is I think periclase (= Magnesium Oxide = MgO). It is a salt that does only very slowly dissolve in water. If the whole product (not a nanofactory - rather drinking bottles and the other stuff that usually ends up swimming in the sea) shall be biodegradable one must forgo using diamond for all its internal parts too.
Periclase wont work well for bearings though - anyone any idea whether it is possible to sensibly passivate such a salt for bearing interfaces - it should be usable for structural elements though. Quartz (a substance nature knows well) might be better usable for bearing surfaces. Oxygen bridges make it much more porous than C, Si, SiC and similar compounds though.
Btw I collect and classify many materials that may be suitable for advanced nano-fabrication here:
One of my favorites is the yet mysterious "beta carbon nitride" (no big chunks made yet)
Which is what you get when you make the most out of air.