Print quality still could be better, I had lots of "strings" to remove. While doing this, one wing broke off (I glued it back on).
The thing was printed in 2 parts. Gluing them together did not work as well as I hoped as the pieces had quite some warping.
Maybe it would print better in PLA instead of ABS.
But hey, the kids like it! :-)

http://www.thingiverse.com/bre
http://www.thingiverse.com/thing:3006

The STL and the OpenSCAD files are now on thingiverse: http://www.thingiverse.com/thing:3177

After playing around with skeinforge for some days I had enough of fiddling around with its loads of parameters. It eventually printed my test objects fine with 0.5mm layers, but other objects were not that good. And my 0.3mm layer setup would have needed more work too.

So I bought "netfabb Engine for Rapman Basic",a 99€ (+VAT) software from Germany based netfabb GmbH (http://netfabb.com/). They have a free software called Studio Basic, which is great for examining and fixing STL files, but I does not have any gcode generation functionality.
There's no eval version of "netfabb Engine for Rapman Basic", so you cannot judge the gcode quality before buying, but I did not get disappointed. Calibration is easy, you simply print a test object, take measurement and enter the values back into the software. After that (and probably adjusting the printing temperature) you are ready to go. Awesome! The quality is better then what I was able to do with skeinforge. I'll post some printed examples the next days.

The software uses a unusual "pillar" raft (watch it  get printed), it's a bit hard to get off the print bed, but works nicely. But you still can choose to use a "traditional raft" if you want.

What I find a bit strange is that it does not use the "no string function" implemented in the latest RapMan firmware (version 1.0.8 and later). Using it the prints probably would have less strings you have to remove.

The downside of hiding away all the complexity of generating gcode for you is that you have less influence over the result. There are "expert settings", which lets you adjust a few more things, but being a commercial closed-source tool, netfabb is not a tool for users who want to be able to change anything.

Some quick notes to summarize my experience of assembling a RepMan 3.1 kit:

• It took me about 30 working hours until the first print
• The documentation is good, but not perfect. Sometimes a photo does not match the text or the part list for a certain step misses something. Nothing serious, you can figure it out. But still, if these glitches were ironed out one could build it faster and with more confidence. Now you keep asking yourself: Is this right? Am I missing something or is the docu wrong? Also the docu is split up: there is a text with photos to follow along, there are videos (for certain build steps), there is a 3D view for each build step and there's an animation. And there's a wiki. All is great, but I still wished I had exactly one document I just could have followed along.
• The acrylic parts are laser cut, and therefore have high precision. No need for filing away bits for exact fit or something. Also they look great. Downside: they break pretty easily. The kit includes training parts where you can test how much you can tighten screws on them before they break. This is really needed. Still I now have some small cracks in some (luckily rather unimportant) parts.
• My kit arrived with the large extruder acrylic part broken. I contacted BitsFromBytes and they immediately sent me a replacement part. No questions asked, no extra costs. Nice!
• Depending on where you live, some of the tools need might be hard to get. I live in Germany, which is in the "metric part" of the world. A 1/16" Allen key is not something I had at home. In fact, I still have none and that's why my pulleys come loose from time to time. I ordered one on ebay, but still waiting for it to arrive. If you are in the US you might have a 1/16" Allen key, but note that you also need a 1.5mm and 2mm Allen key (among other tools). So it still might be worth checking what tools you have before starting the assembly.
• As reported in an earlier post, I did not use fire cement for my extruder. My fire cement was dry and hard by the time I wanted to use it. I didn't bother getting new one, but instead tried attaching the heating wire and the thermistor to the hot end using Kapton tape. Didn't work on the first try (thermistor lost contact thus the hot end was heating up too much) but now works well. I wound some wire around the hot end to hold the thermistor in place. The thermistor is wrapped in Kapton too for electrical insulation.
• After I finished the assembly I printed countless test objects to get my skeinforge setting right. This really is not easy. The BfB wiki explains most of the parameters but not all. There are helpful hints on the net, but there are different approaches to reach a good combination of parameters and most of them are for older versions of skeinforge, where the parameters where different.

In the photo above there are some of my better test objects. To get a feeling for the size, the square (it actually a cube where I stopped printing after a couple of slices) is 2cm wide.
None of them is perfect, but they get better every time.

Overall I'm pretty pleased with the RepMan. It's more expensive than a Makerbot, but it offers a larger printing area. I also like the electronics which allow printing without a PC attached. Compared to building a Mendel it is just nice to have a kit that includes everything. Building a Mendel probably would take much longer.

Completed the first prints. I used (of course!) the mini-mug model.
The very first print had a hole in it and it was not round. I had a problem with the y-direction, too much play, a pulley was not fixed properly.
With the second print the geometry was good, but it was printed too hot it seems, the print dissolved itself at higher layers (I aborted it).
I looks like the thermistor lost its direct connection to the extruder, so it was heated up more than it should. I rebuild the extruder - again using the Kapton tape technique instead with fire cement - and this time the print looked much better (I also switched to black ABS).
The geometry still is not perfect, it has a slight skew in x-direction. I have to check what causes this. The print has some "hair" inside, nothing serious, easy to clean up, but maybe the print temperature was a bit too high.

Yes! Assembly finally finished. All looks good (after fixing  a wiring mistake with the extruder), the "test raft" pattern printed fine.
First real print tomorrow!

I tried using Blender a long time ago, but I found the GUI really combersome. Even after reading a tutorial I wasn't able to create something useful.

Now Blender 2.5 (alpha 2) is available and as they have promised to have a more intuitive user interface, I gave it another try. My goal was to create a 3d-printable model of a Star Wars Tie Interceptor. There are a couple of nice (but short! :-))) Blender tutorials for newbies on thingiverse (http://blog.thingiverse.com/2010/05/05/assembled-blender-tutorials/). They are for previous versions of Blender, but they are easy to adopt and the keyboard shortcuts are the same.

So this is what I did:

Blender 2.5 is really nice! Very stable already, and taken account that I hardly did any 3d modeling before it allowed me to do this pretty fast (a few hours). Bravo!

The printable STL file can be found on thingiverse: http://www.thingiverse.com/thing:3006

Was working on the hot end today. Unfortunately I couldn't use the fire cement which came with the RapMan, it was already dry and hard.
I probably could ask BitsFromBytes to send me new cement, but I don't like the idea of waiting a couple of days to finish my 3d printer.
So I tried the RapRap way and used Kapton tape to fix the nichrome wire and the thermistor to the nozzle. I'm a bit concerned about how good the thermistor is (thermally) connected to the nozzle. We will see.

So extruder including hot end is finished, only a few things are missing now.

Starting assembling the extruder. Got the filament feed ready, but the hot end is still missing.