My idea of low cost is different, my main reason being there is a big chance that it'll sit in a draw doing nothing if I don't get round to using it.
[as an aside: Is this of any use for experimenting with, anything else required? http://uk.farnell.com/digilent.....tt=2061840
The issue I can see here is that in order to do FPGA stuff there are a number of things you have to have available when you do it (high end PC, non-cheap dev board, installed tools and plenty of time). Each of those push it out of the casual hobbist window (such as I am in - unless my work requires otherwise, it'd be good to build background knowledge).
Now is there a need for a RPi FPGA board? Perhaps there is...but it seems like it is not the platform suited for design of the FPGAs itself.
Could there be value though in creating a shield which holds an FPGA which can be reconfigured by the RPi itself?
If it is possible to do so, you'd then have a re-programmable hardware block which could be switched on to suit a range of applications on the fly.
Could be an interesting way to work...probably impractical though.
Also I've no idea how the abilities of FPGAs scale to cost etc, would a $10 part ($20 board) do anything useful, or are we looking at $50+ for a good FPGA?
As a learning tool, on a much smaller scale, is there a programmable logic chip which effectively allows you to program hardware logic circuits, such a implement a shift-register, or a memory cell etc (I know this can be done with software simulation, but hardware makes it much more rewarding). OT from FPGA, but an excellent way to learn about hardware logic circuits...I've not worked in this area so I don't know what is available.
So all you have to do is find about 50 people who can agree on the FPGA and the I/O each willing to pay ~$75-$150 dollars (Price depends on FPGA chosen). If you are still alive after that battle, talk to me again with hard figures and I will see if I can knock out the 50-ish boards in about eight weeks time.
That might not be that impossible (in theory)...if you consider asking 5 universities which are interested in teaching FPGA design, it'd be easy to expect each to invest in 10 boards (if they were serious about teaching). Of course, when there are 1million+ RPi's out there, perhaps there would be plenty of interested parties.
But then what advantage would that have over a normal dev board?