Hi everybody,
It's that time again -- I think we should make a new release of VIFF.
That would be a nice Christmas present for all the VIFFaholics around
the world! :-)
I suggest that we aim for a release after the weekend, on the 22nd.
Does people think this is a good idea?
We still have some outstanding patches:
* I wrote a patch which makes the normal passive multiplication
protocol divide the work between the players so that only a random
subset of 2t+1 players do any work.
The patch is not yet in VIFF since it slows down the performance for
the (n, t) = (3, 1) case:
http://article.gmane.org/gmane.comp.cryptography.viff.devel/495
http://thread.gmane.org/gmane.comp.cryptography.viff.patches/49
I guess it needs more testing and thinking before it can go in.
* Mikkel wrote code to replace the insecure marshal module, but that
turned into a grammar fight :-) and I'm waiting on it to be updated
based on my review:
http://thread.gmane.org/gmane.comp.cryptography.viff.patches/54
This one should be almost ready to go in -- Mikkel you can push it
yourself if you want. But please add the comments I ask for in the
above review.
* Sigurd wrote a faster equality test based on Fermat. I reviewed it
but haven't seen an updated patch:
http://thread.gmane.org/gmane.comp.cryptography.viff.patches/50
This one is also almost good to go!
* I wrote a patch which speeds up and streamlines the handling of
incoming network data. Unfortunately it breaks the test setup so I
haven't include it in VIFF. It's six weeks old, but comments are
still very much appreciated:
http://thread.gmane.org/gmane.comp.cryptography.viff.patches/66
This one is not ready yet unless someone can untangle the test setup
and make it pass with the patch.
Apart from the patches my checklist says that we need to:
* update the code to use the new input/output methods instead of
shamir_share and open. We'll keep the old method names as
depreceated aliases (right now it is the other way around).
* update the NEWS file to reflect all the changes since 0.7.1. This is
mostly done, but someone needs to look over the Mercurial changelog
and update NEWS as needed.
It would also be nice if someone else than me would read through the
file *before* release so that we can avoid any silly typos :-)
* Test the example applications. I've just tested these:
- int-bit-conversion.py
- millionaires.py
- two-fields.py
Pick a couple of programs and test them on your platforms.
You are *very* welcome to write more test programs and to *comment*
the ones already there. The millionaires.py program is probably the
best program to use as a template.
It you can improve the programs so that they test themselves then
please do! What I mean is that some of the programs automatically
figure out if the answer is correct, whereas some require you to
compare the three outputs and figure this out for yourself.
Let me know what you think -- if nobody thinks anything, then we'll
just delay the release.
--
Martin Geisler
___
viff-devel mailing list (http://viff.dk/)
viff-devel@viff.dk
http://lists.viff.dk/listinfo.cgi/viff-devel-viff.dk