Hi George,
On Oct 29, 2009, at 11:57 AM, George A Pace wrote:
1. Binary File processing - It looks like FOSSology is reading
them - I just wanted to verify that this is indeed true - and what
changes are coming in future versions
Anything with a binary type, we run through strings (1) and then use
that output to scan for licenses.
2. Timeframe to the next FOSSology version (I guess version 1.2) -
Is this looking more like a 1st half 2010 ?
Yes on 1st half of 2010. We really don't have an estimate, but my
guess is 4 months. Emphasis is on "guess".
3. Having FOSSology use binary and/or source signatures during the
analysis phase - I know we had a conversation on this a while back,
but how far away is FOSSology from being able to use a library of
existing signatures when doing comparisons ..
Are you talking about discovery - finding out what packages are used,
or code clone detection? I think we talked about clone detection
before. There is lots of great research on code clone detection. One
of the best tools is CCFinderX which is going to be open sourced
soon. It is not really suitable for incremental detection like what
we need. We need to compare a relatively small set of code to a large
library of code to find code clones. CCFinderX, along with most other
academic research (DECKARD comes to mind) finds clones within a single
set. However, these could be modified. DECKARD (UC Davis) is also
not freely available, but probably could be made so. My plan was a
bit different though.
My plan was to sponsor some research to get us a state of the art
solution, for this type of detection, that we could plug into
fossology. I put out an academic RFP and got many excellent
proposals. At this time (several months ago) we had money in our
budget to sponsor this and I awarded the gift grant ($20,000 USD).
Unfortunately, this coincided with the market bust and we were not
allowed to send the money. Do you have $20,000 (or $5k, $10k, $15k)
you could donate to a University?
You mentioned binary scanning. I also received a proposal from a
couple of experts in binary license scans. Unfortunately, even
though they have a great deal to benefit by having this in fossology,
they would like to be paid for their efforts. Again, we have no money
for this.
So I'd dearly love to implement a state of the art clone detector and
a super binary scanner, but it's just Mark and I writing code and it's
hard to look past the next version (http://fossology.org/task_list).
This is why I was looking for help from our academic friends and
experts.
If anyone has expertise, time and/or money who would like to support
one or both these agents (or any other agent), please contact me.
Bob Gobeille
b...@fossology.org
_______________________________________________
fossology mailing list
fossology@fossology.org
http://fossology.org/mailman/listinfo/fossology