Rule Change: Source Code IS Required

Skip to first unread message

James Bowery

Aug 20, 2018, 12:50:48 AM8/20/18
to Hutter Prize
Source code is required as of 2018.

Matt's punctuation was correct.  My apologies for the confusion I created.  

Here is the FAQ entry.

Why do you require submission of documented source code?

A primary goal of this contest is to increase awareness of the relation between compression and (artificial) intelligence, and to foster the development of better compressors. The (ideas and insights behind the) submitted (de)compressors should in turn help to create even better compressors and ultimately in developing smarter AIs. Up until 2017 the source code was not required for participation in the contest, and has also not been released voluntarily. The past submissions are therefore useless to others and may be lost forever. Furthermore this made it difficult for other contestants to beat the as of 2017 four-time winner Alexander Rhatushnyak. Making the source available should rectify these problems. Therefore, as of 2018, the source code is required, which should help to revive the contest, make it easier to build improved compressors by combining ideas, foster collaboration, and ultimately lead to better AI. Contributors can still copyright their code or patent their ideas, as long as non-commercial use, and in particular use by other future contestants, is not restricted.

Mirek Kaim

Aug 21, 2018, 10:40:40 AM8/21/18
to Hutter Prize
For me it just means Prize Committee expects an evolution, not a revolution. In that case it may work, assuming you'll force Alex to release his source code, because - as I've proven in another thread - that change on the rules page was made between may and june last year, so he violated the rules by not making the source code public from the beginning.

Personally though, I was always seeing this contest as a search for potential revolution. Source code requirement will just nuke it.

Of course you may argue that a revolution isn't possible. It's a mistake many scientists made in the past, and many are making it to this day.

Neil Harding

Aug 21, 2018, 11:16:43 AM8/21/18
If you don't want people reading your code, you could always write it
in Perl :) I actually think the biggest drawback is the size of the
data, 100MB is too small for a generalized learning program, I think
the 1GB size may allow for more intelligent algorithms. I spent a long
time on the problem (although I was using Huffman style coding rather
than arithmetic coding for encoding the symbol probalities).
> --
> You received this message because you are subscribed to the Google Groups
> "Hutter Prize" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to
> To post to this group, send email to
> Visit this group at
> For more options, visit

Mirek Kaim

Aug 21, 2018, 6:49:37 PM8/21/18
to Hutter Prize
I prefer Forth :)

As for the data size, 100MB seems enough for me, altough i would welcome 1GB for another reason - such data size, coupled with harsh resource limits, could force innovation. The future is in optimization, not in throwing more computing power at the problem.

just looking

Nov 9, 2018, 4:59:03 PM11/9/18
to Hutter Prize
the tale becomes more and more interesting:-) since alex rhatuyshnyak continue to post right optimised source of inspiration on , rule "source code is required" seems enough reasonable! especially for me there are no matter what to throw in public elf, exe or c/c++ source it all can be decompiled and readable by skilled person! " well documented" is still a big question! is it school literature classes? just give us exact definition:-)
Reply all
Reply to author
0 new messages