[Corpora-List] discussion on reproducibility at ACL 2011 business meeting

[Min-Yen Kan]

Dear all:

I want to go back to Ted's original post and describe what we are
planning to do with the ACL Anthology with respect to software and
code, and in general (for those who may have missed out on the status
of the Anthology).

In the updated ACL Anthology, we are providing faceted search (the
ability to narrow a search by a particular venue, year, SIG, etc).  We
are planning to allow users of the Anthology to identify papers that
have annotations, awards, datasets, software and errata as part of
this faceting system.  This is part of the official development path
of the Anthology, and I'll be very interested to hear what others have
to say about these planned capability.  The "beta" version of the
Anthology (along with other "bugs" that still need to be handed, can
be seen here:

http://aclanthology.heroku.com/

It's not quite there yet, but we are working on it.  Probably a better
way to see faceted search in action is to visit the ACL Searchbench,
provided by DFKI (link from the Anthology home page).

For code and dataset quality, I agree with some of the comments that
the paper review process may not be the best time to do code review.
It seems that post publication use and review might be better avenues.
 In the Anthology, we can eventually support this with commenting for
both the paper as well as the its associated elements (i.e., dataset /
software).

Right now, for the Anthology, the ACL Exec has approved just the
storage of datasets and software, but nothing as to its long-term
availability or whether to offer a service level beyond this.  It'll
be interesting to hear more about proposals on how this can be
enhanced.

-Min
Your current ACL Anthology Editor

_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora