Report from the COLING/ACL 2006 Workshops Chair Suzanne Stevenson The Workshops Program Committee for COLING/ACL 2006 included five people from a range of geographical regions and representing a range of research areas in computational linguistics and natural language processing. The committee consisted of: Suzanne Stevenson (Chair), University of Toronto Ann Copestake, University of Cambridge Pascale Fung, Hong Kong University of Science & Technology Jamie Henderson, University of Edinburgh Ingrid Zukerman, Monash University Proposals for workshops and other co-located events for COLING/ACL were due on December 9, 2005. All workshop proposals were reviewed by at least 3 members of the Workshops Program Committee. The Workshops Chair collated all ratings and read all reviews, and made the final acceptance decisions. Workshop proposers were notified on January 17, 2006, as to whether their proposal was accepted. We received 24 proposals by the December 9 deadline, plus a late proposal from the EMNLP organizers, for a total of 25 proposed workshops. Of these, 4 were what we termed "co-located events", as established, on-going workshops/conferences: EMNLP, SIGdial, INLG, and TAG+; all of these proposals were accepted. The other 21 proposals were either one-time workshops or on-going workshops not quite of the regularity or size of the other established events. Of these 21, 10 were accepted as is, and 7 were rejected. The 4 remaining proposals consisted of two pairs of proposals that highly overlapped in content and potential audience. The Workshops Chair felt that a stronger program could be achieved if each of those pairs were merged, and the organizers agreed to work together to do so. Thus, we had a final count of 16 events: 3 2-day pre-conference co-located events; 1 2-day post-conference co-located event; 12 post-conference workshops (11 1-day workshops and 1 1.5-day workshop). All accepted events were confirmed for the conference. These events represent a very broad range of research topics in our field, and provide a strong program of participation beyond the main conference. Some topics for discussion by the ACL Exec: 1. Even after putting together all the information available from various sources (prior Workshops Chairs, published ACL info, etc.), numerous questions arose throughout the process of workshop review, selection, and organization. I would suggest that the guidelines published by ACL be expanded and clarified, with "institutional memory" (often in Priscilla Rasmussen's head) set down for future Workshops Chairs, and workshop proposers. This would include clarification of financial arrangements for SIG-sponsored workshops. I also expanded the Workshop Proposal Form and Workshop Proposal Review Form, and it would be nice to have a site where these could be posted for the use of future Workshops Chairs. 2. The treatment of on-going, well-established co-located events and "one off" workshops needs to be distinguished. Events such as EMNLP, SIGdial, INLG, or TAG+ may want to start publicizing earlier than the workshops reviewing schedule allows. Moreover, such established events are not likely to be rejected, and indeed are depended on for helping to create a strong overall conference program that attracts broad participation. The ACL Exec needs to decide on a policy for handling such cases, such as a "fast track" reviewing schedule. The tricky aspect of this will be deciding which events qualify for such fast tracking. For example, the four co-located events for this year (mentioned above) are broadly recognized as on-going workshops or even conferences, while other workshop proposals are clearly one-time events. However, others (such as the fairly regular SIGHAN workshop, or several workshops that were in their 2nd or 3rd incarnation) may not neatly fit into one category or the other. Some form of fast tracking is clearly needed, but guidelines are required for the size and regularity of an event to be considered for this. => Please note that this issue has been raised before, but has not => been resolved. It is very important that the ACL Exec make a => decision on this, to avoid problems that arise with the co-located => events. 3. ACL has a policy of allowing the (workshop) registration fee to be waived for exactly one invited speaker per workshop. A number of organizers complained about this policy, particularly in the case of multiday workshops. I would recommend that the ACL Exec consider a modification to the policy that would allow one workshop registration fee to be waived for each full day of a workshop/co-located event. 4. Two workshop proposals had organizers on the Workshops Program Committee. The Workshops Chair respected conflict of interest issues in these and other cases in making reviewing assignments. The Workshops Chair was a co-organizer of one of the workshops; this was cleared with the conference organizing committee prior to my involvement. It turned out that this workshop proposal received very high reviews, making acceptance of the workshop an easy decision. However, given 7 rejections out of 21 "one-time" proposals, it is clear that some policy needs to be in place for cases in which the Workshops Chair is involved in organizing a workshop. A suggestion from a prior Workshops Chair was to have the General Chair called on to make decisions in such situations. This may involve a reasonable time commitment, since to make a fair decision would require some knowledge of the range of reviews and ratings assigned to all proposals. => This is another issue that has been raised previously with the ACL => Exec, and not resolved. 5. There is no set policy for maximum length of workshops papers. I had assumed that the maximum was the same as for the conference (8 pages), but there is no policy on this. For this year, given the very small impact on publication costs, we allowed workshops up to 12 pages maximum, with strong encouragement to limit papers to 10 pages. However, this was discussed very late in the process (long after most submission deadlines and very close to the camera-ready deadlines). It would be preferable to have the policy decided on prior to the workshop CFPs going out, so that workshop organizers can plan accordingly. (It is worth noting that co-located events may have differing typical paper lengths from the main conference standard; for example, this year TAG+ wanted to have 12-page papers.) => This is another issue that has been raised previously with the ACL => Exec, and not resolved.