Report of COLING/ACL 2006 Publications Chair


As the publications chair for this conference, 
I have directly benefitted from the hard work of the 
previous publications chairs Jason Eisner and Philipp 
Koehn, who have prepared a set of excellent tools which, 
when used in combination with data from START, streamline 
the publication process considerably.  I have made certain 
adjustments to the scripts and various templates to 
customize them for A4 size and with appropriate headers 
to suit this year's conference.  

1. Main conference publications

The printing of main conference proceedings is different 
this year.  As we step forward to protect our environment, 
we are moving away from hardcopies.  Participants do not get 
hardcopy proceedings, unless they explicity request during 
registration and pay extra for them.  It turned out that out 
of 600+ pre-registered participants, only 100+ had opted for 
them.  The demand for the printed companion volume was even less, 
and considering the budget, it was finally decided not to print 
hardcopies of the companion volume.

2. Workshop publications

All 16 workshops have printed proceedings.  The publication 
tools were distributed to the workshop organisers in advance.  
They were responsible to assemble their own volumes, for which 
I had done some overall final editing and proofreading to fit 
everything harmoniously together.

All proceedings covers were designed by the local organizing committee.

3. CDROM

All papers in the main conference and workshops are available 
on CDROM, which are distributed to all participants.  We have 
only included pdf files this time as we have a lot of papers 
and it happened that the ps files alone had almost taken up 
the full capacity of a CD.  Moreover, the conversion of files 
from pdf to ps had not been smooth at all (see technical 
issues below).

All papers are also expected to be available at the ACL 
Anthology by the time of the conference.

4. Technical issues

(a) Deviant paper formatting (e.g. paper size, font 
embedding, etc.) has always been a problem, and it is no
 exception this time.  About one-third of the main conference
  papers and posters were incorrectly sized, although it has
   been emphasised again and again that A4 size is required.
     I had to ask those authors to re-submit a correctly 
     sized version.

(b) The font-embedding problem had been a frustrating one.
  Nimbus fonts cannot be retained when using pdflatex to 
  concatenate individual pdf files.  As my system did not 
  seem to have tetex-3.0, I had to find work-arounds by 
  manipulating the font files and the font-mapping files.

(c) The conversion of the pdf files to ps files was not smooth.  
Workshop organisers had reported problems of various kinds using 
acroread for the conversion, such as missing pages.  I myself also 
experienced unpredictable behaviour of the conversion of papers with 
Asian fonts, not only with acroread but also with pdf2ps.  
It seems Acrobat Professional is more reliable but it works on 
windows.

(d) The publication tools/scripts are unix/linux-based.  Several 
workshop organisers do not use unix and therefore have problems 
running the scripts to assemble their workshop proceedings.  I had 
asked them to supply the necessary materials (e.g. frontmatters, 
paper ordering, etc.)  and run the scripts, and let them proofread 
the product.

(e) The most unexpected hiccup, after all, comes from the printing 
house in the last minute.  Although they managed to solve the problem
 by changing to some more tolerant printers (as they called them), 
 they had to ask the local organisers to proofread the printouts.  
 Thanks to Robert Dale and his team for doing this extra work in the 
 last minute.

5. Suggestions and discussions

(a) It might be better to ask authors to submit all source files, ps, 
pdf of the final papers, in case the publications chair needs a quick 
fix to formatting problems, and it saves the risky process of converting
 pdf to ps.

(b) We should consider whether we will go on to print hardcopy proceedings
 in future conferences.  The problem this year is we could not decide on
  the quantity until very late, after the early registration period.  
  This introduced uncertainty in the publication process.

(c) I wish to re-echo the suggestion made by Jason and Philipp last year
 for centralized computing.  It is important to make sure the up-to-date 
 and correctly configured and mutually compatible software needed to run 
 the scripts are available, for the publication process to go smoothly.  
 The publications chair and the workshop organisers should have access to
  them.

(d) Early communications with the printing house is apparently important.  
Obviously what can be printed on simple office printers is not necessarily
 printable elsewhere.

Olivia Kwong
July 2006