These scripts are designed for the Publication Chair in an ACL-style conference. They fill in some gaps not covered by the ACLPUB scripts that you get from the START conference system (https://www.softconf.com/).
The proceedings.tgz
that you can download from START has already generated
the HTML and PDFs for the main conference and each workshop.
You still need to generate an overall proceedings that combines all of these
pieces together.
-
Download the ACLPUB scripts and the current main conference proceedings from the START system:
ACLPUB
->Generate
tab ->All
button ->[zip]
link. Save the zip file to the same directory as this README file. -
Unzip the downloaded
all.zip
. -
Set the
ACLPUB
environment variable to point to the directory that you just unzipped. In bash, for example:export ACLPUB="$(pwd)/all"
-
Run the script to download the workshop proceedings and combine them with the main conference proceedings.
./make-cdrom.sh <conference-title> <START-conference-name> <START-book-name> [<START-book-name> ...]
The
<conference-title>
is the name to be used in the combined bibliography, the<START-conference-name>
is the short name of the conference on START, and the<START-book-name>
s are eitherpapers
, for the main conference, or the short name of each workshop on START. For example, at EMNLP 2013 the command looked like:./make-cdrom.sh EMNLP-2013 emnlp2013 papers SPMRL2013 TextGraphs2013
The script will download each of the
proceedings.tgz
files for each of thebook
s, extract them to abooks
directory, and execute a Makefile that will generate the finalcdrom
directory. -
(Optional) Remove .ps files from the
cdrom
directory. These are not used by the ACL Anthology, and it's unlikely that any of the conference attendees will want them, so this will reduce the size of thecdrom
distribution.find cdrom -name "*.ps" -exec rm {} \;
NOTE: If you get an error message like Attention: The file "cdrom/index.html" is out of date
this may mean that you have generated the main conference proceedings after the workshop proceedings.
Try going to the workshop proceedings on START and re-generating them.
There are several common bugs that you should check for in the cdrom
directory.
-
Check that all references to ACL 2010 have been replaced with the name of your conference:
find cdrom -name "*.html" -exec grep 2010 {} \;
-
Check that there are no non-ascii characters in any of the BibTeX files. (If you find any, you will need to contact Softconf Support to have the conversion scripts updated.
find cdrom -name "*.bib" -exec file --mime {} \; | grep -v 'ascii'
If you do find some files that are not ASCII, the following command may help you to identify exactly where the non-ascii character is:
grep --color='auto' "[^ -~]" <filename>
If you find any of these bugs, you will need to make the appropriate fixes on
the START system, re-generate all of the proceedings on START, and then run
the make-cdrom.sh
script to download the new proceedings.tgz
files.
Acronyms and some other words (such as language names) should always be capitalized, and therefore need to be wrapped in {} in BibTeX titles. There is a script that will help you do this.
-
Open
escape_bibtex_titles.py
and edit the lists at the top which identify system names, proper names and language names that need to be capitalized in BibTeX titles. (There are some entries in there now, but these will most likely need to be adjusted for each conference.) -
Run the script on the
books
directory created by themake-cdrom.sh
script:python escape_bibtex_titles.py books
-
Re-generate the
cdrom
with the updated .bib files:touch books make cdrom
NOTE: These changes will be overwritten if you re-generate and re-download any books from START. You must re-run this script each time you re-download.
The START scripts generate the PDF proceedings, but they do not generate the conference handbook (a.k.a. booklet) that is printed and handed out to attendees of the conference. This procedure is not fully automatic, but there are several scripts and LaTeX templates that should help you.
-
Specify various metadata not available via the START download, by modifying
settings.py
. You will need to specify short names for each day of the conference, workshop titles and locations, locations for each session on each day, and session chairs. -
Generate a partial draft of the handbook from the main conference papers:
python generate_handbook_tex.py books/<abbrev> handbook/handbook.tex
where
<abbrev>
is theAbbreviation
used for the main conference on theCDROM
tab ofACLPUB
in the START system. For example, for EMNLP 2013, the command looked like:python generate_handbook_tex.py books/EMNLP handbook/handbook.tex
This will generate, for each day of the conference, an overview page showing the schedule, and a list of talk abstracts for each session in the conference. Note that the generated
handbook.tex
depends on the macros defined inpreamble.tex
, which is also in thehandbook
directory. -
Generate talk abstracts for each of the workshops:
python generate_workshop_tex.py books/<abbrev> handbook/handbook-<abbrev>.tex
Again,
<abbrev>
is theAbbreviation
that the workshop used on theCDROM
tab ofACLPUB
in the START system. This will generate a filehandbook/handbook-<abbrev>.tex
that contains a single session for the workshop, containing all the workshop talk abstracts. -
Manually add all remaining material. This will typically include at least:
- Filling in
handbook/frontmatter.tex
with pages for:- Conference title and logo
- Welcome message from the general chair
- Organizing committee
- Conference sponsors
- Adding a schedule for the workshop day(s), followed by
\input
ing thehandbook-<abbrev>.tex
files you created above. - Adding sessions and abstracts for keynote speakers.
- Adding talk abstracts for TACL papers (which are typically not on START).
- Adding a conference venue map.
- Filling in