Advice to Authors of Extended Abstracts_

CampusHR

Advice to Authors of Extended Abstracts_
William Pugh
Dept_ of Computer Science and Institute for Advanced Computer Studies
Univ_ of Maryland_ College Park
March ___ ____
This article stems from discussions among the
program committee for SIGPLAN___ PLDI_ The
program committee thought it might be useful
to put together some advice for authors_ To give
some context to these suggestions_ I_ve also pro_
vided a brief description of the process by which
the conference papers were reviewed_ partially
from my perspective_ This process is similar to
the way most SIGPLAN conferences are run_ al_
though the details di_er for each conference_
How the Papers Were Reviewed_
There were ___ extended abstracts submitted to
the SIGPLAN ___ PLDI conference_ At the re_
quest of the program committee chair_ program
committee members  and their graduate stu_
dents refrained from submitting any abstracts
to the conference_ This allowed us to avoid hav_
ing to deal with direct con_ict of interests_
Each program committee member was as_
signed __ abstracts_ based on his or her areas
of expertise_ Since all abstracts were sent to all
committee members_ members could review any
abstracts they wished_ so long as they reviewed
at least the abstracts assigned to them_ Program
committee members could review the abstracts
themselves or have others review them_ although
in most cases the program committee members
at least brie_y reviewed all the abstracts they
were assigned_ even if they had colleagues review
some of them in detail for them_
_Several program committeemembers helped with the
preparation of this article_ I am particularly grateful for
the assistance of Steve Muchnick_
I wasn_t able to read any abstracts until after
the semester ended in mid_December_ and I al_
lowed myself a week o_ from reading abstracts
for Christmas_ Thus_ I had about four weeks to
read the abstracts_ and I couldn_t spend much
more than _ hours a week reading them  due
to limitations both on available time and the
amount of reviewing I could do in a day before
I su_ered burnout_ Since I read more abstracts
than I was assigned_ this came down to an aver_
age of one hour per abstract_
In reading an abstract_ I had to try to un_
derstand the work presented_ the signi_cance of
it_ and possible problems with it_ I spent at
least _____ minutes on almost every abstract_
sometimes coming back to an abstract several
times_ I spent over four hours each on several
abstracts_ In one case this was because the ab_
stract looked interesting but was badly written_
in another case_ because the abstract dealt with
a dense subject_ In several cases_ I spent several
hours on a paper simply because I had expertise
or interest in the topic described by the paper_
The program committee met for two days to
discuss the submitted abstracts and choose those
to be accepted_ A preliminary numerical rank_
ing provided by the reviews received in advance
of the meeting helped structure our discussions_
On each of several passes through all the submis_
sions_ some papers were eliminated from consid_
eration_ others were retained for further discus_
sion and some were accepted_ Finally_ we had a
total of _ accepted papers_
_
What is an Extended Abstract_
An extended abstract is not simply a long ab_
stract_ An extended abstract should contain ref_
erences_ comparisons to related work_ proofs of
key theorems and other details expected in a re_
search paper but not in an abstract_
An extended abstract is a research paper
whose ideas and signi_cance can be understood
in less than an hour_ Writing an extended ab_
stract can be more demanding than writing a
research paper_
Some things that can be omitted from an ex_
tended abstract_ future work_ details of proofs
or implementation that should seem plausible to
reviewers_ rami_cations not relevant to the key
ideas of the abstract_
Some Issues Considered by the Commit_
tee_
_ Are there any major technical _aws
in the abstract_ In a few rare cases_ re_
viewers found serious technical _aws in a
submission_
_ Is the work a signi_cant advance over
previous work in the area_ by the same
authors or others_ The abstract should
give a clear description of the advantages
o_ered by the new technique over previous
techniques_ Simply describing an interest_
ing new way of doing something that could
be done as simply and e_ciently by pre_
vious techniques won_t get an abstract ac_

cepted_ The best abstracts gave a clear de_
scription of what their results allowed that
couldn_t be done previously and why that
is signi_cant_ Examples and measurements
are great for this_
A related problem is not citing relevant
work in the area_ Don_t rely on the program
committee realizing that X_s work in this
area doesn_t apply because you are consid_
ering a slightly di_erent problem that ren_
ders X_s methods unusable_
If you have additional current papers on
topics related to your submission  accepted
by or submitted to other conferences or
journals_ be sure to discuss the contribu_
tion of your submission over that of your
other papers_
_ If the work involves a specialized ap_
plication_ does it make a more general
contribution_ Some abstracts described
interesting specialized applications_ Much
of the content of these abstracts involved de_
scriptions of the context of the work or ap_
plying standard techniques in the new con_
text_ In some cases_ it was unclear if the
resulting paper would be useful to people
not interested in the author_s speci_c appli_
cation_
If you submit an extended abstract involv_
ing a specialized application_ be sure the sig_
ni_cant contributions of your work don_t get
lost in the details of your application_
_ Does the abstract o_er an interest_
ing perspective on a problem or de_
scribe experience that might be use_
ful to others_ Several committeemembers
lamented that although several authors had
built substantial systems_ and tried several
approaches to learn which ones worked and
which ones didn_t_ the authors only wrote
abstracts about narrow technical results re_
lated to their systems_ Relevant comments
about practical experiences attempting to
apply new technologies can signi_cantly in_
crease the value of any paper_
_ Is the abstract well presented and un_
derstandable_ We didn_t reject any ab_
stracts for being poorly presented_ How_
ever_ all other things being equal_ the
program committee was more enthusiastic
about abstracts that were clear and well
presented_ www.gxqingyuan.com _ Is the abstract too long_ There are
many methods of trying to _t _ pages
of material into the __ page limit on ex_
tended abstracts  reducing margins_ using
__point type on ___point leading with dou_
ble columns_ etc_ They are all strongly
discouraged_ The page limit is to en_
courage authors to write abstracts that can
be absorbed quickly_ not to save trees_  al_
though our request for double_sided copies
of the abstracts did have this as its goal_
No abstracts were rejected purely for rea_
sons of length_ but none of the accepted ab_
stracts signi_cantly violated the spirit of the
__ page limit_ consider this a strong hint_
Several program committeemembers stated
that after reading __ pages worth of mate_
rial_ they felt free to stop reading at any
point if they were not truly excited by the
paper_
Don_t let the page count limit prevent you
from providing _gures or examples that
make the paper easier to understand_ The
page count limit should be considered an
upper bound on the number of full pages
of text_ exclusive of _gures and examples_
One program committee member disagreed
and thought that the page limit should be
strictly adhered to_ noting that if a picture
is worth ____ words_ a picture is worth more
than the __ words it displaces_
In exceptional cases_ it may be appropriate
to put additional material in an appendix
that extends past the length limit_ This is
acceptable only if the extended abstract it_
self stands on its own without the additional
material_ Given their time limitations_ most
reviewers probably will ignore the appendix_
Acceptable material for an appendix could
include background material for committee
members not familiar with the details of the
research area and details of proofs and im_
plementations omitted from the body of the
abstract_
_ Does the abstract address the obvi_
ous questions raised by the research_
For example_ if an abstract claims to de_
scribe _an e_cient_ practical algorithm_ for
something_ it should give empirical timings_
asymptotic analysis or both_ If the tech_
niques described require solving a problem
that is NP_Complete or undecidable in gen_
eral_ the abstract should discuss the di__
cultly of solving the problem_ It may be
that in practice the problems that arise in
the author_s application can be solved e__
ciently_ but if the abstract doesn_t discuss
it_ the program committee doesn_t know if
the author is even aware of the potential
problem_
The program committee was sympathetic
about not expecting data that ought to have
been very di_cult to collect_ However_ the
committee was disappointed in several in_
stances by abstracts that failed to report
data that ought to have been easy to collect
and would have answered obvious questions
about the work_
Final Comments for Authors_
_ An ideal submission should have a reviewer
intrigued within the _rst _ minutes of read_
ing_ excited within __ minutes and satis_ed
within __ minutes_ If your abstract fails any
of these tests_ it might be rejected no matter
how good the research is_ Committee mem_
bers may spend more than _____ minutes on
your abstract_ but you shouldn_t rely on it_
Before you submit an abstract_ give it to
a programming languages colleague who is
not familiar with the details of your research
or your research area and ask him or her
how much they can get out of it in less than
an hour_
_ Don_t overlook the importance of the intro_
duction_ _gures_ examples_ and conclusions
and measurements if applicable in an ex_
tended abstract_
_ Remember that some program committee
members_ of necessity_ are not experts in
your area of research and that when they
_
pick up your abstract they may have already
reviewed _ abstracts that day_ Material that
may take an expert in your area _ minutes
to go through might take some committee
members _ minutes or more_
_ There are some types of research that are
di_cult to publish in a conference simply
due to the amount of time and e_ort that
would be required for the program commit_
tee members to review the abstract prop_
erly_ If you can_t prepare an extended ab_
stract of your work that can be digested
and its signi_cance understood in an hour_
it may not be possible to accept your paper_
no matter how good the described research_
For some types of research  particularly re_
search on new topics_ it may be impossible
to meet this standard_ no matter how well
you write_ This is an unfortunate _aw in
the system_ and we have no remedy except
to suggest that you submit your paper to
a journal where more time can be taken to
referee it properly_
_ Please remember that we cannot give as
much attention to a submission as would be
given to a journal submission_ and we do
make mistakes_ If you get back comments
that suggest the program committee mis_
understood your abstract_ you can use that
information constructively_ If the program
committee misunderstood your work_ other
readers may misunderstand it as well_
_ This note has placed a lot of emphasis on
the idea that an extended abstract need to
be clearly written and easy to understand_
Of course_ whenever possible that standard
should be applied to full papers as well_
_ For additional advice_ read the excellent ar_
ticle by Mark Wegman that inspired this
report_ _What it_s like to be a POPL ref_
eree_ or How to write an extended abstract
so that it is more likely to be accepted__
SIGPLAN Notices_ Vol_ __ No_ __ May
_____ pages ______
_