[Frequently Asked Questions]
Volunteering for the Papers Program
I've been doing graphics for years. May I be on the Papers Committee?
The Papers Chair selects the committee. I'll be selecting with several goals in mind, including: coverage of areas in which I anticipate submissions, getting some "old hands" who have been on the committee before and bringing some new folks into the process, recruiting people who will work well together and treat papers with respect and enthusiasm, and getting representation from diverse communities. If you'd like to participate, send email to the Papers Chair and tell me about yourself and your areas of expertise.
I've volunteered to be on the committee for three years now, and I've never been chosen. What's up with that?
It may be that others are better qualified, that we already have Committee members with expertise in your area, that the Chairs have not felt that they can work effectively with you, that the chairs do not feel that you've been in the field long enough to be an effective Committee member, or any number of other reasons. The Committee composition does change from year to year, though. Please keep offering your services.
Just what sort of workload is involved in being on the Papers Committee?
You must review about 18 papers. For about nine papers, you must find two additional reviewers, and for the other nine you must find one additional reviewer. You must attend a Papers Committee meeting in late March 2003 from Friday morning to Saturday evening. During this time, you'll discuss papers, possibly be called on to provide additional reviews of a couple of papers, and be expected to listen carefully to a lot of discussion that has little to do with you. You may also be asked to act as a referee for a paper that's been conditionally accepted or conditionally accepted with minor changes, to verify that the final version meets the requirements set for it.
What do I get for all the work that I'll be doing as a Committee member?
In material terms, you get free registration for SIGGRAPH 2003. You also receive the recognition of your colleagues, the gratitude of authors, and the sense of satisfaction that comes from knowing you've given something back to the organization that helps disseminate research in graphics.
Can I submit after the deadline?
No. The deadline is absolute.
But I had a major life event (birth, death, divorce...) just two days ago!
The deadline is absolute. You may, of course, submit the work in its current form by the deadline, even if it's not the paper that you'd like it to be.
But my fancy color printer stopped working at 4 pm and the FedEx deadline is 7 pm!
The deadline is absolute. Equipment failures are common, and SIGGRAPH
2003 cannot adapt its schedule to accommodate them.
I submitted a paper and six VHS tapes. Unfortunately, in our rush to meet the deadline, we incorrectly set our gamma during taping, so we sent a fairly poor-quality video. I have since corrected the problem. May I substitute new tapes for the ones I submitted? The video is identical, except for the gamma correction.
No. The submission deadline is absolute; all materials must be submitted by the deadline; no material will be accepted after it.
I'm using the SIGGRAPH 2003 English review service, and they didn't get back to me soon enough, so it's SIGGRAPH's fault that my paper isn't ready. Can I submit late?
The deadline is absolute. The English review service makes no guarantees about turnaround, and it's up to you to make contingency plans.
I'm not in the US, and US Customs often holds up submissions, so I have to send it off two weeks earlier than US researchers would. Can I send it by the deadline instead, and you'll receive it about two weeks late after US Customs has had a chance to process it?
The deadline is absolute. If your paper must pass through various hurdles to get here, you must plan in advance how to submit it early enough to ensure arrival on time. Note: in a recent year, there were about 300 submissions, and only six or eight were returned for late arrival. Some of those were from the US.
I would like to submit my paper to conference X or journal Y as well as to SIGGRAPH 2003.
You must submit to just SIGGRAPH 2003 and await our response before submitting elsewhere (should your work not be accepted by SIGGRAPH 2003). If you submit your paper to another conference or journal simultaneously, we will reject your paper without review. We'll be in contact with the editors of several graphics journals, and Chairs of other graphics-related conferences, swapping information. We usually find four or five double submissions each year.
But I want my paper to be in SIGGRAPH 2003. I promise that if it's accepted by SIGGRAPH 2003, I'll withdraw it from the other conference or journal.
We appreciate your eagerness to have your paper published by SIGGRAPH 2003. There is only one route to having this happen: see ACM Copyright Form.
Where can I find a list of the Computing Reviews categories? Also, are we restricted to using keywords specified on the Web site?
See ACM's Computing Classification System, which includes a "how to classify" document that should answer all your questions.
How do I include a reference to myself without identifying myself?
Preparation for Conference Proceedings for a sample SIGGRAPH paper that shows how to include self-references. The general rule is to use the third person, so that if Fred Brooks were to write a paper, he might, in his "related work" section say: "Brooks et al.  discuss a system in which molecular visualizations are ... Our work builds on some of the ideas presented there, and on the ideas of Smith et al.  and the interaction techniques described by Wolford ." He would NOT say: "The authors, in prior work , discussed a system in which molecular visualization ..." The only case in which anonymous references are appropriate are unpublished manuscripts, in which case he might write: "The authors have also developed closely related techniques for molecular manipulation , but that work is outside the scope of this paper." Reference 15 would then read:  Anonymous Authors. Molecular manipulations through computer graphics, submitted to CACM.
I have a problem with the video material I need to submit. My work involves manipulation of video, which has already been captured in PAL format. I can use a conversion service to go to NTSC, but this will negatively affect the quality of the video material. Would it be OK to submit the video as a quicktime movie on a CD-ROM, instead of as a VHS tape?
While there is no guarantee or requirement that the reviewers will look at your CD-ROM, it is a reasonably safe bet that they will. I suggest that you submit a converted video in any case, and indicate in a note attached to the CD-ROM that you prefer that they view the CD-ROM in place of the video if at all possible.
The Submission Procedure Checklist says pages should be numbered, but the document preparation instructions say that page numbers should not be included. Which is correct?
You should number the pages. For the final version, if your paper is accepted, we'll ask that you not number the pages, but please number them for the submission. The sample documents in Document Preparation for Conference Proceedings all include page numbering.
I know I am supposed to remove my name, company name, etc. from the document, but should I also remove names from the acknowledgments? If the paper is accepted, should I send another copy to you with this additional material?
You should not include an "acknowledgments" section in the submission. If your paper is accepted, you will submit a revised version that identifies you and your co-authors, your affiliations, and any acknowledgments that are appropriate.
I am thinking about submitting a paper to SIGGRAPH 2003, and I'm uncertain about the requirements. Specifically, "Your paper cannot have been previously published in an equivalent or similar form. A paper is considered published if it has appeared in a peer-reviewed journal or in published meeting proceedings that are commercially available afterward to non-attendees, regardless of the language of that publication." Does this include abstract-reviewed published meeting proceedings (for example, SPIE proceedings)? Does your restriction apply to papers accepted for journal publication?
If the meeting proceedings are published in a form commercially available afterwards to non-attendees (as in the case of SPIE proceedings) then it is not appropriate to submit a paper that has been so published to SIGGRAPH 2003. If your work has been accepted for journal publication (or even submitted), it is not appropriate for submission to SIGGRAPH 2003.
We've submitted a paper about a pilot study to conference X, and now
we'd like to submit a paper about the full-blown user study to
SIGGRAPH 2003. How should we go about that to avoid the perception that it
is a dual submission?
Cite the submitted paper in your SIGGRAPH 2003 submission with a note to
the reviewers that either it will be accepted by conference X, or you
will publish it as a tech report and make it freely available on the
Web. Send in an anonymous version with your SIGGRAPH 2003 submission.
Then when you write the SIGGRAPH 2003 paper, treat the pilot study as
already published. Don't repeat text or figures from that paper in the
SIGGRAPH 2003 version.
I sent in a paper to workshop X with the understanding that it was
for review purposes only and the workshop would have no published
proceedings. Now, four months later they tell me that they're going to
publish the proceedings and include it in the digital library.
Unfortunately there is significant overlap between that paper and my
submitted SIGGRAPH 2003 paper. How should I handle this?
We realize that you didn't intend to do anything against the SIGGRAPH
rules, but now that the workshop rules have changed,
you should either withdraw the workshop paper from the proceedings
or withdraw your SIGGRAPH 2003 submission.
Could you tell me how I can "Print the online ID number (from the Papers Submission & Authorization Form) on each page" in a LaTeX file?
Preparation for Conference Proceedings for information on how to do this in LaTeX and Microsoft Word, and how to use these programs to produce properly formatted documents overall.
May I include black-and-white images in the submission, knowing that in case my work gets accepted they are required to be in color? (My color printer is broken.)
Yes. But the reviewers will have to make a guess at what your color pictures will look like, and that may reduce your paper's chances for selection.
Are papers merely published as conference papers, or is there a presentation as well?
There is a presentation; the length of the presentation will generally be related to the length of the paper.
Does the video submitted by 22 January have to be final quality? Or will people whose papers are accepted have the opportunity to prepare a more polished video?
You'll have the opportunity to prepare a more polished video. Of course, the better the submission video looks, the more likely reviewers will be able to see the strength of your work, so early polishing is a good investment of time and energy.
Where can I get the "ACM Copyright Form" on the Web? I need to show it to my employers before I submit.
ACM Copyright Form
My paper was just accepted to SIGGRAPH and I'm thrilled. But now my boss points out that I can't use Mickey Mouse as the example in my paper because I don't have the rights to use him. What do I do now?
The call for papers explicitly stated that you MUST have permissions for all the images in your paper and the footage on your videotape/CD/DVD at the time of submission. You should immediately contact the papers chair and show him/her what you propose to use as a replacement. If the new images or footage are not substantively similar to that submitted for review in the judgment of the Chair and his/her advisory committee, then the acceptance of your paper will be rescinded. The archival record (proceedings and paper DVD) must contain material that is equivalent to what the reviewers saw at the time of review.
I found this cool picture of Madonna on the web, and I'm planning to
use it to demonstrate the results of my system for animating dance
sequences from a single static pose. Will that pose a copyright
Yes, that will cause problems. Do not use any image in your
submitted paper that you do not have the rights to. Finding an image
on the web does not give you the rights to it nor does giving credit
(as in "figure adapted from xxx").
Do we have to prepare the paper in the final format?
Yes, please do so. Seeing a paper in final format lets us verify the page count and allows us to compare it to other papers.
I presented an initial version of my work as a sketch at SIGGRAPH 2002; the abstract was printed in the SIGGRAPH 2002 Conference Abstracts and Applications. Does that mean I can't submit it as a paper?
No a sketch does not count as a prior publication. Neither does a demo
in Emerging Technologies. Of course, if the work was submitted as a paper last year and rejected, but you were advised to submit it as a sketch, it probably won't get accepted this year unless you've made substantial progress since last year. If you have, however, you're welcome to submit a paper describing the work and the new progress. Caution: Do not submit the same abstract you submitted for your sketch, since that has been "published" before.
At SIGGRAPH 2002, I presented my work on topic X. Since then, I've done some more work on topic X. How much new stuff do I need to have done to get the paper into SIGGRAPH 2003?
It's very hard to quantify the amount of additional work that might be necessary. The committee has frequently responded negatively to incremental work, especially, for example, in the case of an eight-page paper where only two pages really present anything new. If you've written about the subject before, the new paper should tell what's new. As a simple guide, there should be no paragraphs in your submission that appear in any other paper you've published, and no images of that sort either, except in cases where you want to show how much the work has improved.
My company has a great new product that is of general interest to the SIGGRAPH community. I'd like to submit the product announcement as a paper.
Please don't. It will be rejected, since its topic does not fit any of the categories for submission.
What are the session titles for SIGGRAPH 2003?
Unlike many other conferences, the annual SIGGRAPH conference organizes the papers into sessions after the selection process. This removes the sort of bias where the committee says: "We really NEED one more paper on topic X to fill out the session, so let's accept this." This means that occasionally there are sessions at the conference that lack a coherent theme. You need not be concerned about sessions during the submission process.
We've independently developed an algorithm that produces results a
great deal like those produced by the fur plug-in to Maya. Of course,
Alias|Wavefront won't tell us the details of how their plugin works.
How do we best handle this piece of related work?
You are responsible for comparing your research results to all
published work (including tech reports, theses, and patents). For
commercial algorithms that are unpublished, you are responsible for
comparing your work to existing techniques as a black box. For
example, show how your algorithm produces better results when the
animal has a really dense coat. Or show how the input specification for
your algorithm is far simpler than that for the existing algorithm
because your algorithm has a particularly intuitive UI.
Patents and Confidentiality
What about patents and confidentiality? Are the two senior reviewers and the three tertiary reviewers under a confidentiality agreement not to disclose the contents of the paper to others? Some organizations like IEEE have all reviewers sign a confidentiality agreement. It's very important that I know for sure, since my employer may want to apply for
a patent, and it affects when I may submit the paper to the SIGGRAPH conference. Can I, for example, get a written guarantee of confidentiality?
Reviewers do not sign a confidentiality agreement. In general, there is wide respect for the confidentiality of submissions, but we cannot promise anything, or provide a written guarantee.
It would not be wise for SIGGRAPH to give you legal counsel on the matter of patents and publication; we urge you to seek independent legal advice. The main issue is that in different jurisdictions (such as Europe) prior public disclosure could invalidate a patent application. The situation is different in North America and perhaps Japan, where you have one year after public disclosure (for example, publication) to file a patent. This is entirely anecdotal, but a common situation is that people generally prepare a patent filing coincidentally with their SIGGRAPH publication.
If you are submitting a patent application on the material, we ask that you mention this in the final version of your paper, if it is accepted.
Review Process, Communications
I am submitting a paper on topic X, which I know is an area of expertise for committee member Y. Can I ask that Y be the senior reviewer of my paper?
I am submitting a paper on topic X, which I know is an area of expertise for committee member Y. Can I ask that Y not be the senior reviewer of my paper, because Y works for a competing company?
No. Indeed, Y may well be the best qualified reviewer for your work, and if so, we may ask Y to be the senior reviewer.
Am I allowed to ask for my paper to not be reviewed by someone from whom we do not expect a fair review?
No. The reviewer selection process includes no such provisions. Surprisingly often during the committee meeting there is discussion such as: "This paper got scores of 5, 4, 5, 4.5, and 2, but let me explain the score of 2. The reviewer picked at small details, was angry that his own work had not been properly cited (although when I looked at it, it appeared to have been treated more than fairly), and then wrote a very cursory review of the main contribution of the paper. It seems as if there's something going on here that doesn't have to do with the quality of the paper and we should discount this score as an outlier."
I'm not sure which category I should submit my paper in. I checked your category descriptions, and it looks like both the Research and Systems categories are possible for my submission. In the paper, I describe a novel concept that can be applied to other systems in addition to the prototype I've implemented. I also discuss previous research work.
If your focus is on the system that you have built and the lessons learned from it, then your paper would be a Systems paper. If on the other hand your focus is on a research result, then you would have a Research paper.
Hey, I know your REAL email address. Can I use that?
No. Here's why: Material sent to the Papers Committee will go not only to me, but also to the Papers Advisory Board, and will be archived as well. Things sent directly to me will be bounced to the Papers Committee anyhow, and will just make my job harder.
Isn't the committee more likely to accept papers by committee members and other insiders? How do you prevent a conflict of interest?
Any paper on which a committee member has a conflict of interest will not be discussed while that committee member is in the room. While each committee member has a list of papers and the committee members who reviewed them, these lists are customized so that the names of the members who reviewed papers on which I have a conflict of interest will not be shown on my list. In general, the acceptance rate for papers by committee members has been slightly higher than the acceptance rate for those in the overall submission pool. But the acceptance rate for these same people has also been higher in years when they were not on the committee; they're invited to be on the committee, in part, because of their expertise in the field.
With respect to related (unpublished, but submitted) papers, the SIGGRAPH 2003 Call for Participation appears to make clear that we should send one copy for the senior reviewers, and we should reference (cite) them anonymously. Is it permissible to include additional (anonymous) copies for the benefit of the reviewers, or is there some other way in which we can make copies available to reviewers who so desire (presumably without destroying the anonymity of the process)?
These are (at least) two separate issues. The first has to do with determining whether a work submitted elsewhere is sufficiently different from the work submitted to SIGGRAPH 2003. The other is whether or not one work is essential to understanding the other. If your work is in the latter situation, you are welcome to submit six copies of the related work, but clearly indicate that this is the case so that we don't mistake that work as a SIGGRAPH 2003 submission.
Reviewer #4 clearly didn't read my paper carefully enough (either
that or this reviewer doesn't know anything about the field!). How should I
respond during the rebuttal period?
We've all received SIGGRAPH reviews that made us mad, particularly
on first reading. The rebuttal period is short and doesn't allow for
the cooling-off period that authors have before they write a response
to a journal review. As a result, authors need to be particularly
careful to address only facts in the rebuttals rather than letting
their emotions show through.
Please don't say: "If reviewer #4 had just taken the time to read my
paper carefully, she would have realized that our algorithm was rotation
Instead say: "Unfortunately Section #4 must not have been as clear as we had hoped because Reviewer #4 didn't understand that our algorithm was rotation
invariant and he was therefore skeptical about the general
applicability of our approach. Here is a revised version of the second
paragraph in Section 4, which should clear up this confusion."
Now that I've read the reviews of my paper, I see much better how to
organize it so it will be clear to the reader. Can I do this
reorganization and upload the new version during the rebuttal period?
No. The rebuttal period is for addressing factual errors in the
reviews, not for getting revised text into the review process. The
committee members will have only a short time in which to read and act
on your rebuttal and it must be short and to the point.
Between January 22nd and mid-March, we've gotten some really cool
new results for our paper. Can I upload those results during the
rebuttal period? I'm sure that they will make the reviewers realize
the importance of our approach.
No. The rebuttal period is for addressing factual errors in the
reviews, not for getting new results into the review process.
Reviewer #2 says that our collision detection algorithm won't work
on concave objects. But it will, as we just demonstrated with the lid
of the teapot. Can we upload a movie showing this new result?
Yes, but the new results MUST directly address a reviewer's
comment. Time is very precious at this stage in the review process, and
you must keep your rebuttal short or the committee members will not
have time to read it.
In the rebuttal, is it appropriate for the authors to
disclose their identity. Such as "in our previous work reference "?
Yes. The rebuttals go only to the committee members
who already know the identity of the authors.
Getting a Paper Accepted
What follows is not "official" SIGGRAPH policy, but rather the Papers Chair's idiosyncratic impressions, based on several years of being on the Papers Committee.
Is it important to submit a video with my paper?
There is a very strong correlation between submitting a video and getting accepted. It's not absolute, but at the Papers Committee meeting, the papers are listed in a sorted six- or seven-page handout. The great majority of those on the first page have videos. The majority of those on the last page do not. An animation paper without a video is almost certain to be rejected.
Why do good papers get rejected from SIGGRAPH?
There's a distribution of quality in the papers submitted, and there's a threshold of quality that's high enough to get accepted. At or near the threshold, the error in classifying a paper as "good enough" or "not good enough" is relatively high. Because the SIGGRAPH threshold has always been very high, this means that some good papers get rejected. Sometimes (though not always) those same papers are accepted the following year, especially if they have been rewritten taking into account the reviewers' comments.
But what sorts of things put a paper on the wrong side of the line?
I can't answer that exactly, because it depends in part on the composition of the Papers Committee, but I can say several of the things I've heard most often as partial explanations. Failure to treat prior work fairly, or to give proper credit to those who had good ideas before you, is very dangerous. After all, those prior authors are likely to be your reviewers. Try starting with the phrase: "Building on the pioneering work of X, Y, and Z, we have found an approach to a new but related problem..." You may find that this looks wrong, but it's a far better place to start than "X, Y, and Z all tried to do this, but their solutions were inadequate in the following ways..."
It's enormously important to promise in the abstract what you will actually deliver. "In this paper, we present a general algorithm for improving rendering time for all scenes by at least 20%" is a very strong claim. If it turns out that your algorithm handles "all" scenes, as long as they only contain axis-aligned equilateral triangles, your paper will die right then and there. Conversely, claiming "this paper presents a method for modeling a new class of polygonal objects," without telling the reader what those objects are or providing examples, is too small a claim; the reader may get bored before finding out what you've REALLY done, and the paper may get sent to the wrong reviewers because it's too difficult to quickly determine the nature of the contribution. It's also important to bound your claims: in a "discussion" section of the paper, I strongly recommend that you clearly state the strengths AND limitations of your work, presenting the criticism that your best friend would give you to keep you from ruining your chances by over-representing your results.
Not only should the paper clearly define the contribution of your work,
but it should also be explicit about the parts of the system that are
not contributions themselves but instead exist only to demonstrate the
power of the rest. For example, if you've implemented a great
interface for painting fur onto models of realistic animals but haven't
improved on existing fur rendering algorithms then you need to make it
clear that the contribution of your paper is the interface while the
fur algorithm is just good enough to demonstrate the power of the
Wordiness, sloppy writing, and careless copy editing are not appreciated. Almost every committee member (and almost every reviewer) has worked hard at some point to get his or her SIGGRAPH paper to a minimal number of pages, with clear and pithy explanations. A disorganized and rambling paper says to the reviewer, "I know many people will spend a lot of time on this, but my time is more important, so I'm not going to bother to make their lives easier."
And don't disclose your identity. SIGGRAPH papers are supposed to be written to disguise the author's identity (and, to the degree possible, his or her institution). A photo that includes the author's face, references to the author's work without appropriate third-person citation, or mention of the author's institution -- all of these serve to disclose the author's identity.
What else leads to troubles?
If a paper cannot be replicated by the members of the Papers Committee (if they say: "This looks really impressive, but I simply cannot figure out how to do most of it from what's written here."), it will probably be rejected. Papers that feel as if they are simply a working-out of the next logical step from a previous paper are also often rejected.
Is there anything positive you can tell me?
Yes. If you've written a paper about some topic, it's always nice to show that the method you've developed is somehow useful for more than the initial goal. Perhaps your fur-rendering algorithm can also be used to render fields of grass, or maybe your new soft-shadow-buffer architecture can also be tricked into doing motion blur over multiple frames. I personally always find such lagniappe very appealing. Also, papers that break new ground and do something that others never thought of trying at all -- even if the initial results are not particularly promising -- tend to appeal to reviewers. There are, of course, curmudgeonly reviewers who say: "This isn't what I'm used to, so it shouldn't be in SIGGRAPH." But they are, at least sometimes, outvoted.
Make sure that all the statements you write are either true, or are clearly indicated as opinion, as in: "We believe that this method can be extended to scenes with millions of polygons without substantial increase in rendering time, but have not yet had the opportunity to verify this." If you need to support your work with mathematical theorems, be certain that the theorems are true. (This sounds silly, but lots of not-quite-right mathematics suffices to make pictures good enough to fool an optimistic author into believing the math is correct.)