Note to readers: This is posted by me, Pat Morin, and represents my personal views on the continued association of SoCG and ACM.
Previous posters have summarized the history of SoCG and ACM’s relationship (long and sometimes frustrating), what will happen to the conference name (we’ll have to change it if we leave), concerns about the ACM Digital Library (it has some usability and timeliness issues), and the primary benefit of staying with the ACM (brand recognition).
None of these posts addresses what I see as the most important benefit of leaving the ACM: free and open access proceedings. I’m writing this from home right now, and accessing any SoCG paper more than one month old would cost me $15.00 for the PDF. Admittedly, this isn’t as much as Elsevier’s $31.50 for a CGTA article, Springer’s $39.95 for a DCG article, or World Scientific’s $30 for an IJCGA article. It is, however, more expensive and inconvenient than every conference with free open-access proceedings and every free open-access journal.
There are a number of good reasons to make our proceedings open-access, some of which are ideological:
- Most of the research that appears in SoCG was funded from public sources. Members of the public, including our academic institutions, should be able to access it without paying for it again.
- There are academic institutions all over the world that, for various reasons, don’t subscribe to the ACM digital library. Should researchers at these institutions be kept out of the loop about what’s happening in computational geometry?
Other reasons are more practical:
- It can be a pain to get access to a SoCG paper when travelling or even when working from home.
- It’s handy to be able to post copies of, or direct links to, papers on a course webpage. Right now posting copies is a copyright violation and links are only useful for students on campus.
- Most people in industry don’t have access to the ACM digital library. They may find that something we’re doing is relevant to their work. Knowing the details, though, requires buying an article which costs money and time (see below).
Suppose someone does spend the time and effort to buy a paper based on the title and abstract, because it appears to describe an efficient algorithm for a problem they have. If we’re honest with ourselves, we have to admit that, in many cases, they’ll be disappointed with what they find in the article: an asymptotically efficient algorithm that (a) doesn’t solve exactly the problem they want to solve, (b) is only efficient for impossibly large inputs, (c) is so complicated it’s (for them) unimplementable, and/or (d) relies on a number of other results that are also only available for a fee from ACM and other publishers.
They won’t do this too many times before writing off our work altogether. Our research would be more likely to find applications if those who might actually implement it were able to freely browse our research.
- Even if a casual reader is willing to pay $15.00 for an article, the process is excruciating. After choosing to buy an article, one has to create a web account (including a security question) and then choose to login. At this point, the web browser is no longer on the article page, so one has to find it again and choose to buy the article again. Finally, it’s possible to enter your credit card or paypal information and buy the article. (I don’t know what happens after this step.)
Contrast this with accessing a LIPIcs paper. The third result I get from googling “STACS 2013 proceedings” takes me to a page from which I can get to any paper within 3 clicks.
On the other hand, what are the benefits of continuing to publish our proceedings as closed access articles with ACM?
Jeff Erickson has already discussed the benefit of brand name recognition which can be an issue for some members of our community. ACM has already pointed out the benefit of getting to keep our conference’s name. The benefit of the value added to our papers by having them in the ACM Digital Library is questionable, at best.
That leaves the financial benefit, which Paul Beame has addressed somewhat: SIGACT gets $39K more from the ACM Digital Library than it pays the ACM. After publishing SIGACT News, $25K is left. Of this $25K, SoCG was allocated $4,500 this year. (To my knowledge, this is the first such allocation.) This $4,500 made for some nice student travel stipends this year, but be warned: Paul does say that conferences “should not count on this” in the future and that “we do expect [Digital Library] revenues to go down.”
By publishing with LIPIcs, the cost of publishing the proceedings would be about $1,200 which is, apparently, considerably less than what ACM charges, so there is some savings to be had. By leaving ACM, we would lose any contribution from SIGACT, which was $4,500 this year, or roughly $22 per attendee. Given all the other factors involved, it seems hard to predict whether leaving the ACM and publishing with LIPIcs would make things more or less expensive.
Please keep the issue of open-access in mind when casting your vote this week.