ACM doesn't understand Open Access

February 06, 2013

The ACM recently posted an article with the slug line “The age of open access is upon us. Increasingly, the consensus of authors of research articles and their funding institutions is that the fruits of taxpayer-supported research should be freely available to the public.”

This article was exciting! Open access is a fairly hot topic right now, and the way that the ACM systematically locks up research in my field behind paywalls is frustrating and flustering. (Luckily, my university is large enough to pay for access to their publications, so if I absolutely must have a paper I can get it on campus.)

The next paragraph, though, falls flat on its face. The authors make the claim that maintaining large digital archives” for a long time “incur significant cost”. They describe an alternative payment plan to this end, proposing that authors pay for these costs.

Ultimately, the ACM article proposes four ways for them to make money in order to maintain these archives, which boil down to author payment, conference payment, or, well, author and conference payment. The people they must pay for are system administrators and web developers. And while these things cost money, they are not impossible to pay for if you abandon paywalls.

For comparison, the Internet Archive group stores over ten petabytes of data. Wikipedia has roughly twenty-nine million articles. These websites take donations and keep their doors open, partially because not all of their pages are served every month (or even every year). And those companies do fine.

As for the ACM and ACM publications, here is a list of things that I think would serve as a much better business model and that few, if any, researchers could complain about:

1. Asking for donations for the ACM Digital Library. It works for plenty of other companies, and I can’t imagine that a company as wide-reaching as the ACM would have trouble finding funding support from its members and users if it came with the promise of no more paywalls.

2. Requiring that at least one author of each paper be a full ACM member. This may seem a little like “pay to publish”, but is distinctly different. It encourages membership in the premiere CS organization among its academic members while also ensuring that work published by the ACM is truly representative of its members (because its members wrote it).

3. Offering premium services as part of the Digital Library. The digital library lacks a slew of features that I’d greatly enjoy having, including author maps (large webs with links between authors that have published together, perhaps with sizes based on publication count), content tags for papers (perhaps with a similar web, and searching on tags), pages for conferences themselves (why does the ACM not more closely tie with its conferences), and possibly even links between follow-up papers for authors (it’d be great to know that the paper I’m reading from 2008 has further results presented in 2010). These are services worth paying for, and these premium accounts could go a long way toward running the ACM Digital Library. If the ACM Publication Board insists that hosting these papers is a service worth paying for, it should be.

4. Charging a small overhead per paper published in conferences and journals in such a way that these costs may be folded into conference fees. There were 42 papers published at POPL this last year. There were at least 500 people in attendence. If each paid $20 in conference fees toward publication, that would be$10,000 toward hosting and maintaining the conference proceedings in the ACM Digital Library. And given the modern costs of webhosting, this should cover at least the next twenty years of hosting 42 PDFs.

If all of these solutions were implemented, I cannot imagine that the ACM would need to continue with its paywall behavior (or at least greatly reduce costs). Their arguments and justifications feel reminiscent of the MPAA and RIAA arguments in favor of their failing business models as we move into the digital age, and it’s a google away to know how well it’s working out for those companies.

Using LaTeX for Programming Language Semantics

LaTex is a fantastic tool for typesetting, but there seem to be a serious gapin documentation for using it to lay out programming semanti...… Continue reading

A Small ALU in Haskell, Part I

Published on April 24, 2015

The Refined Gradual Guarantee and Compilation

Published on April 15, 2015