Re: ALPSP statement on e-publishing.

[Eric Hellman <eric@openly.com>]

Peter's argument about the University of Chicago Press experience is
incomplete without noting that the University of Chicago Press's
production process has always been oriented towards very high-quality,
carefully redacted titles. In other words, the electronic-first process
saved money on a process that started out expensive.

Many other publishers have found that in order to go electronic, they have
had to clean up a lot of messes that the print-first process was letting
them ignore. So, in fact they are not lying when they say electronic costs
them more.

I have previously observed that journal publishing is now a technology
business, and like all other technology businesses will experience an
exponential decline in cost of production. The chip industry has been on a
curve of 100% improvement every 18 months; for the steel industry the rate
has been much smaller, but constant nonetheless.

My estimates are that per-article cost of production in journal publishing
will decline by about 10% per year for the forseeable future. Producers
that exceed this rate will do well, producers that lag will not do well.

Eric

At 1:13 PM -0400 4/28/02, Peter B. Boyce wrote:
>Chuck raises a very valid question about the real costs of electronic
>publishing.
>
>During the last decade, the University of Chicago Press made a substantial
>effort to retool the publishing process used for the American Astronomical
>Society journals. By going to an electronic-first concept -- meaning that
>the articles were converted to archival quality SGML as soon as they were
>received -- and deriving the paper and electronic screen versions
>automatically from the SGML, we have made significant savings in the
>publishing process. The AAS now produces both electronic and paper
>versions at a cost less than was needed to produce the paper version alone
>ten years ago.
>
>It took a significant effort and a substantial amount of time to achieve
>this, but we are publishing faster, better, and cheaper than we used to.
>But not all journals will fall into this mode. The AAS journals are
>expensive journals to produce -- that is, they contain significant amounts
>of complex math and they are more heavily copy edited than many other
>journals. Under these circumstances, and with authors submitting
>near-perfect manuscripts using our well defined LaTeX macro package, it is
>possible to achieve cost savings by going electronic.
>
>Yet, I see publisher after publisher claiming that adding an electronic
>version costs them more. In many cases I am skeptical, and would hope that
>by retooling their time-honored processes they could add the electronic
>version far much less than they are charging.
>
>The other benefit of going to archival quality SGML for storage of the
>articles comes in maintaining and providing access to the back issues. In
>our case, the screen version is automatically derived from the SGML at
>very little cost. This means that making technological updates to the
>older articles becomes very simple. Moreover, the cost of storage and
>compute power is dropping much faster than the volume of pages to be
>served is increasing. So, in sum, and even including the increasing number
>of older pages, the overall costs to serve a journal to subscribers should
>be decreasing. Maintaining and serving the past issues is a minor
>extension to the work of serving the current issues. Chuck is right in
>this.
>
>The underlying assumption I am making here is that the journal is designed
>for easy maintenance, using a good, robust version of HTML (or perhaps
>XML). If a publisher does that, and uses automated tools and translation
>programs, then maintaining the complete backlist should have very little
>cost impact.
>
>Yet I do not see either of these outcomes happening. Publishers in general
>are charging additional costs for producing electronic versions of their
>journals which range from small to exorbitant increases. And, they are
>claiming that the cost of maintaining an electronic archive is overly
>burdensome. Going by our experience, I find it difficult to believe either
>of these claims. We, for one, are paying for the long term preservation
>out of current revenues (probably less than 1% of current revenues). Yes,
>serving electronic journals takes some effort, but once you are doing it
>for your current journals, adding the long term preservation and access is
>really trivial.
>
>I don't think electronic journals in the sciences should cost more than
>their paper counterparts used to cost ten years ago.
>
>--Peter Boyce--