The American Chemical Society's Lorrin Garson returns to the discussion with detailed comments about the significant planning and investment course the Society has already taken in moving into non-print publication. He makes the case that scaling up and sustaining production require considerable thought and infrastructura support. More numbers are introduced; Harnad differentiates esoteric publication from other sectors of the information market.
From: lrg96@acs.org (Lorrin Garson)
Subject: Publication costs (cont.)
Date: Sun, 3 Jul 94 15:57:07 EDT
Stevan,
As of yesterday I am on vacation for three weeks, and about to leave for British Columbia and Alaska. However, before going, I wanted to respond to your latest message in our exchange of thought and comments on publishing costs.
Perhaps the disparity of our cost figures is a consequence of the type of material we publish. My impression is that journals in the humanities are much simpler and would therefore be less expensive to create in the front-end process. In fact, chemistry may be the most challenging of the sciences with much information in complex tables, display math, graphics -- including chemical structures and other line art, half-tones and color. Tables, math, and artwork are labor intensive (expensive) to handle whether for print or electronic products. Also, in the sciences, there are many special characters and multi-level positioning which must be handled; we have over 500 special characters for our journals and seven levels of super- and subscripts (on line, 3 levels above and 3 levels below). These special characters must also be handled whether on paper or electronically. I must confess I don't read humanities journals and my experience in this domain is limited to undergraduate textbooks. But even with undergraduate text books, there is a marked difference in manufacturing costs because of the difference in complexity of material.
We are indeed both addressing the issue of what you call "esoteric publications," that is, scholarly journals for which authors submit manuscripts without receiving payment or royalty.
Your statement "The entire superstructure is set up to hurtle headlong toward print on paper ..." is incorrect. Since 1974 the ACS has been publishing its journals on a database structure aimed toward the day when electronic products would be created. We started preserving our journal data in an SGML-like structure long before SGML became an ISO standard. Our print products are spun-off from the database, not the other way around. I am afraid your perception of how we produce journals is quite erroneous. Approximately 80-85% of our costs are for creating this database and 15-20% for printing. The majority is for peer review, processing manuscripts (50% are now done electronically; this will probably reach 60-80% by the end of 1995), editing, copy-editing, proof-reading, etc.
Also your statement "It's natural for them [paper publishers] to do whatever they can to preserve the status quo, or something close to it." is also very inaccurate---certainly incorrect for the the ACS. Let me give you a few highlights of the ACS' electronic publishing activities:
(a) 1980: One thousand articles from the Journal of Medicinal Chemistry were loaded on BRS as the first fulltext file in chemistry, probably the first fulltext file in the sciences. This was an experimental prototype file which was tested by a few dozen volunteers.
(b) 1981: An experimental file of 16 ACS journals was loaded at BRS. The coverage was 1976 to current for the Journal of Medicinal Chemistry and 1980 to current for the other journals. The file was evaluated by about 300 individuals.
(c) 1982: The fulltext of ACS journals file at BRS became a commercial product in November.
(d) 1984: Our colleagues at Chemical Abstracts Service (CAS) established STN International in cooperation with node operators in Karlsruhe, Germany and Tokyo, Japan. This is a true network with files located at any one node accessed from Europe, North America and the Pacific Rim. Users are not aware in day-to-day searching/retrieving on which continent the files are located.
(e) 1985: We developed a prototype CD-ROM in cooperation with OCLC using the chemistry journal Inorganic Chemistry. This prototype was fulltext searchable and provided on-the-fly composition with display of our full character set, including super- and subscripts.
(f) 1986: On September 28th, the ACS made the fulltext of all its chemistry journals available on STN (the CJACS file). This file allows fulltext searching and display, but does not contain mathematics, tables or math. The file is available today and contains our journal data from 1982 to the present. The file is updated every two weeks.
(g) 1987-90: Files from John Wiley (CJWILEY file ), the Royal Society of Chemistry (CJRSC file), VCH Publishers (CJVCH), and Elsevier Science Publishers (CJELSEVIER file) were loaded on STN International. These files are still available and regularly updated. [My group processes the data for these publishers for file loading.]
(h) 1990 to date: The ACS has been involved with colleagues at Bellcore, OCLC, CAS, and Cornell University to create a prototype electronic library at Cornell University. This is called the CORE project; a non-commercial, experimental endeavor.
(i) 1993: The ACS made supplementary material for the Journal of the American Chemical Society available on an Internet server (acsinfo@acs.org). These are TIFF-Group-4-FAX compressed files available for downloading by anonymous ftp or through a Gopher interface. There are approximately 20,000 pages per year loaded on the server. The file is still available and is updated weekly.
(j) 1994-1996: The ACS is a participant in the Red Sage project at the University of California at San Francisco. Approximately 20 publishers are involved (with Springer-Verlag being the dominant publisher) along with UCSF and Bell Laboratories, to create a prototype electronic library in the fields of radiology and molecular biology.
(k) 1994: On June 19th, the ACS/CAS made electronic pages of all its chemistry journals available via STN International, thus tables, mathematics, line art and half-tones are now available by downloading via the Internet, direct dial modem or by FAX.
(l) 1994: Later this month we will ship the first CD-ROMs of two of our titles: Journal of the American Chemical Society and Biochemistry. The CD-ROMs contain fielded, full-text searching capabilities, capability to display and print journal page images, with special processing of half-tone images to accommodate non-grey scale printers, display and printing of color images, etc.
(m) By the end of this year we will have all of the graphics for our journals as separately callable objects, linked to the text, along with SGML encoded data, including tables and mathematics.
Stevan, I assure you the ACS as well as most main-line traditional, commercial publishers of scientific information are not trying to preserve the status quo but rather are very active in developing electronic information products. Other not-for-profit organizations in the sciences, notably physics, astronomy, medicine/biology and engineering, are also very active in this domain.
By the way, the ACS is a not-for-profit organization, but it is also a not-for-loss institution. The Publications and Chemical Abstracts Service Divisions are not subsidized from external sources, nor from ACS members' dues. These two divisions are charged by the ACS Board of Directors to annually return a small net to the ACS' reserves.
I would like to suggest that publishing electronic journals is in fact going to be more expensive than printing. For example, I believe most of the data we currently publish in journals today will in the future be acquired as coherent, digital data. This is starting now in the field of x-ray crystallography and will likely spread to other areas of structure such as spectroscopy (IR, UV, MS, NMR, etc.), biological data, in vitro testing, etc. The journal Protein Science (published by Cambridge University Press for the Protein Society) now publishes with each issue a floppy disk which contains protein/enzyme structure data which can be visualized with a program called Kinemage, which is also provided with the journal. The Protein Society plans to make these data also available on CD-ROM and via the Internet. The collection, maintenance (including indexing and cataloging), and dissemination of these data will, I believe, be more costly than printing, but the information will be much more valuable to the scientific community. Of course, when we get to this point we won't be publishing journals; the output will be called something else.
I am afraid you haven't convinced me to your view point and our cost figures are so diametric we can't possibly both be correct. As I mentioned in my opening, perhaps the great disparity lies in the nature of the information we publish. Have I through my verbiage above changed your perceptions of publishing and associated costs? Probably not . . .
It seems we are unlikely to resolve the issue by merely exchanging messages. Sometime when you are in Washington area, or when I am in Princeton, why don't we sit down and try to thrash this out. If on some occasion you should be in Columbus, Ohio, I would be very happy to walk you through our production facilities (data entry, database building, composition but not printing, which is done in Easton, Pennsylvania). In any case, please count on being my guest for lunch or dinner when and where we might meet.
I won't be responding to e-mail until after July 25th.
Finally, I would like to ask that you forward this message to those to whom you sent your last message. Thank you.
Best regards,
Lorrin
****************** From: Lorrin R. Garson ******************
Publications Division, American Chemical Society, Washington, D.C.
E-mail:
********************************************************
Date: Tue, 5 Jul 94 17:56:39 EDT
From: "Stevan Harnad" harnad@Princeton.EDU
Subject: Lorrin Garson (Amer. Chem. Soc.) Reply to Subversive Proposal
Dear Lorrin,
Thanks for your detailed reply about publication costs and electronic innovations at the American Chemical Society. I am very impressed by the scale of electronic innovativeness you describe taking place at the ACS.
The status quo I should have said that paper publishers would be endeavoring to preserve was the trade model itself: pay-to-see, whether on screen or on paper. You raise a valid point about technical and graphical capabilities and expenses, and you are right that my own data, from a mostly-text discipline, are insufficient to establish the generality of my < 25% per-page claim. I will accordingly allow colleagues in the more technical disciplines to bring forward their own figures in response to what you write below. My own reaction to the impressive panorama of innovations you describe (apart from admiration for what you have accomplished) would be the following:
(1) Many the graphical capabilities you describe are likely to be available on the author/researcher's end these days, as are the technical-text generating capabilities. So what authors submit for publication may be very close to the final product (and they could incorporate editing and design feedback into it in their revision). It is not at all clear that having these functions instead performed by the publisher will be either optimal technically or a justification for sticking to the pay-to-see model instead of the free-access-to-all model for esoteric publication.
(2) The coding will soon be standardized, or near standardized, so that will be provided from the author's end too (guided, of course, by feedback from editors, copy editors and production editors, to which I will return below), and hence no justification for sticking to the pay-to-see status quo.
(3) Powerful public-domain search/storage/retrieval tools are already being developed and made available to all (e.g., wais, www, etc.). So this too need no longer be something the publisher does for the author, and is again not a justification for preserving the status quo.
So what seems to remain in the calculations you describe -- assuming author's end graphics and text-processing plus archive management tools are in place for all -- is (as I suggested) the true cost of quality control: refereeing and editing (include copy-editing and design). I regret that I have to say that I continue to believe that the true cost of this essential service is well under 25% per page in all fields of science and scholarship. I will allow those who are more technically expert than I to follow up on (1) to (3).
One last point: ACS is noncommercial, but is it not worrisome that, as you describe below, it so readily makes common cause with so many others who most decidedly are not? Esoteric publishing simply does not belong in this paradigm.
Stevan Harnad
Date: Tue, 5 Jul 94 23:19:57 -0600
From: Paul Ginsparg 505-667-7353 ginsparg@qfwfq.lanl.gov
To: harnad@Princeton.EDU
Subject: Re: Lorrin Garson (Amer. Chem. Soc.) Reply to Subversive Proposal
I will allow those who are more technically
expert than I to follow up on (1) to (3).
stevan,
essentially your responses are correct, but tentative due to unfamiliarity with publishing technical material including in-line equations, graphics, etc. in physics, we've been transmitting such material without compromise over the networks for close to a decade now so i can make slightly more definitive comments below.
i've lost track of all the different lists (please forward to whichever may be relevant -- feel free to edit if necessary, have been through this many times in many forums and answers grow increasingly abrupt).
Paul Ginsparg
lg> From: lrg96@acs.org (Lorrin Garson)
lg> Subject: Publication costs (cont.)
lg> To: harnad@Princeton.EDU (Stevan Harnad at Princeton University)
lg> Date: Sun, 3 Jul 94 15:57:07 EDT
lg> Perhaps the disparity of our cost figures is a consequence of the type of
lg> material we publish. My impression is that journals in the humanities are
lg> much simpler and would therefore be less expensive to create in the
lg> front-end process. In fact, chemistry may be the most challenging of the
lg> sciences with much information in complex tables, display math, graphics-
lg> including chemical structures and other line art, half-tones and color.
i suspect physics is roughly as challenging as chemistry. who is providing all of the above material? in physics, we the authors produce the tables and graphics ourselves, and can typically integrate them into in electronic end product better than can the publishing companies on paper.
lg> Tables, math, and artwork are labor intensive (expensive) to handle
lg> whether for print or electronic products. Also, in the sciences, there are
lg> many special characters and multilevel positioning which must be
lg> handled; we have over 500 special characters for our journals and seven
lg> levels of super-and subscripts (on line, 3 levels above and 3 levels below).
why would a publisher re-typeset all submissions? in physics, the journal publications are frequently lower quality precisely because of the errors introduced in the typesetting process (it is very difficult to proofread yet again something that has already been proofread hundreds of times for our own versions; especially when many of the conventional pub co's weren't even running spellcheckers to catch their trivial errors.)
lg> These special characters must also be handled whether on paper or
lg> electronically. I must confess I don't read humanities journals and my
lg> experience in this domain is limited to undergraduate textbooks. But even
lg> with undergraduate text books, there is a marked difference in
lg> manufacturing costs because of the difference in complexity of material.
if handled properly, scientific research can be propagated electronically as easily as can non-scientific. this is not conjecture -- the e-print archives on xxx.lanl.gov have from their inception been full text with all in-line figures and equations (and the astrophysicists have begun to submit .mpeg files with on-line animation), all author-prepared, and in no case are any compromises necessary for professional research communication. as i say, the author produced material is frequently superior in quality to the ultimate print form from the publisher.
lg> Your statement "The entire superstructure is set up to hurtle headlong
lg> toward print on paper . . ." is incorrect.
lg>
lg> Also your statement "It's natural for them [paper publishers] to do
lg> whatever they can to preserve the status quo, or something close to it." is
lg> also very inaccurate---certainly incorrect for the the ACS. Let me give you
lg> a few highlights of the ACS' electronic publishing activities:
i am afraid that the litany of "achievements" below tends to support rather than refute stevan's statement. instead they are strawpeople that convey the impression of forward-looking, but remain too firmly rooted in the status quo. essentially this view of the electronic format is literally to repeat the entire process, and then after the final stage, essentially as an afterthought, take an electronic photo (i.e. bitmap) of the finished version, post it somewhere, and suggest that that constitutes vision for the future. from this myopic viewpoint, of course the electronic version appears to add to the overall expense. this just means you'll be hard-pressed to compete when someone else comes along with a better optimized and more streamlined operation.
lg> (a) 1980: ...
prior to 1984 the relevant word processing and graphics simply was not available. any info on usage patterns, cost, etc., is irrelevant. totally different medium. continuing...
lg> (f) 1986: On September 28th, the ACS made the fulltext of all its chemistry
lg> journals available on STN (the CJACS file). This file allows fulltext
lg> searching and display, but does not contain mathematics, tables or math.
lg> The file is available today and contains our journal data from 1982 to the
lg> present. The file is updated every two weeks.
lg> (g) 1987-90: Files from John Wiley (CJWILEY file ), the Royal Society of
lg> Chemistry (CJRSC file), VCH Publishers (CJVCH), and Elsevier Science
lg> Publishers (CJELSEVIER file) were loaded on STN International.
no mathematics, tables, or math. in physics, this would have been less than useless and would convince people of the superiority of paper.
lg> (h) 1990 to date: The ACS has been involved with colleagues at Bellcore,
lg> OCLC,CAS, and Cornell University to create a prototype electronic library
lg> at Cornell University. This is called the CORE project; a non-commercial,
lg> experimental endeavor
isn't this just another scan and shred project to post bitmaps of existing journals? for some reason, many journals seem unable to distinguish superficial appearance from information content and insist that they are defined by their superficial appearance. (the american physical society, for example, proposed an electronic version of its journals which retained every artifact of the paper version -- including a two column format with equations that occasionally cross between columns. [a format that many physicists have grown to despise. aps would likely be subject to a full-scale network attack if they ever ventured to post new material in such an electronic format.]
it is important to rethink the compromises embodied in the current paper format and not robotically propagate them to the electronic format. indeed when i demo-ed a bitmap server to some physics postdocs, the uniform response was incredulity ["my god, it's a picture of each page."] then laughter -- they were just not interested in the static formats promoted in general by OCLC and e.g. Bell's "rightpages". and, again, of course your costs are unaffected or increased -- everything proceeds as before with an extra step added at the end. very soon we will demand functionality (hypertext, in-line links to other resources and applications, public annotation threads, etc.) that can only be embodied in the electronic format from the start.
lg> (i) 1993:The ACS made supplementary material for the Journal of the
lg> American Chemical Society available on an Internet server
lg> (acsinfo@acs.org). These are TIFF-Group-4-FAX compressed files
lg> available for downloading by anonymous ftp or through a Gopher interface.
lg> There are approximately 20,000 pages per year loaded on the server. The
lg> file is still available and is updated weekly.
more after-the-fact bitmaps. not as useful, unfortunately, as text.
anyway, rather than continue point by point, i am just trying to emphasize how all of this substantiates the point that publishers base their cost estimates of the electronic format on an outmoded mentality, viewing it as an "add-on" to existing activities rather than as a means to alter, improve, optimize, and streamline communication of research in a fundamental manner. there is nothing fundamentally different about highly technical scientific material as compared with the humanities -- researchers across the board, once empowered to produce a final format that suits their standards, and given the means of distribution, will take full advantage. the likely outcome is to force established publishers to rethink what they're doing and concede that their cost estimates were based on the wrong analysis.
lg> Stevan, I assure you the ACS as well as most main-line traditional,
lg> commercial publishers of scientific information are not trying to preserve
lg> the status quo but rather are very active in developing electronic
lg> information products. Other not-for-profit organizations in the sciences,
lg> notably physics, astronomy, medicine/biology and engineering, are also
lg> veryactive in this domain.
i have met with a continuous stream of representatives from "main-line, traditional, commercial publishers of scientific information" over the past three years and yes, they are trying to do something, mainly stay in the ballgame somehow, but that something is not necessarily optimized for the interests of researchers, either in cost, functionality, or means of access. no idea to which "not-for-profit organizations" you refer in physics -- there, at least, i believe i know what is going on (perhaps the confusion is over what constitutes "very active" as opposed to "very productive").
lg> I would like to suggest that publishing electronic journals is in fact going
lg> to be more expensive than printing.
i would like to suggest that those institutions and organizations for whom publishing electronic journals will in fact prove more expensive than printing do not have a very bright future in store.
Paul Ginsparg
Date: Tue, 5 Jul 1994 18:54:41 -0400
From: lesk@bellcore.com (Michael E. Lesk)
Steve, Lorrin
I wonder if you both know about an article "Reader rip-off: why are books so expensive" by Tony Rothman in the New Republic for Feb. 3, 1992. He is mostly talking about trade books, and finds most of the cost in distribution. He says that a $20 book costs about $3 to produce. (The author gets $2, the publisher gets $4 for overhead, the distributor gets $3 and the bookstore gets $8). For a 20,000 copy run typesetting is not important -- it is 10% of the production cost. Paper is only slightly more, about 15% of production cost.
Unfortunately, scientific journals have already achieved his most obvious recommendation: eliminate the bookstore retail markup and go to mailorder.
But his overall point is still true- most of the money in the current system is NOT going to run presses. It's distribution and organization that is taking the money, not the production side. I think that's true for scientific journals as well.
Michael
Hi Mike,
I'm sure Rothman's right about those figures, but I think that's probably more general even than book economics and probably gets to the heart of capitalism (and middlemen. etc.).
Rather than take all of THAT on, I think the simple pertinent fact in the case of ESOTERIC (no-market) publication (which makes it different from sell-your-words trade publication) is that it is NOT a "product" from which the author does, can or expects to make money through selling it! That is something peculiar to esoteric publication, independently either of the mark-ups of trade book/magazine publishing or commerce in general: THE AUTHOR WANTS YOU TO READ THE WORK, THAT'S ALL. That motive should never have had to make common cause with an economic model in which there is a MARKET for the work, people ready to pay for it, and the author writing it in hopes of getting part of that revenue -- a model in which it is in the interests of the author as well as the publisher to interpose a price-tag between the author and the readership.
This anomaly in the special case of esoteric publishing is now in a position to be remedied in short order WITHOUT taking on either the inefficiencies of trade publishing in general, or of trade in general.
Stevan Harnad