Electronic and free access to scientific publications

I have last update this page in the beginning of 2003, and even then without spending too much attention. Much more up to date page is here

The main body of scientific knowledge is in articles that are published in refereed journals. Hence, an easy and fast access to these articles is important to effective scientific research. This becomes more and more important, as the number of published articles increases. However, there is a simple solution: make the articles available electronically.

Electronic access to articles is not only fast, it is also very versatile. It allows easy following of references, searches using various criteria, organizing private 'libraries' etc. In addition, it is much cheaper than paper publications, and allows access to casual readers. Thus it is the obvious way forward. Yet, it actually happens very slowly. Why?

Most of refereed journals are published by commercial companies. Once an article is published, the company holds the rights for the article. and can charge royalties for copying or using text from the article. Hence, it has no interest in making the article freely available to the public.

In principle, the companies can charge for accessing the articles electronically. However, The logistical problems associated with this kind of operation cause serious problems:

• In the short term, it simply means that the transition is very slow, because there is a small incentive for a publishing company to make its articles available electronically. Reasonably priced computer power was available from the mid 80's, yet even now most of publication are still not available electronically.

• A long term problem is that the result is so cumbersome, that it is almost unusable. Of the few journals that are available electronically, some are available by CD's, while other on the internet, using paid account. This means that following references is close to impossible, and getting any article is slow and expansive. Searches are, at best, limited to the few journals published by each company, and commonly only to a single journal. You cannot legally build your own 'libraries', or use your own search engines. Thus the main advantages of electronic access are eliminated.
• The financial overhead associated with charging for access is large, and is going to be paid by the scientific community.

Since it is not in the interests of the publishing companies to allow fast and cheap (or free) access to their articles, they are not going to do it. So who will? The obvious answer should be the scientific institutions (I mean national internations ones, like NIH and MRC), because these are the direct beneficiary from this transition. All that is needed is to make all publication by scientific bodies (like 'Science') available free, give financial backing to journals that become freely available, and gradually phase out support for paper journals.

Yet, the scientific institutions (notwithstanding some articles and forums here and there) do nothing. Why? The simple answer is that the decision makers (heads of research institutions) don't think it is important. So the real question is why the decision makers of the scientific establishment don't think this is important.

I don't actually know the answer to this question, but I think the main problem is that the decision makers are essentially computer illiterates. In the scientific establishment, the decision makers are senior scientists. They entered science and learn how to work in the 60's and 70's, before computers were widely available, so never got used to use them. Some of them can use a word processor and mail reader, and may even be able to write some programs, but they don't know how to use a computer as a tool for accessing information.

That is not a reason not to publish articles electronically, because the younger scientists can use computers for accessing information. So why don't the decision makers go for it anyway? There are two possible answers to this:

1. They don't realize how useful is a fast, cheap and versatile access to publications. That seems to me unlikely, because the decision makers are not that stupid.
2. Unconsciously, they don't want to give the more junior scientists an advantage they themselves would find difficult to use. That looks like a very serious charge, as it seems to blame the decision makers of hindering the advancement of science to protect their own interest. However, in reality all it means that they give a very low priority to this subject, and since they have many high priority subjects to deal with, they simply never get around to consider it.

So what is the future of computerized scientific publications? There are several possible scenarios:

1. The publishing companies will do the right thing. Since the publishing companies don't have any incentive to do that, this is not a real possibility.

2. The situation will continues to develop as it is doing now. Slowly more and more publication will become computerized, but in a restricted, fragmented and expansive manner. Science will continue to miss the benefits of electronic access, and continue to waste people time, money and paper.

3. The same as above, but after a while(20-30 years) the decision makers in the scientific establishment will be computer literates, and they will get it right. That still means large delay and a large waste in the meantime.

4. Junior scientists will take the lead, and under their pressure the publishing companies or the scientific establishment will get it right. Currently, that does not seems likely.

5. The bodies that finance science (government departments, large companies, large charities) will do it. The advantage of these organizations is that, in contrast to the scientific establishment, their decision makers are, in general, non-scientists. Thus they are less likely to resist a change. On other other hand, they are less familiar with the subject, and tend to leave it to the scientists.

An idea that is being circulated at the moment is that the solution is to do WWW publishing, with the authors paying the expanses. Financially, this idea is the same as direct support by scientific institutiosn for electronic publishing, because most the authors are supported by these instituution (the loser will be authors which are not upported by scientific institutions). However, it has the large disadvantage that it undermines the peer review process. Since in that case the income of the journal is not dependent on readership, and hence on the quality of the acrticles, it will cause proilferation of low quality journals. In the end, the scientific institutions will have to intereve, and establish the natural solution of free publishing backed by scientific institutions. The only difference will be several chaotic years until reaching the solution.

When I talk with people about this, they sometimes raise various objections:

If You want to read carefully an article, it must be on paper.

This is absolutely true (until we have computers that look and feel like paper). However, that is not a problem, because you can always send an important article to the printer. Currently, scientists normally photocopy any article they need to read more than once, which is more expansive, more time consuming, and gives worst results than even a low quality printer.

What about peer review?

This is completely orthogonal question. Electronic journals can be have a review process like paper journals, provided they are not financed by the authors. In fact, once articles are generally available electronically, it will make the review process much more effective, because it will allow reviewers to scan the reference list of an article in a reasonable amount of time.

Note, however, that this is based on the assumption that authors don't pay for publication. In models where authors do pay for publication, the peer review process will be negatively affected.

What about graphics?

Computers today can handle quite high quality graphics, and the situation is going to improve rapidly. The highest quality graphics are still a problem, but these are rarely of any scientific value. A solution for these is definitely feasible, e.g. several centers that can print high quality graphics on request. Currently, photocopiers cannot cope with high quality graphics, so these are lost anyway.

It involves a large investment.

Initially, maybe. However, the difference in running costs will compensate for this in very few years. In addition, the main benefit is the increase effectiveness of scientific research, which is hard to measure, but is much larger than the explicit expanses/saving.

It should also be noted that some starting moves can be done with very small cost. In particular, making those journals that are published by scientific bodies free has almost zero cost, because the extra cost of supporting the scientific body would be largely, if not entirely, compensated by the saving for the scientists and libraries that buy them.

========================================================

[30Jul99] Things seem to be actually moving by now. See the suggestion by the NIH. They have done a serious error by putting together the refereed part and the unrefereed part, which gives the opponents an oppurtunity to confuse the issue. Hopefully they recover. I did mail them about it, but that should not stop you from mailing them as well.

[25Sep2000] You could have thought that PubMed is a step forwards, but actually it isn't. The reason for that is that the central point of who controls the copyright hasn't changed, and it is controlled by the publishers. As a result, PubMed doesn't have any momentum, because even if it is successful as far as the scientists think, the publishers will not have any intereste in giving more free information, and they are the ones that make the decision.

As the result of that. PubMed is quite lousy thing, where in most of the cases you can at most abstracts and there are no text search facilities. In principle, it gives you "related articles" which may be useful, but you cannot follow references.

It seems to me that as it is, Pubmed is more of a hindrance than help, first because it makes it look like there is progress, where there isn't really, and secondly because of the lousiness of what you get, gives the impression that making journals electronic is not so good.

=================================
[25Sep2000] The latest Nature (12Sep2000) contains a comment by somebody that suggests that instead of having another specialist journals, researchers should start a "online communities". That may seem a suggestion in the right direction, but it misses the main point of scientific journals, which is that they are reviewed, and have respectability. An "online community", at least in the normal usage of the phrase, does not have either of these. It is the review and respected journals which is important to get online.

=================================

[5Apr2001] This text is by now somewhat out of date. By now there is quite strong pressure from scientists on journals to make their artciles available freely after some period of time. However, they still miss the main point, i.e. that the copyright should not be the hand of commercial publishers, which don't have an interest in free access. Getting the publishers to make access easier is a step forward, but very very small one, and by fighting this way the scientists waste their energy.

[12Oc2001] By now this page is really out of date. See Publich Library of Science, where they finally realized that they must publish it themselves. [ 11 AUg 2003] But they went for author-paid publication, thus undermiing the peer-review as discussed. I am not sure whether this is progress or not.

[12Oct2001] The Journal of Machine Learning Research seems to be a step in the right direction, because they leave the copyright with the authors, and explicity encourage the authors to publish it in other ways.

[12Oct2001] BioMedCentral is a commercial publishing house, but they leave the copyright with the authors, and have a start-your-own-journal page. It is not obvious where their profit comes from, maybe from advertising. I am not sure how viable this option is.

[16Feb2002] They may have finally got the point. The Budapest Open Access Initiative explicitly call for research institiutes to established ways for free publishing, completely bypassing the existing journals. The support from Soros may also be useful.

You can sign in support of the initiative, but you must enter some institution. Apparently they don't that people without affiliation count. I signed it and entered the institution as "people that don't have affiliation don't count?". They seem to have filtered it out. I now added my signature with Cambridge University as the institution.

============== related sites ====================

Proceedings of ICSU Press Workshop on Economics, real costs and benefits of electronic publishing in science - a technical study

Nice discussion, but the idea of making the publications free, and not bothering about paper edition, does not appear at all. This makes all the estimation of costs irrelevant.

Harnad's page

Many references about the issue of electronic publishing.

SPARC

Interesting enterprise, but still centered around money-making publishers.