by
(Posted September 5, 1997 ? Issue 15; archived September 19, 1997)
The World Wide Web is only a couple of years old, but because of the massive amount of hype surrounding it, we have come to expect miracles of this technology in its infancy. I confess that I too am prey to this. Sometimes it seems that waiting for the Web to begin fulfilling its promise is like waiting for Godot. (What would Beckett have made of this extraordinary, disembodied, community-building apparatus?)
But if there is one discrete sociological tribe for whom the promise of the Web is more than mere promise, it is life scientists. The irony of this is that many life scientists have never experienced any of the "wonders" the Web is beginning to offer. Why? Because unlike surfers, they think life begins and ends at their benches - or possibly they are too busy desperately filling out grant applications to be browsing Web sites.
In only six months, Science Online has signed up more paying subscribers (18,000 and counting) than most print journals have garnered after decades of effort. And yet I continue to encounter scores of scientists who have yet to visit our site - or have visited us merely to download a particular paper. What they don't know is that Science (and many of the other high-impact life science journals) are now mounting an array of online services that many scholars have only dreamed of.
So it seemed eminently sensible for HMS Beagle to invite me to highlight some of the more significant advances in scientific communication that were coming courtesy of the Web. That is, it seemed sensible until I thought about my audience: Beagle eagles. By definition, you folks must be Web surfisticates. Most of you probably know more about Web offerings than I do. So just to be sure I don't waste your time, I have constructed a pop quiz. If you get it right, don't spend another Web moment with me: hyperlink out of here for greater enlightenment.
Okay, here's the quiz:
Proposition:
Hyperlinking from references in one journal to papers in another will one day be possible.
Your options:
If you said anything but #4, browse this piece; if you picked #4, get back to your grant writing.
Now that I've thinned the herd, here are some of the major and minor advances that I expect to excite most scientists.
Let's start with references. As a life scientist who cares enough about the Web to be perusing Beagle, you're probably a subscriber to one of the high-impact life science journals that collaborate with Stanford University's HighWire Press. The list includes the Proceedings of the National Academy of Sciences (PNAS), the Journal of Biological Chemistry (JBC), the Journal of Neuroscience, Cell, and a score more (including Science). If you've visited the site of any one of these journals, have you noticed that a virtual library is quietly sprouting on your desktop?
Let's go to an example. I'm writing this essay on August 1, and in this very day's issue of Science, there's a paper entitled: "A Cytoplasmic Inhibitor of the JNK Signal Transduction Pathway" by Dickens et al. Now a week before I am likely to get my print copy, suppose I take a glance at Science Online's August 1 table of contents (and let's suppose that I care about signal transduction). In a minute or two, I come across Dickens et al. and decide to cruise it (if you don't want to read this essay and simultaneously surf Science Online, take my word for what I'm about to describe). So I follow the links from the table of contents reference to the full text and, once there, I can do something online that hasn't been possible in the 117-year history of Science in print:
Advance #1: I can link from each of the Dickens et al. references directly to the National Library of Medicine, where I can instantly read the abstract for that reference.
Advance #2: Moreover, if I find a given abstract very useful, I can put the NLM computers to work by clicking on the NLM's "Related Articles" hyperlink above the abstract I've just read. This gets me scores of references to papers culled by a computer from the library's thousands of journals, based on keywords shared among abstracts and weighting the choices by placement of keywords and even by commonality of references.
Then, of course, I can click on the references that seem most promising and order up abstracts of each of these. And I can use the "Related Article" hyperlink again until I've built up an exhaustive folder of all the literature pertaining to the particular area of signal transduction that I seem to be so interested in.
But you might be saying: "Hey, you still have to wend your way to the library to forage for full text, no?"
Advance #3: Not necessarily.
I'm looking at reference #17 in Dickens et al. This cites a paper by Raingeaud et al., published in the JBC in 1995. Instead of offering a MEDLINE link to the abstract of Raingeaud et al., I'm surprised to discover a hyperlink directly to the JBC itself. One click pulls up the abstract as posted in JBC Online. A second click gets me full text!
And here's another technical advance . . .
Advance #4: I got to this JBC Online full-text paper without having to subscribe to the JBC. Normally, you have to pay to get JBC 's full text, so what made me so special? Answer: I arrived from Science Online - a HighWire product. Most of the publishers working with HighWire have agreed to give folks like me one free (for the time being) link to full text from every abstract. I'm not authorized to wander aimlessly (or purposefully) around the JBC site, but I can get a single paper from a reference in my journal and go on from there via a JBC reference to the full text of a paper in another HighWire journal.
And there's more:
Advance #5: I've just shown you how the HighWire time machine takes me back in time to full text papers referenced in Science Online; now watch the time machine sweep forward through the literature!
I'm still in JBC Online, looking at the full text of Raingeaud et al. (which was published, you'll recall, in 1995). Since I like it so much . . . and since I've already plowed backward through its references . . . I now go to the top of the Web page where there is a phrase that reads: "This article has been cited by other articles."
The words "other articles" are a hyperlink that allows me to take a gander at papers published after Raingeaud came out - papers in the JBC and any other HighWire journal that referenced Raingeaud. That's right, every time HighWire posts a new issue for one of its client journals, computers automatically identify any papers citing, say, Raingeaud, and add these to the list of "other articles" citing this article. If you access Raingeaud today, you will find at least two 1996 PNAS papers and one paper published in the Journal of Clinical Investigation (JCI), also in 1996, each of which cited Raingeaud. And from each reference, I can access full text (free, for now) by leaving the JBC Online Web site and effortlessly moving to PNAS Online or JCI Online.
Neat; what's left to be done? Plenty. Getting back issues of the major journals online and getting laggard publishers (they shall remain nameless) to put their full text on the Web, to name a couple of major requisites. After all, you can go cross-journal surfing only with papers referencing each other, all of which have been published in the last two to three years.
And then comes the toughest challenge of all: dealing with information overload. Yes, there's always a downside to these sorts of innovations, and Web-savvy Beaglettes know that none of us has the time to follow significant numbers of these (possibly valuable but potentially endless) strings of related papers. So we're working on . . .
Advance #6: If a computer can beat a Russian grand master, why can't one spare us from having to surf the long rollers these HighWire and NLM computers are throwing up on our e-beaches? Instead, let's make their computers snag a desired fish and toss it up on our virtual beaches whenever they catch one.
Here's what I mean: Suppose you were to agree to spend a few months letting a HighWire server watch you as you trolled the hyperlinks. And suppose each time you came across something valuable, you marked it as such. Suppose you also marked every paper you thought was execrable. Perhaps the HighWire computer could develop a profile of what you like and what you hate - and perhaps it could then deliver directly to your e-mail box all articles you'll love, just as they appear in HighWire journals.
What do you think about this notion? Send feedback to me so I know if Beagle eagles want this service or something like it. I'd like to hear from you because . . .
Advance #7: We are, after all, trying to use the Web to build a community, aren't we?
Of course, we have a long way to go to develop a properly functioning
communal experience, but in the meantime I hope you agree that with
these sorts of services in your future, you'll soon have not
merely a library on your desktop, but the librarian as well. Indeed, in
a weird way, you'll be your own librarian. . . . Reminds me of a tale
from the late, great Argentine short story writer, Jorge Luis Borges. If
you haven't read it, try
"The
Library of Babel" in
Ellis Rubinstein, a reporter, writer, and editor for more than two decades, is the editor of Science, the world's largest-circulation general-science magazine.
Endlinks
Web sites mentioned in this column:
BioMedNet Library - offers access to a collection of over 90 biomedical journals and newsletters, including the Current Opinion series. Abstracts and particular columns are available for free and full-text articles can be downloaded for a small fee (usually about $1). BioMedNet membership is free.
Biochemistry Online: Journals - a comprehensive alphabetical list of biology and biochemistry journals. Legends indicate features that are available online for each journal. Each listing contains all relevant links, including the publisher and editor.
Evaluated MEDLINE - is a unique implementation of the National Library of Medicine's full MEDLINE database. Many of the most significant items have been cited and evaluated by expert reviewers in the Current Opinion journals; evaluations are displayed alongside search results. When the corresponding full-text paper for a MEDLINE record exists on BioMedNet, you can link directly in both directions - from the article to the MEDLINE record and vice versa. But the CitationWeb feature can also follow links from a MEDLINE record to all the articles on BioMedNet in which the item has been cited. This system will soon be extended to journals outside BioMedNet. Read the review of Evaluated MEDLINE on HMS Beagle.
"Dancing on the HighWire" - William Tucker's HMS Beagle profile of HighWire Press.