HighWire Press Archive (preview version)

Publisher: HighWire Press
URL: http://highwire.stanford.edu:4141/
Cost: free abstracts and select full text
Tested: May 3-21, 2005

The Context

HighWire Press (HWP) hails from Stanford University, the alma mater of many of the smartest Web developers whom you hear about every day. It was 10 years ago when HWP had its first client and launched the digital archive of the Journal of Biological Chemistry (JBC). Since then it has facilitated the digital publishing of more than 850 other journals. HWP pioneered the idea of making abstracts accessible to anyone. It now has nearly 2 million free abstracts. And that's just the tip of the iceberg — the full text of about 900,000 scholarly articles are also freely available and fully searchable in the HWP archive.

Other digital facilitators followed suit, such as CatchWord, Ingenta, MetaPress and Extenza. CatchWord was acquired by Ingenta (now known as IngentaConnect). For one reason or another, none of them comes close to what HWP offers in scholarly content (as measured by the impact factors of the journals it hosts) and software capabilities. (In later columns I will review the other digital facilitators that remain active.)

The Content

HWP used to host almost exclusively science journals, but having SAGE as a client the number of social science journals grew to form an impressive sub-domain. HWP practically covers most of the disciplines in science and social sciences, as well as some academic journals in the arts and humanities. HWP volunteers information in the best way you can imagine. It shows the perfect example of informing users, even if they are only using the freebies.

Subject, journals and publishers

From the hub address you may survey the subject scope of the archive. It is obvious that medicine is the largest area, with journals from medical specialty areas, such as emergency medicine, palliative medicine and tropical medicine. Science and Biological (Life) Sciences represent the second largest group, followed by the Physical Sciences and Social Sciences.

The presence of scholarly journals in the arts and humanities is little known and is not obvious from the subject map. Still, you will find titles such as the Medieval History Journal, Studies in History, American Literary History, European History Quarterly, French History, Review of English Studies and Journal of English Linguistics. Music journals are also hosted by HWP, including Early Music, The Musical Quarterly and Music and Letters. There are also the interdisciplinary journals, such the International Journal of Music Education or the Psychology of Music. Just glancing at the journal list makes you want to browse some of them. Actually, you can browse many journals for free, although predominantly in the science and medicine journals. If you want a shortcut in order to see the journals that are entirely free (except for the most current 6 or 12 months), you can visit that list directly.

Looking at the list of publishers gives an impressive mix of society, association and university presses and commercial publishers. Some are represented only by one or two journals (such as the American Medical Informatics Association or the American Genetics Association), but those one or two journals are among the most highly ranked by the ISI impact factor. (You may disagree with the calculation of the journal impact factors, but there is no better alternative to get a quantitative measure for the most-cited 9,000 scholarly journals and serial publications.) Take, for an example, the Proceedings of the National Academy of Science of the United States of America (PNAS). It is a single source from the Academy, but its impact is universal across many sciences. HWP has about 92,000 full-text articles from PNAS and 88,600 of them are free of charge. Only the articles of the most current six months are under embargo — a very reasonable approach.

Many publishers have almost their entire stable hosted by HWP (such as the American Society of Microbiology, the British Medical Journal Publishing Group and Lippincott, Williams and Wilkins), even those that have in-house developed alternative (such as Oxford University Press) or are hosted by an aggregator (such as SAGE hosted by CSA, which handles the cited references masterfully). You will no longer find the publications from Annual Reviews, Inc., whose excellent, top-cited review series have been hosted by HWP software for years, on the list of publishers and journals. They are still graced by the HWP software, but are not part of the HWP archive, even though they could benefit the most from the functionally and visually appealing representation of the citations their chapters receive. The exquisite content is enhanced by about 15 million records from PubMed, offering joint searching of the HWP archive and PubMed.

Abstracts and full-text records

The HWP archive has 1.8 million abstracts and 1,610,000 full-text articles. Many of them are available both in HTML and PDF formats. Almost 900,000 articles are open access.

Once again, HighWire Press comes through with the facts and presents the distribution of the records, journal by journal, showing the number of total and free online articles, the ones in PDF and HTML formats — all the ones with abstracts.

The Software

The new software (still in beta) simplified the interface, especially the search template which was a bit crowded in the earlier version.

Browsing and searching

Although only the journal and publisher name indexes can be browsed, the software automatically provides clues for the most appropriate terms. The best strategy is to start with a broad query of one or two words, such as diabetes, instead of guessing in what form and syntax insulin independent diabetes mellitus should be searched. The software incorporates Vivisimo for on-the-fly clustering of the results. It shows the most common words and phrases in the result set in the company of the query words and the frequency of their occurrence. The list makes it clear that the correct spelling of the term is non-insulin-dependent-diabetes and the subject heading is Non-Insulin-Dependent Diabetes Mellitus (lowercase and uppercase formats make no difference). As these subject terms are linked, you don't need to type them — simply clicking on them will bring up the records with the selected subject terms. I preferred the earlier layout that also automatically displayed the set of index terms found in the new result set.

The advanced form offers more search options, but not as many as the earlier one, which made it possible to restrict the search to the title, tilte+abstract and title+abstract+full-text index. The new version does not offer the title+abstract search option, which is not a good idea. Why? Because using the title+abstract option can help focus a search. When a search term shows up in the abstract, it suggests that it is a major point in the article, rather than just a side comment. I actually asked HWP to make a separate option for an abstract-only search that would have guaranteed that the record have an abstract. It helps in sifting through the results and cherry picking before diving deep into the articles.

Linking and mapping citing and cited articles

The full text may be in HTML and/or PDF format. The former is my preferred option for a first reading and navigating back and forth between the text and the cited references section, then jumping to a cited reference or at least to its abstract. Only the HTML version allows that kind of flexibility, as well consulting the summary list of the articles that cite the one I just read.

It should be noted that only those citing articles are included that are from journals hosted by and present in the HWP archive. This is the case in all publishers' archives, but not in aggregators that offer links to citing articles, such as EBSCO and especially CSA, that have bi-directional links (cited and citing articles) from journals in other databases, such as the bi-directional links between the SAGE archive and Sociological Abstracts.

Even if the links are restricted to a single archive, as is the case with the HighWire Press Archive, it is a power tool in light of the hundreds of high impact journals from various publishers within the HighWire Press Archive. That's exactly the reason why I mentioned that the secession from the HWP membership of Annual Reviews is unfortunate. The citation map for the above article visualizes the genealogy of the target article, showing both its cited and citing articles visually. A click on any of the circles, representing the citing/cited articles, will bring up the article.

For fairness, Annual Review chapters freely show the number of citations received from source journals analyzed by ISI — an excellent clue about the importance of the chapter. You also can look up the list of each citing article if you have (or more likely your library has) a subscription to the Web of Science database. PNAS offers the same features — it did not choose secession from the HWP Archive — for its own benefit, the benefit of HWP member publishers and ultimately the end user.

HWP has been an outstanding software from day one. It kept adding new clients, as well as new software features. It teamed up with Vivisimo at the right time and smoothly integrates its functionality. I would like to see more sort options, downloading results in standard-compliant XML format and a few other features. But overall, HWP is a superior service showing competence and care. I wish more scholars knew about it.