The Astrophysics Data System Abstract and Article Services

A. Accomazzi, G. Eichhorn, C.S. Grant, S.S. Murray, M.J. Kurtz
Smithsonian Astrophysical Observatory, 60 Garden Street, Cambridge, MA 02138, USA
Author's e-mail: alberto@cfa.harvard.edu

Abstract

The NASA ADS Astrophysics Science Information and Abstract Service (ASIAS) is enhancing its presence on the WWW by incorporating additional sources of abstracts and making available online journal articles. In order to provide more complete coverage of our abstract database, we have started using several sources of astronomical references including NASA/STI, SIMBAD, author abstracts obtained directly from journals, and individual contributions. We are also planning to include abstracts from other subject categories loosely related to astronomy (such as spacecraft instrumentation and communication satellites). Our article service, currently providing full-text papers of the Astrophysical Journal Letters from 1975 to the present, will be greatly expanded to include most of the major astronomical journals, creating one of the first implementations of a digital library for astronomers.

Keywords: Abstracts, Information Systems, Astrophysics Data System, World Wide Web, Digital Libraries.

Introduction
Sources of Bibliographic Data
Bibliographic Searches
Full-text Articles
Future Plans
Acknowledgements

Introduction

To allow easy and transparent access to its databases, the Astrophysics Data System (ADS) has been concentrating its efforts in making its data services available on the World Wide Web (WWW). In particular, the ADS Abstract Service has undergone a number of major enhancements and has been integrated into the ASIAS, which now consists of a sophisticated search and retrieval system featuring access to bibliographic references, full-text journal articles, published data tables and catalogs, and object name resolution via access to the SIMBAD database.

Recent advances in the field of distributed database technology, wide-area information systems and network connectivity have allowed much of the functionality of the original (or "classic") ADS system to be ported to network services based on the WWW. Easy access to the ASIAS is now provided via simple HTML forms allowing fielded textual searches of the ADS abstract database with links to the online full-text articles. Browsing and printing of such articles is now as easy as pressing a button or selecting a hyperlink in a WWW document.

In the following sections we describe the contents of the bibliographic database in use with the ADS abstract service, the WWW searching tools and interfaces currently available, and the recent addition of full-text journal articles to the ASIAS.

Sources of Bibliographic Data

We have recently increased the number of bibliographic references in our database by including data obtained from several sources in addition to NASA's Scientific and Technical Information (STI) Database. The need to supplement this data has become increasingly necessary as the STI program has undergone organizational changes which have caused delays and gaps in the coverage of astronomical literature.

At the time of this writing, our abstract database contains approximately 210,000 unique references collected from the following data sources:

NASA/STI: The NASA database contains abstracts from over 200 journals, publications, colloquia, symposia, proceedings, and internal NASA reports since 1975. They include the majority of astronomical journals as well as several sources from journals loosely related to astronomy. To avoid copyright violations, the text of the original abstract as published by the journal is rewritten by STI. Keywords are assigned to the paper as a way to classify its contents.
SIMBAD: SIMBAD provides up-to-date references for papers published in all of the major astronomical journals, kindly made available to us by the Centre de Données astronomiques de Strasbourg (CDS). The SIMBAD references do not include abstract text.
The Astrophysical Journal: Electronic listings of papers published by the Astrophysical Journal are made available by the American Astronomical Society (AAS). These references do not include abstract text.
Journal Scans: For the journals for which we provide online full-text articles, we have scanned the journals' table of contents and applied Optical Character Recognition to the scans to generate journal references in electronic format. These references do not include abstract text.
Individual Contributions: Individual institutions or librarians willing to provide us with references in electronic format can now do so by properly formatting the bibliographic information and sending it to us via e-mail.

In the next few months, we plan to add data from the following additional sources:

Astronomy & Astrophysics abstracts.
Publications of the Astronomical Society of the Pacific abstracts.
Revista Mexicana de Astronomia y Astrofisica
Contributions of the Astronomical Observatory Skalnate Pleso

The system we have now in place enables us to keep multiple sources of abstracts referring to the same publication online. This allows the user to read both the original author abstract in addition to the one provided by NASA/STI. Furthermore, we recently started to collect notes from the authors of papers in our database. A WWW form can now be used to submit author's notes/comments regarding the contents of a paper or the online abstract information.

Bibliographic Searches

Searching the ADS bibliographic data is possible using a number of different WWW form interfaces depending upon the nature of the query. Currently the possible search methods available are Abstract Queries, List Queries, and Bibcode Queries.

Abstract Queries

The Abstract Query Form provides basic access into the ADS abstracts database. Through this interface, users can specify words and names to be searched for in separate query fields, and modify the query logic by changing the default settings on the bottom part of the form. The query fields currently include:

Author names: One author's name may be entered per line, or multiples may be listed on the same line separated by semi-colons. Authors may be entered by last name only, or by last name followed by a comma and first name initial.
SIMBAD Object Names: One astronomical object name may be entered per line, or multiples may be listed on the same line separated by semi-colons. The object name query is forwarded to the SIMBAD database which returns a list of papers from the SIMBAD bibliography regarding the object(s) specified.
NASA/STI Keywords: One keyword may be entered per line, or multiples may be listed on the same line separated by semi-colons. The keywords are standard subject indexing terms chosen by the NASA STI abstract service, such as GALAXIES, COOL STARS, or STELLAR MODELS.
Publication Date: The publication date should be entered as two integers in the form MM and YYYY (e.g. 12 and 1988). Dates go back to the beginning of the century and are current through approximately one month back. For references which had no publication month available, 0 was assigned. If no date is entered, the program will default to the full date range. If no "From" month is entered the program will enter a default value of 00. If no "To" month is entered the program will enter a default value of 12.
Abstract Text and Title: This field allows the user to enter any combination of words or sentences. The search will be done on individual words (except for some of the most common English words such as "the", "a", and "and" which are eliminated during the indexing procedure).

An abstract query is performed by typing text into the input widgets of the WWW query form and submitting it to the abstract server. When the search is completed a ranked list of papers that fulfilled the query requirements is returned. To view the full reference and abstract text for a given paper in the list, the user simply clicks on the hyperlinked bibliographic code.

The reference page for each publication retrieved has a relevance feedback form at the bottom which allows users to find similar references. This is done by generating a new query that uses the relevant text from the selected reference as query terms. This very powerful feature allows users to quickly search through the literature for papers relevant to the topic of interest.

List Queries

The Abstract List Query Form provides the capability to search the ADS abstracts database for author lists, NASA/STI keyword lists, synonym lists, and object alias lists. To run one of these queries, the user can specify one or more of the following:

Author Template: If a partial author name is entered, the list of all authors in the database which begin with that template string is returned. To search for author names containing a template, wildcard characters should be added to the right and left of the template (e.g. "*ste*"). Wildcards work at the beginning or end of strings, but not in the middle.
NASA/STI Keyword Template: As in the case of an Author list query, If a partial keyword is entered, the list of all keywords in the database which begin with that template string is returned. To search for keyword names containing a template, wildcard characters should be added to the right and left of the template (e.g. "*star*"). Wildcards work at the beginning or end of strings, but not in the middle.
Synonym Matching: This allows a user to search for synonyms of words used in the abstract text and title of references in the database. The results list will contain all synonyms for the word(s) entered.
Object Name Aliases: This allows the user to search for aliases of an object, using the SIMBAD database. The search for an object (e.g. "m31") will return all aliases for that object.

Bibcode Queries

The Abstract Bibcode Query Form allows the user to search the ADS abstracts database by partial or whole bibliographic code.

Bibliographic codes, first adopted by the SIMBAD group, are ASCII strings of characters and symbols that provide a unique identifier for publications. The codes are in the form: YYYYJJJJJVVVVMPPPPA, where YYYY is the year, JJJJJ is the abbreviation for the journal (e.g. AJ, ApJ, MNRAS, Sci, PASP, etc), VVVVV is the volume number, M is a qualifier for the publication (e.g. L for the "Letter" issue of a journal, P for "Pink Pages", etc.), PPPP is the page number, and A is the first letter of the first author's name.

A partial bibliographic code may be entered in which case the abstract server will return all bibliographic codes beginning with that string. This effectively allows the user to query by journal title. For instance, a query on "1992ApJ" will return all articles in the database which are from the Astrophysical Journal 1992 volumes. In addition, a wildcard is allowed before the journal title (e.g. "*ApJ" will return all Astrophysical Journal articles).

Table of Contents Queries

The Table of Contents (TOC) Query form allows users to retrieve the list of references published in a journal by entering either the publication date or journal volume number. Similar to a bibcode query but not as general or powerful, this form is easier to use in that it doesn't require users to know the root bibcodes for the journal they are looking up, nor the publication date of a particular volume.

The TOC query is only available for journals for which we have the full-text articles online (currently the Astrophysical Journal Letters).

Full-text Articles

The ADS intends to extend the contents of the ASIAS to include full-text articles from all the major astronomical journals. To this end, we have recently obtained permission from the American Astronomical Society and the Astronomical Society of the Pacific to put online the last 20 years of the Astrophysical Journal, the Astrophysical Journal Letters, the Astronomical Journal, the Publications of the Astronomical Society of the Pacific, Revista Mexicana de Astronomia y Astrofisica, and Contributions of the Astronomical Observatory Skalnate Pleso. We are currently negotiating with the major European and Japanese Journals to obtain permission to carry their articles.

Because none of these journals have electronic versions of the articles containing both the final text and graphics (except for the recent prototype system now in place for the Astrophysical Journal Letters), we decided to create electronic images of each journal page using a high-resolution scanner and make them available on our WWW server. The newly dubbed ADS Article Service now offers two different form interfaces to directly view an article selected by Journal name, Volume and page, or by requesting the Table of Contents of a particular Journal by either volume or publication date. When available, scanned journal articles are also directly accessible through hyperlinks from the result page of an abstract query.

Our pilot program, consisting of putting online the Astrophysical Journal Letters from January 1975 to October 1994, was received with great enthusiasm by the astronomical community on the network. The positive feedback from our users has prompted us to accelerate the project's timetable as much as possible. We now plan to have most of the Astrophysical Journal articles online by July 1995, and the Astronomical Journal by September of the same year.

Currently we provide article pages in three different formats:

a low resolution anti-aliased inlined image for online browsing. This allows users to preview the contents of the paper and read most of the textual information without having to retrieve and print higher resolution versions.
a TIFF or PostScript (tm) document at intermediate resolution that can be retrieved and viewed or printed either page by page or as a single document. Two flavors of PostScript code can be retrieved: a level-1 compliant document which should print on most printers, or a smaller and faster level-2, compressed document which will work with most of the modern printers.
a TIFF or PostScript (tm) level-2 document at high resolution which when printed will yield a hardcopy of quality superior to most copy machines.

By pushing the appropriate button on the HTML article forms, users can retrieve and either save to disk, send to a viewer, or directly print any of the online journal articles.All of these options are also readily available from the abstract service result page, where in addition to the abstract information, a menu for viewing or printing the journal article is available.

Future Plans

Having gone through a major overhaul of the Abstract Service server software, we are now concentrating on increasing the number of references in our abstract and article databases. We are also considering consolidating our services and making them more maintainable and reliable. In particular, these are the major projects that we will work on in the near future:

increasing the number of online references by including abstracts from different NASA/STI categories (such as Space Instrumentation), and creating a separate database for them. In addition, We are soliciting abstract contributions from publishers and librarians to provide better coverage of abstracted references and to add the capability of retrieving original author abstracts when available.
increasing the number of online full text journal articles for the major US astronomical journals. While the implementation of the Article Service has shown the technical feasibility and usefulness of the project, we now have to resolve the practical problems of placing about 100 Gbytes of data online and upgrading to a system with increased capabilities to cope with the expected increase of user requests.
collaborating with data providers and publishers of non-US astronomical journals to obtain permission to make their journal articles available from the ASIAS. This would make our service a working case of a digital library for astronomy with complete coverage of all the major journals.
adding a Z39.50 based front-end to the ASIAS services, and make our current WAIS abstracts database available through it.

We welcome any suggestions on how we can improve the ADS ASIAS services. Please contact us at ads@cfa.harvard.edu with your comments or questions.

Acknowledgements

This work is funded by the NASA Astrophysics Program under grant NCCW-0024.