perlman@cis.ohio-state.edu (Gary Perlman) (06/15/91)
The HCI Bibliography Project
Gary Perlman
Department of Computer and Information Science
The Ohio State University
Columbus, Ohio 43210 USA
614-292-2566 Office, 614-292-9021 Fax
perlman@cis.ohio-state.edu
ABSTRACT
The HCI Bibliography is a free-access online extended bibliography
on Human-Computer Interaction. The basic goal of the project is to
put an online bibliography for most of HCI on the screens of all
researchers and developers in the field through anonymous ftp
access, mail servers, and Mac and DOS floppy disks. Through the
efforts of volunteers, the bibliography has passed 1500 entries,
with abstracts and/or tables of contents; eventually, citation
information and hypertext access will be added.
Keywords: Human-computer interaction, Bibliographic information, Research
Aids, Distributed group work, Networks, Cooperative Research,
Information sharing, Hypertext, Information retrieval
1. Introduction
1.1. Goals of the HCI Bibliography Project: The HCI Bibliography Project is
an initiative to provide unrestricted access to online versions of extended
bibliographic information on the subject of Human-Computer Interaction. The
result of the project will be a database of several thousand bibliographic
records, freely accessible to researchers around the world by anonymous ftp,
electronic mail servers, and as a last resort, on Macintosh and DOS floppy
disks. Longer-term goals include developing more effective ways of accessing
information in bibliographies.
1.2. Philosophy of the HCI Bibliography Project: The primary reason for
compiling the HCI Bibliography is to improve the accessibility of information
used as a resource for research and development. Online access is a major
step in that direction because it allows keywords searching in selected fields
of records. Full-text versions of records (i.e., abstracts and tables of
contents) provide auxiliary targets for search, and additional information to
help a researcher decide if the original source (e.g., journal, book,
proceedings) should be approached. Although some abstracted information is
available from some commercial sources, the expense of access and the low
quality of user interfaces make them unattractive to many researchers,
particularly students. One unfortunate result is the duplication of effort by
researchers to create their own online bibliographies. The HCI Bibliography
addresses those issues by being non-commercial, providing free-access to its
information bases, and by performing basic research to produce better tools
for accessing extended bibliographic information.
1.3. Coordination of Cooperation: The HCI Bibliography Project is designed to
be a cooperative venture to coordinate the efforts of dozens or even hundreds
of people to benefit hundreds or even thousands of others interested in HCI.
2. Contents of the HCI Bibliography
The bibliography will contain most publications on HCI dating back to 1980;
for publications before 1980, we will be more selective. The publications to
be compiled include major journals in HCI, major conference proceedings, and
selected publications from other sources. At first, we sought donations of
any bibliographic information, but a large collection of multiple-format
overlapping records presented problems. Our current strategy is to gather
complete sets from published units like books, journals, and conference
proceedings.
2.1. Structure of Records: Extended bibliographic records consist of four
categories of information:
key information identifiers, data entry/modification dates, ...
basic information author, title, date, ...
summary information keywords, abstract, (table of) contents, annotations, ...
link information references, citations, containments, republication, ...
In addition to common summary information such as keywords and abstracts, many
records will include tables of contents of section headings. Another
extension that will allow interesting hypertext access is that records will
(eventually) include a section containing the identifiers of referenced
publications.
2.2. Current/Planned Coverage:
Index of the HCI Bibliography Project (6/14/91)
Special Files:
index the current contents of the HCI Bibliography
confer.bib extended information about conference proceedings;
includes order numbers, editors, number of entries/pages, etc.
publish.txt extended information about copyrights, with ordering details;
this file should be read carefully by all HCI Bibliography users!
Key to STATE Information in the Following Tables:
B=Basic, A=Abstracted, C=Table of Contents, V=Validated, R=Released
(lower-case means that only some are included/completed)
* abstracts and/or contents release permission from publisher is pending
Conference Proceedings:
FILE SIZE CNT STATE VER DESCRIPTION
CHI82.bib 25867 76 BAVR 1.4 Gaithersburg Conference
CHI83.bib 57759 58 BAVR 1.9 ACM CHI'83, Boston, MA
CHI85.bib 39091 41 BAVR 1.9 ACM CHI'85, San Francisco, CA
CHI86.bib 53219 59 BAVR 1.9 ACM CHI'86, Boston, MA
CHI87.bib 62996 59 BAVR 1.4 ACM CHI+GI'87, Toronto, Canada
CHI88.bib 50812 51 BAVR 1.10 ACM CHI'88, Washington, DC
CHI89.bib 68447 76 BAVR 1.6 ACM CHI'89, Austin, TX
CHI90.bib 64103 72 BAVR 1.7 ACM CHI'90, Seattle, WA
CHI91.bib 85947 130 BAVR 1.5 ACM CHI'91, New Orleans, LA
video.bib 12429 21 BCR 1.3 CHI' Video Reviews/Demos, UMd UIS videos
INTACT84.bib 35527 153 BaVR 1.3 INTERACT'84, London, England
INTACT87.bib - 164 B --- INTERACT'87, Stuttgart, Germany
INTACT90.bib - --- --- INTERACT'90, Cambridge, England
CSCW86.bib - 33 BAv --- Computer-Supported Cooperative Work
CSCW88.bib 26479 32 BAVR 1.3 Computer-Supported Cooperative Work
CSCW90.bib 27499 32 BAVR 1.5 Computer-Supported Cooperative Work
HYPER87.bib 28602 29 BAVR 1.5 ACM Hypertext'87, Chapel Hill, NC
HYPER89.bib 25080 37 BAVR 1.7 ACM Hypertext'89, Pittsburgh, PA
ECHT90.bib 8662 31 BAVR* 1.3 European Conf. on Hypertext, Paris
UIST88.bib - 21 --- UI Software & Technology, Banff, Canada
UIST89.bib - -- --- UI Software & Technology, Williamsburg, VA
UIST90.bib 19179 19 BAVR 1.3 UI Software & Technology 90, Snowbird, UT
Journals:
FILE SIZE CNT STATE VER DESCRIPTION
HCI.bib 95122 66 BAcVR 1.8 Human-Computer Interaction
IJHCI.bib 7927 35 BaVR 1.1 Int'l Jl. of Human-Computer Interaction
TOIS.bib 33501 161 BVR 1.3 ACM Trans. Information Systems
IWC.bib 43119 38 BAVR 1.1 Interacting with Computers
Books:
FILE SIZE CNT STATE VER DESCRIPTION
books.bib 55967 137 BcR 1.5 Books and Reports on HCI
helander.bib 7911 53 BaCVR* 1.6 Helander's Handbook of HCI
salvendy.bib 16180 69 BCVR* 1.3 Salvendy's Handbook of Human Factors
baecker.bib 11731 60 BvR 1.1 Baecker & Buxton's Readings in HCI
2.3. Planned Coverage: Some summary information for materials will not be
available until permissions are obtained from publishers (e.g., the Handbooks
edited by Helander and Salvendy). We expect that by the end of 1991 that the
bibliography will grow to about 3000 extended records, levelling off after
another year to about 5000, with 500-1000 records being added each year. We
would welcome suggestions about what sources should be included in the
bibliography, with priorities of what should be done in the short term. Some
current priorities include:
(all) Hypertext and CSCW Conference Proceedings
(all) IFIP INTERACT and HCI Conference Proceedings
(all) Interacting with Computers
(all) Human-Computer Interaction (journal)
(all) International Journal of Human-Computer Interaction
(much) International Journal of Man-Machine Studies
(all?) Behaviour and Information Technology
(all?) ACM SIGCHI Bulletin
(all?) ACM SIGLINK and SIGCSCW Newsletters
(all) ACM Transactions on Information Systems
(some) ACM SIGOIS Newsletter, OIS Conference
(some) ACM SIGGRAPH Newsletter, Conference, & TOGS
(some) AAAI, IJCAI, AI Journal
(some) BYTE & other Magazines
(some) CACM
(some) Cognitive Science
(some) Ergonomics
(some) Human Factors Journal and Conference Proceedings
(some) IEEE Computer and IEEE Software
(some) IEEE Systems, Man and Cybernetics - Transactions and Conference
(few) Journal of Applied Psychology
(few) Behavior Research Methods, Instruments, and Computers
We would welcome publishers to provide us with this and other information and
for volunteers to provide large chunks of these resources. Potential donors
should check with the HCI Bibliography project before making a large time
investment; a chunk may be in construction by another party.
3. Storage/Interchange Format of the Bibliography
Choosing a format for the bibliography has been a problem, with UNIX refer
format being selected because of its widespread use, its simple format that
eases exchange, and the ability of programs like EndNote to import it.
Eventually, we expect that the master copy of the bibliography will be in
SGML, the Standard Generalized Markup Language, because it can represent all
the types of information we need, and because we will provide tools to convert
it (as best as can be done) to other formats, such as Scribe, TeX, troff,
ProCite, and refer. We also plan to provide first a basic set of software
tools for searching and viewing bibliographic records, but the release date
has not been set.
3.1. Standard and Extended Refer Codes: The refer format is limited to
single-character field descriptors. This means that some of the field
descriptors are non-mnemonic, and that realistically, we are limited to 26
field names, even though refer programs are case-sensitive. There are several
fields that serve no general bibliographic purpose, but are used for
computational purposes or for obscure information: %F %G %H %M, and %U is not
mentioned in the documentation for addbib, a program for `interactive'
addition of entries. Extended bibliographic entries with single-character
field-names require unused or little-used fields to represent extended
information. Re-allocating fields is at best controversial and still non-
mnemonic, and at worst may cause problems for some users of existing refer
tools. In the worst case, it is always possible to rename problematic fields
to make programs ignore them.
The representation of formatting is another problem for the HCI Bibliography.
The refer manual suggests using troff codes, but this would create search
problems for all users and display problems for non-troff users. Formatting
information such as filling, bold, italic, underline, subscript, superscript,
and paragraph indentation are all problems that should be addressed, but there
are no conventions. As is discussed elsewhere, the eventual representation of
the materials in SGML will allow much greater accommodation of different
formats/formatters.
The following table describes the standard codes used by the refer system and
extensions used in the HCI Bibliography. The description of the refer code is
based on UNIX documentation, while the comments discuss current and future
conventions for the HCI Bibliography. If a comment begins with an equal sign,
then it means that the HCI Bibliography is redefining the meaning of that
field. If a comment begins with a plus sign, then it means that the HCI
Bibliography is augmenting the meaning of that field.
CODE DESCRIPTION OF REFER CODE COMMENTS
%A Author's name pointer to author record
%B Book containing article referenced %T of edited book, proceedings
%C City (place of publication) could be place of presentation
%D Date of publication
%E Editor of book containing article one preformatted line in refer
%F Footnote number or label (computed) USELESS but must be kept clear
%G Government order number =HCIBib ISBN or other number
%H Header commentary, printed before ref. USELESS?
%I Issuer (publisher) pointer to pub. record, imply %C
%J Journal containing article pointer to journal record
%K Keywords to use in locating reference init. upper-case, comma separated
%L Label field used by refer -k should be kept clear
%M Bell Labs Memorandum (undefined) =HCIBib modification/id
%N Number within volume +HCIBib chapt/sect/rept #, # ents.
%O Other commentary, printed after ref. the last resort
%P Page number(s) +HCIBib book/conf/rept # pages
%Q Corporate or Foreign Author unreversed author name
%R Report, paper, thesis (unpublished) %T of report, thesis
%S Series title +HCIBib book/jour/conf sect. title
%T Title of article or book mixed case title format
%U not listed in addbib manual entry =HCIBib user annotations, separate
%V Volume number with %N
%X Abstract filled text, matching line-by-line
%Y ignored by refer =HCIBib Contents, unfilled
%Z ignored by refer =HCIBib references (ids)
%* copyright accompanies extended fields
%$ purchase price =HCIBib member/other cost
%^ contained parts or containing doc book of papers, inherited fields
4. Compilation of the Bibliography
4.1. Donations of Bibliographic Materials: The HCI Bibliography project would
greatly appreciate the contribution of large chunks of HCI bibliographies,
with or without abstracts and/or tables of contents. We are still not sure
what to do with individuals' collections, but we may eventually be able to
make good use of them. Bibliographies on specific topics (e.g., menus,
hypertext) will be representable as a list of identifiers, with which it will
be possible to merge separately stored annotations. Ideally, a group of
people at a site would volunteer to provide a complete set of records for a
journal or series of conferences. If ten to twenty sites (e.g., with graduate
students) each took on one journal, it would be easy to keep the bibliography
up to date. Potential contributors should write to hcibib@cis.ohio-state.edu.
The first line of the message should be:
Contribution:
4.2. Quality Assurance of Records: To try to provide accurate records, the
HCI Bibliography project puts each collection of records (e.g., proceedings
from a conference, or several volumes of a journal) through a data validation
process. A series of scripts look for easily-detected formatting and spelling
errors. Volunteers, usually responding to requests over networks, are given
validation instructions and collections to validate over a period of a few
weeks. To estimate the quality of data validation, a set of `bugs' are placed
in the records to see what percentage of such known problems are detected.
Results suggest that it is not possible to remove all problems from such
collections, but the number of errors can be reduced to one in every 5-10
abstracted records. Typical missed errors include commas, incorrect plurals,
and unfortunately, the occasional missing word, like ``not,'' but fortunately,
such omissions are unlikely data entry errors. Any errors in records should
be brought to the attention of the HCI Bibliography Project for correction of
master files. Send corrections via electronic mail to hcibib@cis.ohio-
state.edu. The first line of the message should be:
Correction:
followed by the corrected entry.
4.3. Permissions from Publishers: We are sensitive to the rights and needs of
publishers, and we intend that materials will be included in the HCI
Bibliography only if their publishers (i.e., copyright holders) permit it.
Certain publishers automatically grant free non-commercial use of even
extended bibliographic information (e.g., IEEE and ACM, publishers of journals
and conference proceedings). Other publishers have graciously granted
royalty-free non-commercial use of their copyrighted material:
Lawrence Erlbaum Associates
Human-Computer Interaction (journal)
Ablex Publishing
International Journal of Human-Computer Interaction
Cognitive Science (journal)
Human Factors Society
Human Factors (journal)
proceedings of the annual HFS meetings
Butterworth & Co. Publishers Ltd.
Interacting with Computers (journal)
Taylor & Francis Ltd.
Behaviour and Information Technology (journal)
Ergonomics (journal)
The contents of the HCI Bibliography can be copied and distributed free of
charge, and only copied and distributed free of charge.
We request that publishers grant the HCI Bibliography Project the royalty-free
non-exclusive license to distribute and provide free-access to, on a non-
commercial basis, extended bibliographic entries of their copyrighted
publications. In return for this license, the HCI Bibliography Project will:
+ include a copyright statement with each record;
+ include a notice that any republication or commercial use of the copyrighted
material requires written permission from the publisher;
+ provide detailed information about the source of each record, including
publisher address and purchase or subscription cost;
+ provide to the publisher online versions of records of its publications,
with no restrictions on any commercial uses.
Although the HCI Bibliography project cannot pay fees for these rights, we
think that distributing this information will promote the purchase of all
included works. We have heard of some research that supported this
conclusion, but ironically we have been unable to locate a published source.
We hope that the free dissemination of online bibliographic information will
increase the subscription or purchase and use of the cited materials. On the
other hand, exclusion from a resource that we think will be used by thousands
of researchers and developers might diminish the popularity of published
works, even if they are of the highest quality and importance.
5. Procedures for Accessing the HCI Bibliography
5.1. Registration: To allow us to inform users of additions and corrections,
and to keep track of who uses the HCI Bibliography, we request that users
register with us. Send electronic mail and/or postal address to
hcibib@cis.ohio-state.edu. The first line of the message should be:
Register:
followed by a line with electronic mail address, followed by lines with other
identifying information (e.g., a mailing label, in electronic form).
5.2. Internet/Anonymous FTP Access: To access files in the HCI Bibliography,
internet users can use FTP (file-transfer-protocol) to copy files and programs
to their machines. See the file named README for more information about the
contents of files and for information about how to contact publishers of the
works. To log in to the archive machine, use the login name: "anonymous" and
provide your internet account name as your password. The messages provided by
ftp are unbelievably cryptic; many users can not distinguish between positive
feedback and messages about unrecoverable errors, so ignore them and plod
along. In the following example, where much of the output from ftp is left
out, the following conventions are used.
"$" is your system's prompt
text after "#" is a comment
you type in text after the ":" and the ">"
"ftp>" is the prompt from the file transfer program
File transfer rates will vary, so try something small first.
$ mkdir mybib # make a directory where you will keep your copy
$ cd mybib # set transfer directory before ftp
$ ftp ftp.cis.ohio-state.edu # to reach our archive; obscure text will follow
Name (...): anonymous # user logs in with standard anon ftp name
Password: yourname@yoursite # use your name and your site for identification
ftp> cd pub/hcibib # go to HCI Bibliography directory
ftp> dir # to get a listing of what's there
ftp> get README # to retrieve a file
ftp> prompt # to toggle (turn off) interactive prompting
ftp> mget * # interactively retrieve all files
ftp> quit # to leave ftp when done
5.3. Electronic Mail Server: During the first few months of 1991, access via
a mail server will be added. To use this mail server, users will send special
formatted messages to a special mail alias that will automatically generate
responses. Send your first request to hcibib@cis.ohio-state.edu. The first
line of the message should be:
Send: index
Other requests will be of a similar form, with ``index'' replaced by other
file names.
5.4. Floppy Disk: Users who do not have internet/ftp or electronic mail
access to the bibliography can request Macintosh or DOS floppy formats.
Transferring files and making disks is tedious work, so such requests may only
be filled once a month. The requester will send us (1) a prepaid self-
addressed (international) mailer and (2) enough formatted Mac or DOS floppy
disks to hold the bibliography, which was over one megabyte at the end of
April of 1991. We will be able to accept any format of DOS diskettes. This
setup will allow us to provide the bibliography without charge, while
minimizing our own effort.
6. Acknowledgements
This is a volunteer project, so there are many people to thank. It is
interesting that most of the volunteers have never met face to face. We hope
that by having an extensive acknowledgements section, users of the
bibliography appreciate their efforts. We hope that we have not left out any
people who have contributed, and we apologize to any that we may have
overlooked. We hope that by acknowledging contributions, we will encourage
others to volunteer as a provider or validator, particularly to work on the
priority areas listed above.
6.1. Contributions: The following have contributed large amounts of data,
either by typing it, or by independent efforts using unknown methods. Some of
the contributions have not yet been put to good use, but we hope that they
eventually will be integrated. If you want to contribute data to the project,
send an exploratory message to hcibib@cis.ohio-state.edu. The first line of
the message should be:
Contribution:
Most of the HCI Bibliography has been entered by work-study students,
including Nghia Pham, Matthew Erickson, and Hoa Truong at OSU. Other
information was compiled for the annotated bibliography of the Software
Engineering Institute's curriculum module on User Interface Development; it
was the utility of this compilation that provided the motivation for the HCI
Bibliography project. Large collections of bibliographic information have
been contributed by:
Ben Shneiderman (Univ. of Maryland)
Patricia Sheehan (Xerox PARC)
Craig Hartley (Martin Marietta)
Alison Lee (Univ. of Toronto)
David D. Stubbs (Tektronix)
Jakob Nielsen (Bellcore)
Liz Sonenberg (Univ. of Melbourne, Australia)
Lorraine Normore (Chemical Abstracts Service)
Bob Allen (Bellcore)
Karen Ward (Oregon Graduate Institute, Beaverton, OR)
Ellen Yu (Univ. of Maryland)
Fabio Vitali (Univ. of Bologna, Italy)
Scott Robertson (Rutgers Univ.)
Tami L. Crawford (Canisius College, Buffalo, NY)
Deb Galdes (Silicon Graphics)
Scott Overmyer (George Mason Univ.)
Saul Greenberg (University of Calgary)
Patrick Holleran (Apple Computer)
Michael Wilson ()
and by others at the Ohio State University.
6.2. Validation: The following people have volunteered to validate records,
usually in batches of 50 or more. Validating a set of records typically
requires a commitment of 5-10 hours, and we are always interested in getting
more volunteers.
Alan Edmonds (Ohio State Univ.)
Deb Galdes (Silicon Graphics)
Thom Gillespie (Univ. of California, Berkeley)
Jonathan Grudin (Aarhus Univ., Denmark)
Sharon Irving (Univ. of Colorado, Boulder)
Alison Lee (Univ. of Toronto)
Paul M. Mullins (Youngstown State Univ.)
Bill Ogden (New Mexico State Univ.)
Karen Ward (Oregon Graduate Institute, Beaverton, OR)
Virginia Peck (Carnegie-Mellon Univ.)
Steve Franks (Texas Instruments)
and Gary Perlman. If you would like to volunteer to validate, send a message
listing the works you would be able to validate to hcibib@cis.ohio-state.edu.
The first line of the message should be:
Validate:
6.3. Ideas: We thank the following for their ideas and/or feedback: Don
Norman, Darrell Raymond, Jonathan Grudin, and there are probably others we
have missed, for which we apologize. To send suggestions to the project, send
a mail message to hcibib@cis.ohio-state.edu. The first line of the message
should be:
Suggestion:
--
Name: Gary Perlman | Computer and Information Science Department
Email: perlman@cis.ohio-state.edu | Ohio State University, 228 Bolz Hall
Phone: 614-292-2566 | 2036 Neil Avenue Mall
Fax: 614-785-9837 or 292-9021 | Columbus, OH 43210-1277 USA