INFORMS Online
OR, Computers and the World Wide Web
By Matthew Saltzman
The history of OR and the history of computing have been closely tied since the early days of both fields. Dantzig's original paper, "Programming in a Linear Structure," appeared in 1948 and the first published benchmark of the simplex method (on problems with up to 10 variables!) appeared in 1953. While advances in hardware and algorithms have continued to pace each other since then, other developments in computing have also had profound effects on how we conduct our business.
In keeping with this issue's anniversary theme, I will survey some of these developments that have occurred during my career. Of course, I've been in the profession for only a little more than 20 years, not 50. But I've been computing for more than 35 years, and I believe that this period covers the most dramatic and important advances.
As a Cornell undergraduate in the 1970s, my computing environment consisted of punch cards and batch submission from remote job entry stations. My compiler course textbook was typeset on a chain printer using one of the first text formatting programs — the text was punched on cards. By the time I graduated, the environment had advanced to interactive mainframe computing using DECwriters and "glass TTYs" (80x24-character CRT screens).
When I started graduate school in 1980 at Carnegie Mellon, interactive mainframe computing was the norm. Production of papers was still primarily the province of technical typists using IBM Selectric typewriters or specialized word processing machines with daisy-wheel printers. But several important revolutions were underway.
The advances with the most immediate impact on me were laser printers, the EMACS editor and the Scribe document production system (developed by CMU grad student Brian Reid). Together, these tools allowed authors to produce near professional-quality mathematical documents.
At about the same time, Don Knuth was taking time off from writing "The Art of Computer Programming" to develop TeX and Metafont, tools for creating typefaces and typesetting mathematical documents. Leslie Lamport was developing LaTeX, a document production system (like Scribe) written for TeX. The free TeX/LaTeX combination eventually became the tool of choice for markup-based document production.
Another revolution was the development of the commodity personal computer with office tools. The most successful one, of course, was the IBM PC. In 1981, a typical configuration supported a 4.77 megahertz (MHz) CPU, an eight-bit bus, up to 640 kilobytes (Kb) — yes, that's Kb — of RAM, two 5 1/4-inch, 360 Kb floppy disk drives and a monochrome monitor. It cost about $4,000 to $5,000 (1981 dollars). Early PC hard drives were five or 10 megabytes (Mb) — yes, Mb — and cost several hundred dollars. The first color displays were marketed almost exclusively for gaming — they had four colors and 320x240 screen resolution.
Since then, PC computing power has improved by about three orders of magnitude in almost every dimension. Top-of-the-line processors are two gigahertz (GHz) with a 32- or 64-bit bus architecture, and 256 Mb of RAM, 60 gigabytes of hard drive space, and 32-bit, 1280x1024 color displays are the norm. Currently, systems cost in the hundreds, not thousands of dollars.
For the past decade, Microsoft has dominated the PC office suite market, and the most common document production application is Microsoft Word.
The advent of the PC also dramatically changed teaching. As hardware became affordable, computing in the classroom became ubiquitous. Business classes use spreadsheets for statistics, modeling and analysis (including optimization). Engineering and science disciplines use specialized statistics, optimization and simulation packages as well as computational tools such as Matlab, Maple and Mathematica. Use of these tools in the classroom drives their use in industry, as students graduate ready to integrate computation into their jobs.
These innovations are not without drawbacks. Large-scale spreadsheet models are difficult to validate, and plug-in solvers have been considered less robust than specialized programs. Over-reliance on computational tools can lead to lack of understanding of concepts. Effective integration of technology into the classroom is difficult and time consuming, and often is not rewarded by academic institutions.
The next revolution, the opening of the Internet, fostered the growth of e-mail as the most efficient person-to-person communication medium. The late 1980s saw the introduction of text-based file browsing at remote locations. In 1989, Tim Berners-Lee developed a standard for graphical display of remote files containing appropriately marked-up text, including hypertext linking among files at different locations. Since the mid-1990s, the World Wide Web has had profound effects on how we communicate with each other and disseminate our research results.
INFORMS Online (www.informs.org) and Mike Trick's OR Page (now www.informs.org/Resources), dating from about 1995, were among the earliest Internet sites devoted to OR. The Resources page now lists more than 1,200 Web sites with OR-related content. Abstract submission and registration for INFORMS meetings, elections and member profiles are now managed with Web tools, and renewals and new-member signups are on the way. Preprints can be easily distributed by authors. Our archival journals are published on the Web as well as in print. Web-based manuscript tracking is in the works.
But we have barely begun to exploit the potential of online publishing. Hyperlinking, motion, interactivity and "living documents" kept current by their authors are just beginning to appear. Hyperlinked references are coming soon to institutional versions of INFORMS' electronic journals. The now-defunct ITORMS (itorms.pubs.informs.org) and its living document concept were, I hope, only slightly ahead of their time. The COIN-OR project (www.coin-or.org) supports publication of software. Good examples of what can be done appear in the online journal INFORMS Transactions on Education (ite.pubs.informs.org).
I expect the synergy of OR and computing technology to produce developments in the near future fully as exciting as those of the past 20 years.
Matthew Saltzman (mjs@ces.clemson.edu) is an associate professor of Mathematical Sciences at Clemson University and the editor of Informs Online.
OR/MS Today copyright © 2002 by the Institute for Operations Research and the Management Sciences. All rights reserved.
Lionheart Publishing, Inc.
506 Roswell Rd., Suite 220, Marietta, GA 30060 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com
URL: http://www.lionhrtpub.com
Web Site © Copyright 2002 by Lionheart Publishing, Inc. All rights reserved.