|
OR/MS Today - December 2002 Cyberspace A Thousand Years of Archives By ManMohan S. Sodhi I am sure you found the last issue of OR/MS Today (October 2002) celebrating 50 years of operations research delightful. Now imagine a similar issue a hundred years from now (in the year 2102), a thousand years from now (year 3002), even ten thousand years from now (year 12002)? What would we see in such an issue? How will we ensure this information actually gets there? In an age of increasingly shorter lifecycles and even shorter attention spans, one project that captures my imagination is the Clock of the Long Now, a project to design and construct a clock that would last 10,000 years. [1] The obvious questions pertain to materials and technology, but there is a lot more to think about. One such question is that if we are to have a library adjoining the clock, what should that library contain? As OR/MS professionals, we create, gather and archive information in the form of articles, books, computer programs, lecture notes and even our memories. In this article, I raise questions about our ability to preserve this information for future generations and challenge you to think of answers in terms of hundred years, a thousand years, even ten thousand years from today! File Formats Even if we have standard ways of archiving articles digitally today, looking at the near past gives little confidence about our abilities for the long term. One of the more popular file formats for archiving articles currently is the PDF format; you can get OR/MS articles from INFORMS PubsOnline, Elseivier's Science Direct or Kluwer's KluwerOnline in this format. Adobe Acrobat, the reader software, is freely available for the most commonly used operating systems, so think of it as a standard. But what was the standard five years ago? How easy is it for you to read digital files from five years ago? For instance, I used WordPerfect and now I cannot open older files unless I reinstall that software. You may have Wordstar or VisiWord files. Even with Microsoft Word, you may not be able to read older Word files or those created on a different operating system, like Apple's. So, if we have such a mess after just five to 10 years, how can we assume our current "standard" solutions will take us through the next decade, let alone a hundred or a thousand years? A simple solution is to archive articles in plain text files that can be opened by any software so that future generations can read them. LaTeX and TeX, typesetting software used for technical or scientific documents, are elegant solutions. A newer solution is a variant of eXtensible Markup Language (XML) called MathML intended to facilitate the use of mathematics on the Web [2]; many vendors, including Wolfram Research, are backing this format. But what should the OR/MS standard be? Computer programs are already in text files, as well as data when not in a database. There are many data file formats. Among the standard ones for linear programming are MPS and the one for ILOG's CPLEX, so the problem is not as acute as with archiving articles. Of course, changes in format, computer languages and even compilers will remain problematic. The Medium Even if files can be read in the future, the medium for archiving is a headache. It is hard to retrieve older OR programs written on punch cards, magnetic tape, etc. And how would you open any of your 5.25-inch floppy disks on your PC? Old 3.5 single density disks on your new Apple? For the future, you can save your information on a CDROM, but there are plenty of questions about whether CDs will be used longer than a few years. Selection and Filtering Even if we agreed on standard text-based formats and found a wonderful medium for storage, the question of what we should store remains. True, digital memory for archiving is cheap and getting cheaper. Could we not store everything? Think again. According to Hal Varian and Peter Lyman [3], the world currently produces between one and two billion gigabytes of unique information per year, growing at 50 percent per year! Even if we assume zero growth, there is going to be a lot of "unique" information in a hundred or a thousand years. And even if we did manage to save it all, who will go through it in the future? The same applies to paper archives, but how many issues of Operations Research and of Management Science would we have in ten thousand years? Who should filter OR/MS archives and how? Do we let "market forces" determine that? And if the invisible hand of the market is to determine everything, should we leave our entire field passively to the same? Can we really then look forward to an OR/MS Today issue celebrating a hundred, a thousand or even ten thousand years of operations research? References
ManMohan S. Sodhi is a member of the operations management faculty at Cass Business School in London. He welcomes your comments at M.Sodhi@city.ac.uk. OR/MS Today copyright © 2003 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 2003 by Lionheart Publishing, Inc. All rights reserved. |