Authors Referees APS Members |
Speaker: Bob Austin (Princeton University)Panel Discussion: Tracking of PapersA Humble Experimentalist's View of Electronic PublishingI. A Day in the Life. I assume in my remarks that the scientist is a PC user, TeX literate, rather broke physicist (about $100/year) who teaches and does research in a small lab with 1-4 other people, usually grad students and a post doc. Lock-ins take precedence over Sun workstations!!! This usually means that you have a few 486 based PCs with maybe 500 megabytes of hard-disk, a laser post-script printer. Ethernet link to a VAX or maybe UNIX based server. This means that the physicist is working his/her ass off all the time, running between classes, committee meetings and precious hours in the lab. Grant writing is a serious drain on time. Getting 1/2 an hour free each day to handle e-mail is not always possible. Learning cute UNIX tricks, when compared to trying to write coherent lectures or carry out that critical experiment, is way down on the priority list. ii. What is the advantage of electronic publishing for our heroic foot-soldier? RAPID and COMPLETE scanning of the literature would not only help one keep abreast of the literature, but also would ease the pressure to publish in the few journals that people actually read (Science, Nature, PRL). Papers could be quickly down-loaded and printed (WITH HIGH RESOLUTION 600 DPI GRAPHICS) to spare time spent trying to track down papers and xerox them. High resolution image files and appendixes with large amounts of experimental details would make papers much more useful. Cores of the paper would be readily available, with options (as in a Hypertext environment) for detail IF DESIRED. I am not sure however how much the electronic publishing community realizes how much data is involved in high-resolution graphics. I showed about 1 minute of video footage of DNA moving in a maze. Since the image quality must be high to see the molecules, a minimum of 250 kilobytes of binary data per frame is necessary, 250 kbytes of binary data turns into an ASCII Postscript file of about 2.5 megabytes length. The 1800 frames I showed then represe nted between 0.5 (binary) to 5 (ASCII)gigabytes of data! Compression algorithms do not work well with scientific data. A single frame image is available in postscript (2.5Mb eps image) and pdf (41kb pdf image). Example of Video Sequences Included in Paper. I don't expect a great cost savings due to the medium! E-publishing for an experimentalist will consist of value added options (such as animation and detailed appendixes) which will be surely require more effort and technology on the publisher's side. Perhaps the theorists may see a brave new world of cost-free publishing, but I don't. III. What about the ideal of a non-refereed literature? I, myself, make big mistakes quite often. I need good referees to shake me up, point out my mistakes, keep me honest. Occasionally I met a referee that is an IDIOT, but that is usually obvious to all. I wish that the people who review grants were as good as the typical Phys. Rev. Referee. Preprint servers are interesting, but you must remember that an experimentalist must be very sure that a result is correct before committing valuable time to pursuing an observation:
Review Articles, based upon a hypertext layering system, would be an extraordinarily productive way for novices and experts alike to keep current with the flow of a field. These "living books" if you will could be quite current and quite critical of the literature. Reviews of Modern Physics at present takes too long to come out with reviews. These could be monthly up-dated resources which can be probed down to the basics or up to the most current results. IV. Whatever we do, the software must be extremely user friendly and tolerant of harried people with lousy memories. UNIX freaks must be controlled at all cost! Note added in proof: the author has just purchased an SGI Unix based Indy computer and is busy trying to become a Unix freak. |
