Source

Automatically imported from: http://commons.somewhere.com:80/rre/1996/computists.html

Content

This web service brought to you by Somewhere.Com, LLC.

computists

``` Date: Thu 4 Jan 96 23:46:38-PST From: Ken Laws To: ;;@FM_14 Subject: TCC 6.2 (Full Moon Edition)

---

AI Vol. 6, No. 2 IS January 4, 1996 CS THE COMPUTISTS' COMMUNIQUE

1> Information retrieval. 2> Internet reference services. 3> Web indexes and spiders. 4> Research software. 5> Electronic journals.

---

We are "web weavers and dancers at the dawn of the Meso-Electronic Period." -- Paul Evan Peters. [Karen Campbell , 12/95.]

Goood Morning, Full Moon Subscribers!

I've split the Communique into two issues per week, of about 12KB each. You should find it more readable, less intimidating, and closer to real-time news reporting. Full Moon subscribers still get 12 issues per year, but Computists will get at least 88 issues. Ask me about our rates and services if you're ready for the full Communique stream. New members are always welcome, and we have discounts for departments and other groups.

Our APJ digest for graduate-level, non-AI faculty, and applied jobs has been renamed the Computists' Applied Jobs (CAJ) digest. Likewise, our RSW digest has become the Computists' Research Software (CRS) digest. They will usually be about 45KB/week, and are free to anyone who asks. Our MS/PhD-level Computists' Career Jobs (CCJ) digest, however, is only for members of Computists International.

We don't have a website yet, as I'm still checking out service providers. (Recommendations are solicited. :-) Resource leads for webmasters will be a topic of continuing interest this year, along with tips for computer scientists, programmers, and entrepreneurs.

1> Information retrieval:

CMU's Ken Lang has built a "Newsweeder" that filters Usenet newsgroups for messages of interest. See , or for the commercial version. Lang had a paper in Proc. 12th Int. Conf. on Machine Learning, pp. 331-339. [Tom Dietterich , comp.ai, 12/26/95.]

NetSumm is an experimental WWW page summarizer produced by the Language Group at BT Laboratories. It highlights the "most important" sentences of a web page, or extracts these sentences to abridge the article. ; comments to . [Bill Park , 12/23/95.] (Probably doesn't work with non-graphic browsers such as Lynx.)

For indexing your personal Mac files, try EndNote Plus 2.0 by Niles & Assoc. [Toby Moore , Mac*Chat, 10/20/95.] (That's primarily a bibliographic system. Other programs such as Locate or ON Location can search random files for keywords.)

2> Internet reference services:

Internet Findex offers a comprehensive list of dictionaries, gopher pointers, Internet indexes, and Web search engines, updated weekly. . [John Hart , net-hap, 10/13/95.]

Internet guides, service lists, reference pages, FAQs, and search services are listed on the Galaxy site, . [Susan Klopfer , net-hap, 12/1/95.]

A comparison of several Web search engines, plus articles about searching, has been posted to . [Karen Campbell , web4lib, 10/31/95.]

An Internet guide, tutorial, and over a dozen search engines can be found on Robert Kabacoff's Inter-Links, at . The guide lists several thousand hand-picked WWW, gopher, telnet, FTP, IRC, MUD, and BBS resources. [Scout Report, 12/15/95.] (Formerly .)

The Internet Sleuth is a listing of over 750 searchable Internet databases. . [, net-hap, 12/17/95.]

Sources is an e-journal for the intelligence community -- as in "private eye," information broker, or spook -- examining methods of intelligence gathering and information interpretation. . [Glenn Davis , newjour, 11/28/95.]

InfoSeek at has recently added the CorpTech Directory of 40K technology companies and the CSA Worldwide Market Research Database of company profiles and market research reports in defense, finance, and basic industries. Also Microcomputer Abstracts (with reviews from 90 publications since 1989); SoftBase: Reviews, Companies & Products (200 periodicals); and databases covering work-group computing and AT&T. One-month free trial, plus $10 in trial credits if you enter *biz in the "referred by" box when you register for the business databases, or *comp for the computer databases. [Julian Stewart , net-hap, 11/20/95.]

3> Web indexes and spiders:

The WAIS search engine at simultaneously searches all public WAIS databases. [Sam Sternberg , net-hap, 12/17/95.]

The European Directory searches European website descriptions. Access to search or to register new services. [Jim Kissel , net-hap, 12/17/95.]

NET Navigator is a new URL search engine, similar to Yahoo!, Lycos, Web Crawler, and Infoseek. . [, net-hap, 12/17/95.] (URL search lets you find who is linking to your own site.)

Webcatcher notifies you of new sites matching your specified topics. . [Liz W. Tompkins , net-hap, 12/15/95.]

The Excite search service at is very powerful, but results can vary from day to day. [Internet-on-a-Disk, 11/95.]

SavvySearch is CSU's experimental server, able to search email addresses, Usenet, FTP sites, gopher space, and web space. 19 search engines are used. . [Daniel Dreilinger. Sam Sternberg , net-hap, 12/5/95.]

The Open Text Index and search engine indexes every word of almost 1M web pages. You can use Boolean, phrase, proximity, weighted, and KWIC searching, with relevance feedback. Also, searching of just parts of pages. . [Network News, 12/9/95.]

UCB's Inktomi searches 1.3M documents on the web -- five times more than Infoseek -- and it's faster than Lycos because it harnesses four Sun workstations and 32 networked computers in UCB's CS building, Soda Hall. There's no Boolean logic, but you do get to see the actual keyword hits within each document. . Eric Brewer and his student Paul Gauthier developed the site. Inktomi ("ink to me") is named after a mythological trickster spider of the Plains Indians. [UC NewsWire. , net-hap, 10/18/95.]

Digital's new Alta Vista high-speed search engine scans 13K Usenet discussion groups in addition to Web sites, indexing every word. The prototype sends out "a brood of spiders" (or "threads"). . [NYT, 12/18/95, C2. EDUPAGE.] Alta Vista is said to be 100 times as fast as competitors, and indexed 16.5M pages in 8 days. A Digital engineer estimates that the web includes 40M-50M pages -- not counting graphics, audio, and video -- on 130K servers, and that Digital has the horsepower to maintain a current index of the whole thing. [Eric D. Bauer , online-news, 12/18/95. net-hap.] (Bauer ran a search on "Middlesex News": InfoSeek found 72 matches; Alta Vista found 1,000.)

4> Research software (in our CRS digest this week):

FUZZLE20: fuzzy logic software.

TDL for NN and GA programming.

WebBase: WWW CGI forms input and FoxPro/DBF database for WebSTAR.

Mitek OCR for hand-printed character recognition.

RECORE 4.0 OCR/ICR Toolkit: OCR for Windows 95. Genetic Algorithms and Genetic Programming at Stanford 1995: Stanford student papers, edited by John Koza.

Pioneer Programmable Computer Conveyance: SRI robotic platform for $1995.

WeboDex Organizer: browser bookmark organizer for MS Windows.

SURFACE: scattered-data surface generator/viewer for Windows 3.1.

MacroLife: an alife program with 4B cells and built-in lifeforms.

Ofront Evaluation Shop (OES): Oberon-2 to C translator.

LitProg Archive: resources for literate programming.

QUOTS 1.2: motivational quotation screen saver.

XRAYDIF: X-ray powder diffraction pattern simulator.

Wily: Unix/X text editor and user interface.

5> Electronic journals:

Forbes magazine suggests that the first major journal to fall victim to online publishing may be Elsevier's $10,775/year Nuclear Physics B. The physics preprint exchange set up four years ago by Paul Ginsparg now handles 70K transactions/day. Formal review will be added soon. [Robert L. Park, WHAT'S NEW, 12/22/95.]

The J. of Functional and Logic Programming (JFLP) from MIT Press is now available at . "New, peer-reviewed, electronic, reader-powered." JFLP articles are in LaTeX and Postscript, augmented by refereed, forward references to improvements and subsequent, related work. MIT Press plans to develop a cost-based -- instead of market-priced -- model for nonprofit, electronic, paid-subscription journals. JFLP is $30 for 1966, or $125 for institutions. ISSN: 1080-5230; , , 617-253-2889, 617-577-1545 Fax. [newjour, 12/9/95.] (The Andrew W. Mellon Foundation contributed $150K to this effort and The Chicago J. of Theoretical Computer Science.)

The Nordic J. of Philosophical Logic (NJPL) from UOslo is online at , in addition to its hardcopy publishing. Mailing lists, newsgroups, and live discussion groups are planned. [Johan W. Klüwer , newjour, 11/14/95.]

The Very Large Data Base (VLDB) Journal will add an online version starting 1/96, at . Subscribers also get the hardcopy version, which will lag by several months. Back issues for 1992-95 are being offered for $50 (normally over $500) while they last. New subscriptions are $48/year, from , +49-30-82070, +49-30-8207448 Fax. [Arie Shoshani , dbworld, 12/26/95.] (Full online papers are free during most of 1/96.)

The Int. J. of Intelligent Systems in Accounting, Finance and Management offers abstracts at . [newjour, 11/28/95.]

ACM's J. of Experimental Algorithmics, , covers the design, analysis, implementation, testing, evaluation, and reuse of algorithms and data structures. JEA will also distribute research code and testbeds. [Michael Allen Clore , newjour, 11/20/95.]

If you're really into math, the Electronic J. of Linear Algebra (ELA) is at and mirrors. Free to ILAS members. [Daniel B. Szyld , newjour, 11/15/95.]

Another new peer-reviewed e-journal is the Int. J. of Neural Regulation, at . Electronic bits only. [newjour, 11/15/95.]

The Journal of Higher Education is experimenting with an electronic version, on . It will lead the print version by 4-6 weeks. [, newjour, 11/1/95.]

Configurations is a journal of literature, science, and technology, from Johns Hopkins University Press. Theories and practices of science, technology, and medicine, and their relations to literature and the arts. . [, newjour, 11/3/95.]

Name that publication: "X and Y Journal is the first publication to feature X and Y; and it especially focuses on the integration of these two paradigms as well as on their common foundations. The journal's articles range from theoretical study of X and Y to "real world," applied investigations of X and Y. An international distinguished board will edit the journal." Act now to take advantage of this special offer. :-)

-- Ken

I am generally not interested in using Internet to see work that wouldn't interest me in person or in a book. That fact that something is on Internet usually doesn't improve it. -- Ken Friedman , "Language and Culture in the Information Age," Art and Design, 11/95. [Joe Raben .]

---

ISSN 1084-015X. Publisher/Editor: Dr. Kenneth I. Laws, 4064 Sutherland Drive, Palo Alto, CA 94303; (415) 493-7390. Internet: laws@ai.sri.com (courtesy of SRI International). Copyright (C) 1996 by Kenneth I. Laws. Computists' Communique is a service to members of Computists International. Members may make copies for backup, direct mentoring, or recruiting, and may extract occasional articles if attribution is given.

--- ```

This web service brought to you by Somewhere.Com, LLC.