Date: Sat, 5 Jul 86 17:30:44 edt From: vtisr1!irlistrq To: fox Subject: IRList Digest V2 #30 Status: RO IRList Digest Saturday, 5 July 1986 Volume 2 : Issue 30 Today's Topics: Reply - Diff at the word level [See issue 24, May 21] Discussion - FERRET [see issue 25, May 25], SMART, SIRE Announcement - NSF's Project EXPRES Announcement: Office Information Systems Conference Program Available COGSCI - The Oncocin Project, An ES to assist with Cancer Treatments ---------------------------------------------------------------------- Date: Mon, 9 Jun 86 11:24:44 EDT From: seismo!harvard.HARVARD.EDU!macrakis (Stavros Macrakis) To: nlm-vax.arpa!donna Subject: word-wise diff [This is a reply to query of Donna Williamson appearing in Issue 24 - Ed] The classic Unix-style approach to this is to translate the original files to one-word-per-line, use diff, and then translate the diff back. The main problems with this are: 1. Translating so that you can recover the original is finicky; 2. `diff' is (very) slow; for simple changes, `diff -h' may work; 3. Mapping the diff results back to source file terms is painful. You might want to look at Walter Tichy's paper `The String-to-String Correction Problem with Block Moves' ACM Trans. Comp. Sys. 2:4:309 (Nov 84) for an algorithm that seems far superior to that used in `diff'. His implementation ran at about the same speed as `diff' in tests, delivering 7% smaller output; it promises to be much faster and much better in environments with more general kinds of changes. -s ------------------------------ Date: 17 Jun 1986 22:12-EDT From: seismo!cad.cs.cmu.edu!Michael.Mauldin Subject: Re: try again to send msg, now with correct address Dr. Fox, I will send you a written copy of my proposal. Thanks for the comments, and thanks for pointing out IRlist. I started subscribing to IRlist just this month, so I am already in your tables. Dr. Salton's ``single term indexing'' for comparison sounds like it would reduce my workload and also give me a better comparison. Do you have his net address so that I may contact him? My opinion is that when my ``fancy conceptual text skimmer'' works, the ``right'' solution will be a hybrid of both keyword based and conceptual based retrieval systems. They each work best on different types of documents, and the combination of approaches will have better recall AND precision. That is similar to the SIRE approach: keywords to get EVERYTHING that might be related, and something more intelligent to do the final winnowing. I will be submitting to IRlist as sooon as I feel I have something important to say. In the meantime, I look forward to reading it (and, of course, storing the contents in FERRET). Michael L. Mauldin (Fuzzy) Department of Computer Science ARPA: Michael.Mauldin@CAD.CS.CMU.EDU Carnegie-Mellon University Phone: (412) 268-3065 Pittsburgh, PA 15213-3890 [Note: G. Salton is, on the ARPANET, at gs@cornell.arpa and Mr. Mauldin is, I believe, now in touch with him. - Ed] ------------------------------ Date: Wed, 2 Jul 86 18:38:48 edt From: vtopus!fox (Ed Fox) Subject: Project EXPRES On 4/17/86 the National Science Foundation (NSF) announced 2 solicitations in the Commerce Business Daily. The 2nd, RFP 86-104, has been suspended. The first, which now covers both topics, has been distributed as EXPRES Project Solicitation for Research Groups, NSF 86-34. A meeting was held at NSF on 13 June for EXPRES Research Group Proposers - I attended along with almost 40 others. The following summary of the solicitation and meeting is based on documents released by NSF, and my own impressions. I. Solicitation: "The Office of Information Systems of the National Science Foundation (NSF) has initiated a project called "EXPerimental Research in Electronic submission" (EXPRES). Its purpose is to improve the ability of the Nation's research community to exchange documents containing text, image and graphics. The vehicle for experimentation will be the NSF proposal development, submission and review process. Awards will be made to research groups to study the problems of compound document transmission among different hardware/software environments. The project duration will be three years, beginning in September, 1986. NSF anticipates making awards to two research groups. ... Proposal must be received on or before Friday, August 1, 1986 to be considered for award under this solicitation. BACKGROUND ... The Goal: To move to a situation in which the scientific and engineering research community, building on computing equipment and facilities currently available, will benefit from emerging technologies and will be able to exchange compound documents among dissimilar [multivendor, multiuniversity, multidisciplinary] hardware and software environments. [Compound documents are those which may contain text with numerical and symbolic data, images, drawings, charts and graphs, and photographs.] ... The Approach: The National Science Foundation has initiated a project called "EXPerimental Research in Electronic submission" (EXPRES). The vehicle for experimentation will be the NSF proposal development, submission and review process, which embodies all aspects of the problem [high volume, geographically dispersed community, and multivendor environments] but permits a measure of control necessary for experimentation. The National Science Foundation will make awards to groups to perform research on overcoming the barriers to compound document transmission in the multidisciplinary, multiuniversity and multivendor environment. The research effort will include the design and installation of pilot systems with the aim of expanding them. Systems characterized as "high-function, single vendor" or "mediaum/limited-function, multivendor" could be used as a basis for the project. The awardee[s] will cooperate with industry and government initiatives advocating and promoting interoperability [communication among multivendor equipment and software] based on the International Standards Organization Open Systems Interconnecttion [OSI] standards. The National Science Foundation: NSF faces a subset of the communication problems of the Nation's research community. Scientists and engineers at about 2,400 organizations submit proposals to the National Science Foundation each year; the vast majority of these organizations is educational. NSF receives approximately ten copies of 37,000 proposals annually. Each proposal averages 50 pages and may contain more than 100 pages. Most contain not only text but images and graphics; some contain black and white or color photographs. After their initial review at NSF over half of these proposals are mailed to 6-8 reviewers for analysis. ... ... Research Areas: It is anticipated that research in the following areas will be required during the project: - document representation and management - networking protocols - authentication, certification and security - data compression - data management - high volume image transmission - graphics - human-machine interface - user acceptance and impact on work and social patterns Electronic proposal submission: The EXPRES experimental design will be based on the following model for electronic propsal preparation, submission and processing. - The principal investigator prepares the proposal using a computer workstation. - The proposal is forwarded through institutional channels in the Research Administration Office where it is approved by the authorized officials - It is then submitted over NSFNET or another wide-area network to NSF - A receipt acknowledgement is sent to the submitting institution - The proposal is entered into the computerized processing systems at NSF and assigned to a program manager - After initial review, the program manager selects reviewers - The proposal and reviewing instructions are sent to the reviewers - Reviews are completed and sent back to NSF - The program manager evaluates the reviews, and decides to recommend award or declination of the proposal - For recommended awards, the necessary forms are completed and sent with the proposal through division and directorate approvals, then to the administrative divisions for final processing - For declinations, the decision justification and proposal are sent to the division director for concurrence - The notification of award or declination is transmitted to the institution It is expected that paper copies of the proposal and forms could be generated at any stage in this process, but transmisison and handling would be primarily of electronic media. Cost Sharing: The current total level of NSF funding available for this EXPRES activity is approximately $2 million per year. Accordingly, it will not be possible for NSF to bear all the costs associated with the project, and significant cost sharing by the research groups may therefore be necessary. ..." [Note: the solicitation advertised the meeting of 13 June, and gave David Staudt at (202) 357-7448 as the contact.] II. Minutes of Meeting "... What are your really looking for? What will constitute "success"? As stated in the Management Plan, "The primary objective of EXPRES is to provide an environment in which compound documents can be transmitted efficiently and economically." What we want is a system that will eventually work, that can be used by the academic scientific and engineering research community to communicate electronically quickly, efficiently, economically, fully and robustly. ... ... What will happen to the installed equipment and software at the end of the project? Disposition of any equipment provided to colleges and universities as part of the EXPRES project is a matter between them and the vendors. Any equipment installed at NSF must either be returned or purchased separately. ... Will the pilot system have to handle fascimile or some other method of inputting paper proposals? No. The purpose of the pilot is to research and demonstrate an electronic system. ... What percentage of proposals contain photographs and other image? The percentages vary by discipline, but about 75% have graphics or image. Some have color. We would expect that an electronic system would ultimately be able to handle color. ... How important is it to be able to edit the data contained in the proposal? We wouldn't necessarily need to change the proposal, but we would want to be able to search for words, for example, or otherwise be able to scan the document using its contents. ..." III. My comments 1. Interested parties a- Companies Serious interest was expressed by vendors like IBM, Digital, Xerox Organizations like BBN, Aspen have done work in related areas b- Educational Institutions MIT, Carnegie-Mellon, Berkeley are expected to compete States like NY, NJ, Michigan may submit proposals 2.Communications NSFNET will be used; there is interest in moving to the OSI protocols but that may take longer than it will take to complete this part of EXPRES 3.Importance The Director of NSF, and several of the program directors, attended the meeting. There is keen interest at NSF in this venture! 4.Becoming involved Due to the short time before proposals must be submitted, and the orientation to having consortiums send proposals, it seems imperative for anyone interested to become part of a large consortium of say 2-5 universities along with 1-3 vendors. 5.For more information: Contact David Staudt at the phone number given above. I hope that the information I included is enough to help people decide if they are interested in getting involved. But the intent of the explanation is really for those who would like to know that such an effort is underway! - Ed ------------------------------ Date: Tue, 24 Jun 86 00:55:09 edt From: rba@PETRUS.ARPA Subject: Office Information Systems Conference Program ACM CONFERENCE ON OFFICE INFORMATION SYSTEMS October 6-8, 1968, Providence, R.I. Conference Chair: Carl Hewitt, MIT Program Chair: Stan Zdonik, Brown University Keynote Speaker: J.C.R. Licklider, MIT Distinguished Lecturer: A. van Dam, Brown University Panels and Sessions Advanced Computational Models AI in the Office Impacts of Computer Technology on Employment Organizational Analysis: Due Process Future Directions in Office Technology Comparison of Social Research Methods Organizational Analysis: Organizational Ecology Models of the Distributed Office Interfaces For more information, call the Conference Registrar at Brown U. (401-813-1839), or send electronic mail to mhf@brown.CSNET. ------------------------------ Date: Sat, 21 Jun 86 06:53:08 edt From: DEJONG%OZ.AI.MIT.EDU@MC.LCS.MIT.EDU Subject: Cognitive Science Calendar [Extract - Ed] Thursday, 26 June 1:00-4:00pm Room: E51-329 EXPERT SYSTEMS SEMINAR The Oncocin Project An Expert System to Assist Physicians with Cancer Treatments Curtis P. Langlotz Stanford University The Oncocin Project is overseen by a collaborative group of physicians and computer scientists who are developing an intelligent system that uses the techniques of knowledge engineering to advise oncologists in the management of patients receiving cancer chemotherapy. The general research focus of the group members include knowledge acquisition, inexact reasoning, explanation, and the representation of time and of expert thinking patterns. Much of the work developed from research in the 1970's on the MYCIN and EMYCIN programs, early efforts that helped define the group's research directions for the coming decade. The prototype ONCONCIN system is in limited experimental use by oncologists in the Stanford University Oncology Clinic. Thus much of the emphasis of this research has been on human engineering so that the physicians will accept the program as a useful adjunct to their patient care activities. ONCOCIN has generally been well-accepted since its introduction, and we are now testing a version of the program which runs on Xerox 1100-series professional workstations to further enhance its acceptability and to facilitate its evaluation at sites away from the University. Hosts: Sloan School of Management, MIT Xerox Arificial Intelligence Systems Info: (603) 881-7028 ------------------------------ END OF IRList Digest ********************