Date: Mon 8 Aug 1988 00:22-EDT From: AIList Moderator Nick Papadakis Reply-To: AIList@mc.lcs.mit.edu Us-Mail: MIT LCS, 545 Tech Square, Rm# NE43-504, Cambridge MA 02139 Phone: (617) 253-6524 Subject: AIList Digest V8 #40 To: AIList@mc.lcs.mit.edu Status: R AIList Digest Monday, 8 Aug 1988 Volume 8 : Issue 40 Queries: Consistent Labelling Problem Response to - Ornithology as an AI domain Response to - Gardening ES Response to - Dave Goldbery's address Response to - RightWriter and Grammatik II (AILIst v8 #27) ---------------------------------------------------------------------- Date: 2 Aug 88 16:14:39 GMT From: mcvax!ukc!reading!onion!isg.cs.reading.ac.uk!rmb@uunet.uu.net (Rob Bodington) Subject: Consistent Labelling Problem I am researching into the Consistent Labelling Problem. Does anyone have any references on this topic? I will summarise the response on the net. Thanks. Intelligent Systems Group JANET: Rob.Bodington@reading.ac.uk Dept. Computer Science University of Reading Whiteknights, Berks. U.K ------------------------------ Date: Wed, 3 Aug 88 00:15:03 EDT From: Marvin Minsky Subject: Response to - Ornithology as an AI domain Have you considered ethology, rather than ecology? There is a world of things to do in that realm, because the traditional models of animal behavior are based on virtually no methodology. For example, Tinbergen and later workers have described hypothetical mechanisms involved in building bird nests - but no one knows how adequate they are. A great research thesis would be to see if the sorts of computational structures proposed by ethologists could actually build a respectable nest, using some sensors, robot manipulators, and suitable behavioral control structures. I have a couple of students simulating various sorts of animal behavior, but there is a whole unexplored universe there. - minsky ------------------------------ Date: Fri, 05 Aug 88 16:58:20 GMT From: IT21%SYSB.SALFORD.AC.UK@MITVMA.MIT.EDU Subject: Response to - Gardening ES Re. parvis@gitpyr.gatech.edu and otter!ijd@hplabs.hp.com (Ian Dickinson) Parvis wanted details of gardenign packages, and Ian Dickinson replied that there was a farming ES by ICI plc. The system was called COUNSELLOR, and provided advice to farmers via a ViewData system on the use of fungicides. It predicted the disease profile and then recommended sprays throughout the season. It allowed the farmer do to what-iffing for cost-benefit analysis. I worked on it in the early days, and am pleased to see that it is one of the ESs that went into use (about 3 years ago) and, I believe is still being used. Looking back, I rather wish I had built an ES to advise on organic farming instead! For fuller details see Jones M.J., Crates D.T. (1985) 'Expert Systems and Videotext: an application in the marketing of agrochemicals' in 'Research and Development in Expert Systems - proceedings of the Fourth Technical Conference of the British Computer Society Specialist Group on Expert Systems' ed. Bramer M.A., Cambridge University Press. Andrew Basden. ------------------------------ Date: 5 Aug 88 13:47:00 EDT From: "NRL::MEISENBACHER" Subject: Response to - Dave Goldbery's address David Goldbery is now at the University of Alabama Dept of Engineering Mechanics 210 Hardaway Hall University, Alabama 35486 Phone: (205) 348-1618 His dissertation is availabe from Disseratation Abstracts International, 44(10), 3174b. (University Microfilms No. 8402282) ------------------------------ Date: 5 Aug 88 14:17:00 EDT From: Nahum (N.) Goldmann Subject: Response to - RightWriter and Grammatik II (AILIst v8 #27) In the following text I'll give both packages every chance to demonstrate their capabilities. I was using both packages for several years. Both are strong on the evaluation and improvement of readibility (sentence length, average word duration, presence of esoteric words), and not as strong on suggesting alternative grammar rules (several thousand rules in each package are obviously not sufficient for a thorough evaluation of a typical technical text). User interfaces in both packages could also be somewhat improved. I can't comment whether they use expert system technology in both packages, simply because I do not know how one defines what is proper ES and what is not. My impression is that they use something similar. As far as readibility evaluation goes, their results are more or less consistent. In grammar analysis I found them complimentary, with very little correlation between their suggestions for text improvements. I found both packages useful for two reasons: 1. They impose on a writer the discipline of writing prose in short sentences, and help to eliminate complex worlds, making text more readable. This, however, is important only for a short period, after which an average writer does it more or less automatically. 2. They are invaluable in dealing with colleagues and students of mine, who otherwise would not believe that their latest report is totally unreadable (for some reason or another everybody believes it when told by computer!). Overall, both packages provide good value for money (about US$100-125 per package). I have not seen any announcement on Grammatik III so far, but when it comes I will probably order it. Greetings and love. Nahum Goldmann e-mail: (613)763-2329 Analysis by Grammatik II: Subject: In response to Robert Dale's message on RightWriter [#Capitalization : don't mix cases] and Grammatik II (AILIst v8 # [#Capitalization : don't mix cases] 27) In the following text I'll give both packages every chance to demonstrate [#Overstated or pretentious : show or prove] their capabilities. I was using both packages for several years. Both are strong on the evaluation and improvement of readibility (sentence length, average word duration, presence of esoteric words), and not as strong on suggesting alternative grammar rules (several thousand rules in each package are obviously [#Hackneyed, Cliche, or Trite : use this word sparingly] not sufficient for a thorough evaluation of a typical technical text). User interfaces in both packages could also be somewhat improved. I can't comment whether they use expert system technology in both packages, simply because I do not know how one defines what is proper ES and what is not. My impression is that they use something similar. As far as readibility evaluation goes, their results are more or less [#Misused often : use less for nonnumerical quantity, fewer for nu mber] consistent. In grammar analysis I found them complimentary, [#Misused often : this means flattering or free; complementary is completing] with very little correlation between their suggestions for text improvements. I found both packages useful for two reasons: 1. They impose on a writer the discipline of writing prose in short sentences, and help to eliminate complex worlds, making text more readable. This, however, is important only for a short period, after which an average writer does it more or less [#Misused often : use less for nonnumerical quantity, fewer for nu mber] automatically. 2. They are invaluable in dealing with colleagues and students of mine, who otherwise would not believe that their latest report is totally unreadable (for some reason or another everybody believes it when told by computer!). Overall, [#Hackneyed, Cliche, or Trite : total or general] both packages provide good value for money (about US$100-125 per package). I have not seen any announcement on Grammatik III so far, but when it comes I will probably order it. Greetings and love. Nahum Goldmann e-mail: (613)763-2329 SUMMARY FOR gram.out Suspect problems marked: 9 --------------------------------------------------------------------------- Grade School High School College Graduate School 3 4 5 6 7 8 9 10 11 12 Fr So Jr Sr +1 +2 +3 +4 PhD ---------------------------*----------------------------------------------- * - Flesch Grade Level (Reading Ease: 55) Sentence Statistics Number of Sentences: 16 Short (< 14 words): 8 (50%) Average Length: 18.0 words Long (> 30 words): 3 (19%) End with ?: 0 ( 0%) Shortest (# 12): 1 words End with !: 1 ( 6%) Longest (# 3): 47 words Word Statistics Number of Words: 288 Average Length: 5.0 letters Special Statistics (as estimated % of Words or Sentences) Passive voice: 0 ( 0% S) Prepositions: 33 (11% W) RightWriter Analysis: Subject: In response to Robert Dale's message on RightWriter and Grammatik II (AILIst<<*+36. UNUSUAL CAPITALIZATION? *>> v8 #27) In the following text I'll give both packages every chance to demonstrate their capabilities.<<*+17. LONG SENTENCE: 28 WORDS *>> I was using both packages for several years. Both are strong on the evaluation and improvement of readibility (sentence length, average word duration, presence of esoteric words), and not as strong on suggesting alternative grammar rules (several thousand rules in each package are obviously not sufficient for a thorough evaluation of a typical technical text)<<*+17. LONG SENTENCE: 47 WORDS *>><<*+31. COMPLEX SENTENCE *>><<*+39. CAN SIMPLER TERMS BE USED? *>>. User interfaces in both packages could also be somewhat improved<<*+21. PASSIVE VOICE: be somewhat improved *>>. I can't comment whether they use expert system technology in both packages, simply because I do not know how one defines what is proper ES and what is not.<<*+17. LONG SENTENCE: 29 WORDS *>> My impression is that they use something similar. As far as readibility evaluation goes, their results are more or less consistent. In grammar analysis I found them complimentary, with very little correlation between their suggestions for text improvements. I found both packages useful for two reasons: 1. They impose on a writer the discipline of writing prose in short sentences, and help to eliminate<<*+7. REPLACE eliminate BY SIMPLER cut out *>> complex worlds, making text more readable.<<*+17. LONG SENTENCE: 32 WORDS *>> This, however, is important only for a short period, after which an average writer does it more or less automatically. 2. They are invaluable in dealing with colleagues and students of mine, who otherwise would not believe that their latest report is totally unreadable (for some reason or another everybody believes it when told by computer!<<*+17. LONG SENTENCE: 36 WORDS *>><<*+31. COMPLEX SENTENCE *>>). Overall, both packages provide good value for money (about US$100-125 per package). I have not seen any announcement on Grammatik III so far, but when it comes I will probably order it. Greetings and love. Nahum Goldmann e-mail: (613)763-2329 <<** SUMMARY **>> OVERALL CRITIQUE FOR: g:cocos.doc READABILITY INDEX: 12.12 Readers need a 12th grade level of education to understand. Total Number of Words in Document: 290 Total Number of Words within Sentences: 285 Total Number of Sentences: 15 Total Number of Syllables: 499 STRENGTH INDEX: 0.32 The writing can be made more direct by using: - the active voice - shorter sentences - more common words DESCRIPTIVE INDEX: 0.92 The writing style is overly descriptive. Many adverbs are being used. JARGON INDEX: 0.26 SENTENCE STRUCTURE RECOMMENDATIONS: 1. Most sentences contain multiple clauses. Try to use more simple sentences. 14. Consider using more predicate verbs. << WORDS TO REVIEW >> Review the following list for negative words (N), colloquial words (C), jargon (J), misspellings (?), misused words (?), or words which your reader may not understand (?). ACOUSTBNR(?) 1 AILIST(?) 1 COLLEAGUES(?) 1 COMPLIMENTARY(?) 1 CORRELATION(J) 1 DALE'S(?) 1 ES(?) 1 ESOTERIC(J) 1 GOLDMANN(?) 1 GRAMMATIKII(?) 1 GRAMMATIKIII(?) 1 NAHUM(?) 1 READABLE(?) 1 READIBILITY(J) 2 US100125(?) 1 V8(?) 1 << END OF WORDS TO REVIEW LIST >> ------------------------------ End of AIList Digest ********************