.Open RJBottings Research Web Log Number 4 . Mon Jan 3 08:21:53 PST 2005 Archived this Blog For the later entries see .See http://www/dick/blog.html my current blog. . Mon Dec 20 08:10:09 PST 2004 End of Quarter Grades all posted, web sites .See http://www/dick/cs320/ .See http://www/dick/cs375/ and syllabi .See http://www/dick/syllabus.html for next quarter's class ready... And brought my .See http://www/dick/symphony.html up to date. . Tue Dec 14 07:24:51 PST 2004 Please have a Fortune Cookie! Finally enought time to translate a broken CGI into PHP: .See http://www/dick/cookie.php which picks out a randomly chosen fortune cookie from the famouse Cyber cookie jar implemented on the CSU mainframe network of the 1980's. . Mon Dec 13 15:48:45 PST 2004 Bakers Dozen of new items for bibliography See .See http://www/dick/newb2004/newb1213.html for stuff on networks, project risks etc. . Tue Dec 7 14:49:23 PST 2004 Review Published My review of $ArlowNeustadt04 has been published as review#0412-1461 in $CR V45n12(December 2004) pp791-792. . Mon Dec 6 16:15:26 PST 2004 Graphics Here is an odd thing... I've been proof reading a chapter I wrote on the co-evolution of SSADM and UML for publication some time `real soon now`. And the editor asked me particularly to check the figures in case the page layout program had changed the figures. It had. In every diagram one or more items were displaced about 1/8th of inch to the left. The looked ok when viewed by MS Word. It seems likely that grouped objects are handled strangely by it and the page layout program does it differently. So I tried to read one of the figures by half-a-dozen different programs and something went wrong with every one. For example Corel have QuickView program that can not disply the figure. Wordperfect gives up. MS WordPad objects to the figure several times, on Linux Kword displays a blank screen and Open Office shows all the objects in a pile at the bottom.... Looks like MicroSoft have done a pretty thorough job of forcing me to use their product to display and print figures prepared in MSWord. . Mon Dec 6 07:22:11 PST 2004 Doubts I've always liked the idea of using logic to document systems. It turns out the Microsoft is backing an approach to data base design that is based on Formal Predicate Logic. This is the .Key Object Role Modeling method of Terry Halpin. Its exciting to hear of a tool and praise from users. However it casts doubt on some of the research I documented back in the 1980's on using logic. For example .As_is Use predicates not relations and sets and maps. and .As_is Divide up the universe of discourse into types of objects. .See I'm tempted to reconstruct the whole of my notes on Logic .See http://www/dick/maths/logic_10_PC_LPC.html .See http://www/dick/maths/logic_11_Equality_etc.html but first I'll need to review th consequences... . Tue Nov 23 15:21:56 PST 2004 A bakers dozen bibliographic items Uploaded .See http://www/dick/newb2004/newb1122.html a batch of new items including one on ways to encourage creativity when developing (inventing!) requirements. Also Scott Ambler's definition of " .Key normalized documentation " where information is recorded once, and once only... so a business rule may be expressed in code and visible as a decision table in another document..... but it only stored in one place. . Fri Nov 19 17:42:28 PST 2004 Another CGI bytes the dust Gack...... sorry about the pun. I've translated my `kgrep` shell script into PHP and am testing it in .See http://www/dick/cs360/notes/09.commands.html#Questions where you select a question mark to see if that is after a good answer to the question or not.... . Sat Nov 13 06:57:30 PST 2004 Experiment withhighlighting Key phrases in a paragraph Inspired by Dr. Nielsen's .See http://www.sun.com/columns/jakob/ observation that people do not normally read web pages I'm experimenting with the ability to include .Key a key idea in each paragraph. This will be highlighted () and made into a name (...) in HTML. In turn, this means the key phrase can be used as an internal and external link. Finally, my search engines rely on HTML names to extract significant content in a document. The syntax in my $MATHS language is (as usual) trivial. Include a like line this: .As_is .Key an important phrase or word as part of the paragraph. Semantics: no formal meaning, but should be rendered to attract the attention of readers and search engines. The code seems to work... Next to see to see if I can create a link to the key .See a key idea in the previous paragraph. I've noticed that .Key new features are often used in strange ways so I'll now have to see how I use it. . Mon Nov 8 15:38:46 PST 2004 New items New: .See http://www/dick/newb2004/newb1108.html that includes the thought that Business Process ReEngineering fails becaus people are expected to fit the process rather than the other way round. . Fri Nov 5 09:47:15 PST 2004 Lack of sleep leads to strange behavior Somehow got to work unscathed. Scared a few people... including me. Posted grades and a proposal for a data modeling profile for the UML: .See http://www/dick/cs372/14.html Then added a feature to my search engine to output the keywords I've attached to bibliographic items. Please try at .See http://www/dick/bib.php and Email me if you hate it or like it a lot. Time for a meeting.... and then down a tunnel, and then to a demo, and then to another meeting. Ah, the fun of an academic life! . Fri Oct 29 13:04:09 PDT 2004 Paper accepted for SIGCSE Presentation My paper on teaching Ethics .See http://www/dick/papers/index.html#2004 has been accepted. I've controlled my enthusiasm having scare a neighbor's cat on getting the EMail... Now to revize it, and finish a review for money, and draft some papers for OOPSLA, and propose a special SIGCSE teaching tips session on using cards to introduce algorithms and data structures, and... something on some research on teaching and learning the UML: .See http://www/dick/cs690/20041105Farquhar.txt . Mon Oct 11 06:57:58 PDT 2004 Babel 17 Spent part of the weekend re-reading Samuel Delany's `Babel 17`. I first read this way back in the 1960's. One of the few SF novels that refers to Fortran and Algol, and has interesting characters. I noted the following quote: .Set It's the most analytically exact language imaginable. But that's because everything is flexible, and ideas come in huge numbers of congruent sets, governed by the same words. This just means that the number of paradoxes you can come up with is staggering. .Close.Set Delaney went on to write a fantasy set in a time when Goedel's Theorem has had wider implications than Einstein's mathematics... . Sat Oct 9 08:02:58 PDT 2004 Experiment: no indexes on translated MATHS pages Since the very first tool to translate MATHS into other forms an index of defined terms has benn provided at the end of the translated document. In HTML the terms and labels became links into the file. I've modified the mth2html tool to not do this. If no one complains (and I like it) this change will become permanent. . Thu Oct 7 08:34:32 PDT 2004 Scheduling is a Hard Problem (1) trying to find a gap in 3 faculty schedules and a student's schedule takes abotu 3 weeks of Email... There has to be a better way! (2) Developed a personal schedule as an HTML table .See http://www/dick/plan.html please let me know if it doesn't make sense to you! . Wed Sep 29 17:17:13 PDT 2004 Submitted Review Chen vs SSADM experiment Submitted a review of "Comprehension of diagram syntax: an empirical study of entity relationship notations" Purchase H., Welland R., McGill M., Colpoys L. International Journal of Human-Computer Studies 61(2): 187-203, 2004. Note: Wed Nov 10 16:40:05 PST 2004. Online published CR# CR130246. .See ./blog.html#2005-03-10 Thu Mar 10 11:03 Review Published I was on the team that designed SSADM and I'm pleased to see we did something right... even if the test is biased in the direction of SSADM and away from Chen's ERD notation. . Fri Sep 24 12:45:41 PDT 2004 five new items See .See http://www/dick/newb2004/newb0924.html for five new items. Three are from the procedding of the 2004 "ISSTA" International Symposium on software testing and analysis. One shows the kinds of file that harbored bugs in one large project. Who would think that makefiles would be a good place to look! ANother shows how highly abstract requirements can be proved to have problems. Meanwhile meditating on the difference between evolutionary projects and revolutionary projects. . Tue Sep 14 11:01:18 PDT 2004 Mutilated Chessboards Not a curse! But I've just had Review number 0409-1085 published in the September Computer Reviews journal. . Tue Sep 14 10:17:36 PDT 2004 Back in office Updating a Windoze laptop, synching the Palm Pilot Tungsten, collecting Buddy the Plant, .... Upload a dozen new items of software development research: .See http://www/dick/newb2004/newb0914.html that includes a sad result for Object-Oriented fanatics: .See [ArisholmSjoberg04] show that a `good` OO design may confuse less experienced maintaners more than a `bad` (centralized) design. I guess this is good news for OO consultatns and trainers since it implies that everybody in an organization needs to be trained thoroughly in understanding `good` OO designs. Meanwhile, here is the citation for the Lewis Carrol monograph. It was written in Guildford public library on my Palm and just uploaded .List Lewis Carroll (ed. William W Bartley III) Symbolic Logic Pts I & II Harvester Press 1977 Hassocks Sussex England =MONOGRAPH LOGIC syllogisms SORITES EXAMPLES Pt II uses something like resolution and semantic tableaux! .Close.List . Mon Sep 13 10:45:10 PDT 2004 Back from the UK I just had a smooth flight back from LHR yesterday. Oddly the new security systems seem to make getting into the USA faster than previously (at least for resident aliens and BEMs). I had planned a week of friends and family with no computer science. All went well until I needed to check part of a railway journey that involved some walking.... on a day forecast to be rainy. The reference librarian at my old home town of Guildford was most helpful and found the relevant map with out leaving her window (Scott Adams is right: reference librarians know `everything`!) The walk was half-a-mile which is easy for me except if I hit a downpour. These are not uncommon in southern England at this (or any) time of year. So I knew I would be wise to change my route.... Out of nostalgia I returned to the Dewey classification of my childhood obsession with mathematics -- they have moved it and pruned it. Where has "Principia Mathematica" gone. And for that matter how come a typical county town should have such a very good mathematics collection just when I needed it? What I found this time was a previously unseen edition of Lewis Carrol's "Symbolic Logic". This edition has both parts. It is quite clear that under the heading of Aristotelean logic, Carrol was inventing what I call "Semantic Tableaux" as a way to analyze the possibilities in a series of assumptions: .See http://www/dick/maths/logic_2_Proofs.html#Semantic Tableaux I plan to get a copy and document the history, and then for the fun of it, take some of Carrol's examples, and solve them as a sample of my approach. What has this to do with computer science: (1) requirements contain hidden contradictions and consequence that need to be discovered before they become misfeatures. (2) Proving that something is true and/or works can be done by using symbolic logic -- if you have the time and want to be very sure you've done something correctly. Here is a translation of Carrol's `Barber Shop` problem in the second book: .Let (1): There are three processes A, B, and C. (2): At any time, each is either running or halted. (2b): At any time, No process is both running and halted. (3): If C is running then, if A is running then B is halted. (4): If A is running then B is running. Is it possible, necessary, or impossible for C to be halted? .Close.Let . Thu Sep 2 11:55:58 PDT 2004 Ready for a break Spent some time this morning improving my notes on concurrency .See http://csci.csusb.edu/dick/maths/math_76_Concurency.html .See http://csci.csusb.edu/dick/maths/math_73_Process_algebra.html and .See http://csci.csusb.edu/dick/maths/math_14_Dynamics.html I planning to take the next week off with a trip to England to visit my white haired mother and some friends. Also to sort out some things to do with retirement. I'm hoping for a break from academic things like publications and teaching. . Tue Aug 31 06:29:14 PDT 2004 Prep I've been preparing for the new quarter .See http://www/dick/cs202/ .See http://www/dick/cs372/ and a trip home to England. But I was able to chase down a paper .See [MockusFieldingHerbsled00] analyzing how the Apache server was developed. A very different model for successful software development. The "sweet spot" for an Apache style process seems to be when you have a core of expert developers, a discipline of coding, a large number of users who can code, and an even larger number of users willing to use the developing software, all communicating by Email and the Web. According to another study .See [Worley04] the users don't have to be computer geeks. For more on the Apache project look up .Find APACHE in my bibliography. .Open Wed Aug 25 12:47:13 PDT 2004 AmblerS04: new item Scott W Ambler Agile Database Techniques: Effective Strategies for the Agile Software Developer John Wiley & Sons 2004 ISBN 0-471-20283-5 =UNREAD AGILE DATA Basis on the web at .See http://www.agiledata.org/ .Close Meanwhile working on notes on documenting ontological hierarchies by using conditional documentation: .See http://www/dick/maths/notn_13_Docn_Syntax.html#Conditional documentation (Partly Baked Ideas). . Mon Aug 23 09:22:25 PDT 2004 DFDs and Ethics Posted a draft of a paper on teaching Ethics and some notes on expressing DFDs in the UML. See .See http://www/dick/papers/index.html#2004 for more. . Fri Aug 20 10:52:41 PDT 2004 Correspondence and Catch Up Shared some thoughts about UML and DFDs with Scott Ambler. Waiting to read his new book: I want to know how one can easily change a live data base as part of a agile process: suppose that 10,000 students all need a new Student Identifier added to their records? The Big Design Up Front texts on data base don't talk about this kind event! Brought my monograph on methods up to date with agile and extreme methods. I have a bad feeling that rather than being a useful way to reach for better ways to produce software (my original plan for the book) it is rapidly becoming an archaeological document suitable for "Raiders of the Lost Art". If you want a view of how we got where we are see .See http://www/dick/monograph/ and be charitable. . Thu Aug 19 10:06:24 PDT 2004 Digging in the library Pleasant day yesterday looking into .See [KarpMiller66] which was an early model of highly parallel computation. An elegant idea: a set of nodes that hold functions, and connected by queues of data items. Each node must wait until a certain number of items (depending on the queue) is waiting for input and then it inputs a fixed number (depending on each queue) from each input. It applies the function and then, at an unknown time later outputs a fixed number of items to each output queue. These rules mean that the graph always does the same calculations.... but in different orders. There are some very elegant arguments about the behaviors the graphs exhibit and an example of solving a Partial Differential Equation. You can show (I figure) that the function can have a state and so be a restartable process of some complexity. Then, at home, noticed that the nodes can not choose where to get there input (obvious), and so you can not have one that does a collation of merge operation on two (or more) ordered streams. This is a very useful for expressing some solutions to problems (including a lot of Data Processing). In the morning I realized I'd seen a similar restriction some where else, and found it in .See [CommonerHoltEvenPnueli71] defining "Marked Directed Graphs". So it looks as if the two models are equivalent at some level. Now this suggests that to get a complete theory of DFDs one is forced to go colored Petri Nets of some kind or other. 12:30:04 PDT 2004 Digging into these pages uncovered about 15 items that had vanished from my bibliography... luckily I keep backups and was able to replace and update them. . Wed Aug 18 09:36:14 PDT 2004 Better than QuickSort?!? Another batch .See http://csci.csusb.edu/dick/newb2004/newb0817.html of a dozen or so bibliographic items. It includes an answer to a question that has bugged me for a decade or so: "Is there a variation of quicksort that has both good average and good (O(n log(n))) worst case performance? .See [Chen04] has developed a proportional partition sort and implemented in C with some of the speeding up tricks use in qsort. The author claims that both the average and worst case performance in O(n log(n)) and has empirical data showing that `psort` is faster than qsort. . Thu Aug 12 13:02:52 PDT 2004 More reading... Worked through a backlog of 3 years of ACM SIGPLAN and extracted a couple of relevant papers.. plus stuff in IEEE Computer: .See http://www/dick/newb2004/newb0811.html and meanwhile an odd pattern in my other reading: I'm reading a book on the history of mathematics and another book on the mathematics of history.... hmmmm. . Wed Aug 11 07:30:39 PDT 2004 A Quick set of edits of a page on sets Decided to review .See http://www/dick/maths/logic_30_Sets.html which lays down a set of rules for using sets in ASCII documentation: .As_is |- administrator ==> user. .As_is |- dick is_a administrator. .As_is ()|- dick is_a user. Reorganized and spell checked the page. Yesterday San Bernardino hit 106.F. Expecting another day with temperature above 100.F. . Mon Aug 9 11:23:09 PDT 2004 Paper Publication of Review $CR V4n8(August 2004)published review #0408-0939 of "Clustering for Petri nets" By W. Keller, Theoretical Computer Science 308(1-3): 145-197, 2003. . Mon Aug 9 09:56:25 PDT 2004 Summer reading? Just uploaded a dozen or so bibliographic items .See http://www/dick/newb2004/newb0803.html including one joke .See [Ovans04] about life-cycles. Meanwhile much domestic sorting out. Bought a keyboard for my Tungsten E, which has full size keys(good) but only 3 rows of keys(weird). I've continued to correct the errors in .See http://www/dick/maths/logic_8_Natural_Language.html and think about reorganizing it. . Mon Aug 2 09:16:58 PDT 2004 Summer time and the reading is easy I've drafted a paper on my last "Ethics and Professionalism" course that went well. Next step is to format, categorize and submit to SIGCSE05. Meanwhile, some work done on ontologies in $MATHS .See http://www/dick/maths/logic_8_Natural_Language.html and .See http://www/dick/samples/index.html#Ontologies the SUMO project. Just added a large number of links to my XML page .See http://www/dick/samples/xml.html taken from Mark Doernhoffer's SIGSOFT articles. . Fri Jul 2 13:44:31 PDT 2004 Biting the Bullet I decided it is time to bring my definition .See http://www/dick/maths/notn_00_README.html in line with my actual practice. In particular trying to develop a stronger `Document Object Model` for $MATHS. The original ideas involved documents that had a mixture of formal and informal parts and ways to refer and reuse previous documentation. When the Web popularized Hypertext it became clear that hyper-links could do a good job for linking a formal definition to a use of definition. After all why should you reinvent a 200-year-old theory when you could just link to the theory. I also had defined a way to structure documents and developed tools to generate contents lists and indexes. I found I had a useful way to put ideas on the web. This Blog is written in the $MATHS notation and you are probably reading the HTML that is generated from it. So is my $bibliography and most of my teaching materials. But in the process the usage has drifted away from the formal rules. Also, oddly, unexpected errors have appeared in the the original documentation. So this summer is a chance to make these changes to files with names like ..../dick/maths/notn_*. . Tue Jun 29 10:25:34 PDT 2004 Another batch of interesting items Slowly catching up with my reading... ACM TOSEM and Software Development Magazine. I always like the Halloween stories .See [EvansEtal03] in Software development and Ivar Jacobsen has explained .See [Jacobsen03] how use case extensions help to separate concerns, but demand some kind of Aspect Oriented Programming. For more see .See http://www/dick/newb2004/newb0629.html (latest items). . Wed Jun 23 13:44:41 PDT 2004 Updates to Number Theory and Samples I've managed to make some slight improvements to my rough notes on number theory .See http://csci.csusb.edu/dick/maths/math_42_Numbers.html#Divisors%20and%20Factors but more could be done. Also added some links on .See http://csci.csusb.edu/dick/samples/index.html#Ontologies under my samples of $MATHS documentation and listing of links for software developers. . Tue Jun 22 13:43:07 PDT 2004 Gathered a good crop of references Discovered that the Tungsten E preserves tabs in notes attached to Tasks/Todos but not in WordToGo or Memos. Programmers do not think clearly. Meanwhile, read some interesting things .See http://www/dick/newb2004/newb0619performance.html about estimating the performance of software and real time systems. Also some .See http://www/dick/newb2004/newb0619.html more general refs. . Wed Jun 9 17:12:09 PDT 2004 Finals ready... and new items... Busy weeks teaching and medical things, and a nice trip up to San Luis Obispo. Also forced to upgrade from Palm III to a Tungsten E at great speed. Still suffering from losing Gr^f|7| version 1. Also from the loss of the Tab character on the DeskTop Memos. URG. Uploaded .See http://www/dick/newb2004/newb0609.html that includes .See [Jorgenson04] explaining that the quality of estimates improves if you ask for information in a special way: don't ask what the worst possible value for finishing a project -- the reply will be optimistic. Instead ask what the probability is that the estimate is only half of the actual time. Meanwhile: paranoia is easy when you're locked door can be pushed open, and when 50+ machines on campus start to transmit movies with out their users knowing about it! . Mon May 17 15:50:19 PDT 2004 Requirements: straws in the wind Just docketed half-a-dozen items on requirements .See http://www/dick/newb2004/newb0516.html from the IEEE Software Magazine. Other items include a proposal to undo all the object oriented thinking and practice: separate data from function etc. This morning discovered that I or some accident had wiped out my syntax definition of the language Z. Now have to decide whether to reconstitute them from scratch. I guess I may find a copy on a floppy somewhere. I found a spare copy of the HTML and reversed engineered the source code .See http://www/dick/samples/z.syntax.mth and then regenerated .See http://www/dick/samples/z.syntax.html the HTML. ARRRRRGH. . Wed May 12 AM PDT 2004 Submitted a review to Computer Reviews .Open ArlowNeustadt04 J Arlow & I Neustadt Enterprise patterns and MDA: Building Better Software with Archetype Patterns and UML Addison Wesley Longman Publishing Co, Inc., Redwood City, CA, 2003. =HANDBOOK COMMERCIAL ORGANIZATION LITERATE UML MODELS MDA PCL archetype pleomorph MDA::OMG="Model Driven Architecture", using tools to convert models directly to code. PCL::=" Pattern Configuration Language". archetype ::= "a class, attribute, operation, or relationship that appears in many applications" ArlowNeustadt04_pattern ::= "A collection of collaborating archetypes", optional parts shown by the <> stereotype. Names: Party, Party relationship, Customer Relations Management(CRM), Product, Inventory, Order, Quantity, Money, Rule. pleomorph ::= " distinct patterns of archetypes in different enterprises". .Close . Tue May 11 09:17:46 PDT 2004 Not a good month but... Preparing labs and work and teaching is taking a lot of time with two new preps this quarter. Also more colds and diseases. Add less than perfect systems... However the latest reading .See http://www/dick/newb2004/newb0507.html includes evidence .See http://www/dick/newb2004/newb0507.html#LutzMikulski04a .See http://www/dick/newb2004/newb0507.html#LutzMikulski04b that many surprises and anomalies in NASA software are systems problems: bad procedures, missing or misplaced data files, and bad documentation... Perhaps it is time to return to my work on applying formal methods at the system level. . Mon Apr 19 14:36:35 PDT 2004 More items for bibliography and rebuilt lab/search page New items: .See http://www/dick/newb2004/newb0417.html and rebuilt .See http://www/dick/lab.html for searching for stuff on my web site. At long last someone has come up with a clearer way to approach asymptotic Big-O notations: .See http://csci.csusb.edu/dick/newb2004/newb0417.html#Rousos04 via equivalence class of functions that have similar growth patterns. Main weakness is that the definitions only work on monotonic increasing functions. . Sat Apr 17 07:42:52 PDT 2004 Experimenting with CSUSB Calendar .See http://cs1.csusb.edu/command.shtml?view=overview&calid=rbotting&tzid=America/Los_Angeles&security=1 . Mon Mar 8 14:32:03 PST 2004 New bibliography items: COBOL lives Just posted a fresh batch of new items .See http://www/dick/newb2004/newb0306.html and noticed to distinct items providing evidence of COBOL being alive and well: .See [DatteroGallup04] found it in the top 15 most used languages, and .See [AntoniolEtal04] developed a staffing model that fitted a real Y2K project very nicely. . Fri Mar 5 12:00:49 PST 2004 Review published On March 3rd $CR published my review of "Clustering for Petri nets" By W. Keller, Theoretical Computer Science 308(1-3): 145-197, 2003. I forgot to note that I had another review published in $CR online on December 31st: "Behavioral subtyping relations for active objects", by H Wehrheim, Formal Methods in System Design 23(2): 143-170, 2003. . Wed Mar 3 11:07:47 PST 2004 Regular Expressions and PHP Decided to handle the pain of having to use MS Word to prepare a chapter for publication by extending my PHP searches to use the regular expression functions. Lesson Learned. Pattern matching an HTML file causes unexpected false drops. For example searching for "Di.*stra" (case insensitive) found an occurrence inside an anchor: .As_is Therefore: only the bib.php script was upgraded. .See http://www/dick/bib.php Next step is to link the search page to my documentation on how to use regular expressions .See http://csci.csusb.edu/dick/samples/egrep.html#Main%20Content and then revise it. On the other hand I have indexed my home directory and some subdirectories so .See http://www/dick/lookup.php will find way too many references to "Category" if asked! . Fri Feb 27 17:46:29 PST 2004 Improved Bibliographic Search By writing a script that extracts a single item from the 1Mb collection you can quickly see the details of what you've found. Here is a simple and egotistic test: .See http://www/dick/biba.php?search=dick for example. When you search for a string a list of links appear, click on one of them to extract that item: .See http://www/dick/bib.php For a list of interesting searches, see .See http://www/dick/samples/subjects.html which uses the new PHP scripts. The next step is to extract the details on a list of items. Then to add some the functionality I developed in Hypercard such as intuitive Boolean searches and formal concept analysis. . Fri Feb 20 16:22:26 PST 2004 Developed a Faculty Vignette .See http://www/dick/vignette.html . Fri Feb 20 12:08:11 PST 2004 Recent Activities My chapter mapping SSADM into the UML has been resubmitted for final editting.... if successful I will announce here. Otherwise I'll post it in PDF in my $papers. Also submitted a review of a theoretical paper on Petri nets that made extensive use of category theory. This lead to some small improvements to my notes .See http://www/dick/maths/math_25_Categories.html on categories And then volunteered to review a paper on mutilated chessboards. . Wed Feb 18 11:41:25 PST 2004 Graduate Projects I'm opening up a directory for information on Graduate students work that I am advising: .See http://www/dick/cs690/ to hold announcements and abstracts etc. . Mon Feb 16 09:46:24 PST 2004 Updated Vita I've been nominated to `Who's Who in America's Teachers` for the second time: .See http://www/dick/short.vita (txt file). . Fri Feb 13 09:39:26 PST 2004 New search In .See http://www/dick/samples/ I have a collection of sample pieces of documentation. I wrote them for three purposes. (1) I need to make notes on something I was learning for example VRML or Javascript. (2) I wanted to test the notation I invented for documentation. (3) I want to promote the use of mathematical and logical notations in software development. As a result I've created a large knowledge base on a wide range of formal languages, methods, ideas, etc. One of the advantages of my notation is that it maps fairly easily into Hypertext and significant terms are easy to extract and enter in to a consolidated index or dictionary of terms. Today I recreated the search feature that died in September. It will scan the defined terms etc in samples. Please click on .See http://www/dick/samples/lookup.php and input some term in software development that interests you. . Mon Feb 2 08:22:44 PST 2004 treading water Spent the weekend reviewing Category Theory fro a review I've volunteered to write. Also thinking about the process of moving from a static domain model to the dynamic/active object-oriented model of an executable component. Partly for my CSCI375 Requirements Analysis class and partly for my chapter on SSADM and UML. Decided to not confuse the issue by writing about the categorical semantics of SSADM domain models (LDSs). Uncovered a problem in my tool for translating formal MATHS documents into HTML. On some versions of Linux an error is reported: .As_is sed: -e expression #1, char 4047: Invalid reference \2 on `s' command's RHS The script has several `sed` commands, and each has many `s` substitutions. Extracting all the usages of "\2" (about 40 lines...) and then removing all 's' commands that don't define two or more '\(...\)' leaves me with nothing. And the error occurs with an empty input file. What is irritating is that I've seen this misfeature before and can't recall how I worked round it at that time. I've checked back to a Dec 22nd version and found it there. This just in -- the error is reported on different characters depending on the input data: .As_is sed: -e expression #1, char 4095: Invalid reference \2 on `s' command's RHS .As_is sed: -e expression #1, char 4047: Invalid reference \2 on `s' command's RHS .As_is sed: -e expression #1, char 4083: Invalid reference \2 on `s' command's RHS 09:09:25 More... the above effect is because the 'sed' command contains the name of the file being processed in it. Now that I've extracted the bad command and run it in isolation... Next step remove the variables from the expression and try to isolate the problem string. 09:20:39 Looks like the error is caused by having a long expression. Long being about 3712 characters (plus or minus 50). This smells of a buffer problem in the sed code. Probable introduced after April 2003 in Red Hat... Plan: I'll have split the `sed` in two. And that will take thought to re-engineer the logic of the script. If you interested a similar problem is occurring with the July 07 2003 version that is published as .See http://www/dick/tools/mth2html but at a different point in the expression (of course). . Fri Jan 30 16:47:27 PST 2004 Started in the long bibtrek In other words, I'm starting to redevelop my bibliographic search engine .See http://www/dick/bib.php without using any of the UNIX power tools. So far it scans the source code for the $bibliography and extracts a list of the identifiers of keys that uniquely identify the items that match. It makes these into links that can be clicked. Three lines of development: First, the search demands a perfect match and is case sensitive. In the past it could be any `awk` search string. Second, I plan to place the extracted keys in a text box so that in a form that permits boolean operations on search results. Third, I can develop a faster way to link to the found items. Enough for a Friday night... . Tue Jan 20 14:17:33 PST 2004 On the state of the Practice Some interesting items about what software developers really do .See http://www/dick/newb2004/newb0119.html and don't do. . Tue Jan 13 16:33:39 PST 2004 Access Attended a workshop on Web Accessibility. Have a list of about 8 "ToDo"s as a result. For example a tiny downward pointing arrow will be starting to appear in my pages that skips the navigation and cuts to the main content of the page. Similarly if the page has a reference to a PDF file like .See http://www/dick/syllabus.pdf then an extra link to the Adobe reader will be added. . Mon Jan 12 13:56:10 PST 2004 Software Engineering Foundations vs Practices Wound up ESEC and FSE papers and started into a special issue of IEEE Software engineering Magazine on "The State of the Practice": .See http://www/dick/newb2004/newb0112.html including Capers Jones observation (from an 18 year sample of projects ) that no method or language determines success but good quality control does... . Tue Jan 6 12:08:28 PST 2004 Items from ESEC etc Also opened up a directory for new bibliographic items: .See http://www/dick/newb2004/ with .See http://www/dick/newb2004/newb0101.html as the inaugural item. . Mon Jan 5 16:07:54 PST 2004 Items from ESEC and FSE 2003 Added a batch of items from a joint ACM/IEEE Conferences on software engineering including .See http://csci.csusb.edu/dick/newb2003/newb1230.html#HerbsledMockus03b on modeling the intertwinglement of decisions in software developing. . Sat Jan 3 08:10:21 PST 2004 New Format Finally bit the bullet and changed the 6-year old formatting of bibliographic items in $bibliography. A matter of editing more than 3,500 items. The original format defined an item as a set of strings. A mathematical set -- that is. So, no repeated items and no order. It is easy to edit a set by removing and inserting elements: example'= example~{old} | {new} But there is no way in a set to be sure which item is the list of authors and which is the title of the paper or book. For example a autobiography by "Joe Doe" may be entitled "Joe Doe". A dumb but subtle error in logic like this can remain hidden in a mathematical document for years. So, one needs independent review of all mathematical documentation. One might argue that the open source community have rediscovered or applied the precise process that makes science work: Publish or Perish. In the new format each item is a `section` with a header and a sequence of `paragraphs`. The first is the list of authors(who), the second is the title(where), and the third is where to find it. Then follows a list of subjects and then any notes I feel like adding. I can now specify a change of author by: item["Ackenfuss11"].paragraph(1)'="Isambard K Ackenfuss". Next step is to make sure the formal definition of the $MATHS language supports this, and then develop PHP pages to retrieve and search for items. . Thu Jan 1 09:17:56 PST 2004 A brand new year See $Previous for previous items. Have just moved many collections of bibliographic items on software development to .See http://www/dick/newb2003/ in preparation for 2004's batches... and tried to update any links. Web Site Goals for 2004: .List Keep course work and teaching information up to date. To repair the various searching tools for this site mainly by transmogrify CGI shell scripts into PHP. Work on improving the infrastructure: format of source code for the bibliography for example. Revise the syntax and semantics of the source code ($MATHS) and update tools to match. Continue to accumulate the most practical theories and the soundest experiences on software development on this site Construct an online compiler for translating $MATHS to HTML. .Close.List . Purpose The main purpose for this Blog is to record the development of my software development web site. This site is intended to become a place to visit if you want to find computer science and software engineering research that might be useful, or information on current practice that is likely to be good. But it also includes information on jobs, $people (including companies and societies), advertising, etc. Its strongest focus is on $methods and $languages. The key search tool ($lab) will scan the whole site and an extensive bibliography of published books, articles, and research papers from many sources once I get it repaired. . Previous (December 2003): .See http://csci.csusb.edu/dick/blog003.html (July 2003): .See http://csci.csusb.edu/dick/blog002.html (June 2003 and before): .See http://csci.csusb.edu/dick/blog001.html . Latest .See http://csci.csusb.edu/dick/blog.html . Glossary and Links bibliography::=http://www/dick/newbib.html, (source .See http://www/dick/newbib.mth ) a large collection of publications on software development. CR::=http://www.reviews.com/home.cfm, Computer Reviews, a journal reviewing books, papers, etc. about computers. A prime source for the serious researcher in computer science. etc::=http://www/dick/samples/etc.html, a collection of links to sites for software developers, engineers and researchers. lab::=http://www/dick/lab.html a place to search for data on my site. Badly damaged in the later half of 2003. languages::=http://www/dick/samples/languages.html, information on computer languages. latest::=http://csci.csusb.edu/dick/blog.html, MATHS::=http://csci.csusb.edu/dick/maths/, a language for semiformal documentation like specifications, that is also good for weblogs, essays, lecture notes, etc. etc. methods::=http://www/dick/samples/methods.html, links and definitions about software development methods and processes, plus some jokes. Also see .See http://www/dick/samples/methods.glossary.html instead. monograph::=http://www/dick/monograph, a study of software development methods 1940-1990 attempting to show how simple mathematics can remove many errors from software. papers::=http://www/dick/papers, pre-publication drafts, local seminars, unpublished essays, etc.. people::=http://www/dick/samples/people.html, samples::=http://www/dick/samples/, samples of documents prepared using $MATHS. se::=http://www/dick/samples/se.html, links to things about software engineering and software development. source::=http://www/dick/blog.mth, I use my own $MATHS language to write these blogs. standards::=http://www/dick/samples/standards.html, STANDARD::=http://www/dick/maths/math_11_STANDARD.html, my personal standard definitions for $MATHS. subjects::=http://www/dick/samples/subjects.html, tools::=http://www/dick/samples/tools.html, Z::=http://www/dick/samples/z.html, specification language. .Close RJBottings Research Web Log Number 4