home *** CD-ROM | disk | FTP | other *** search
-
-
-
- Comments On
- C2 Support Modules (AI)
- by Software Architecture & Engineering, Inc.
-
- Tool 13_1
- May 21, 1986
-
- COMPILATION
- -----------
- No recompilation of C2 Support Modules (AI) has been attempted to date.
-
-
-
- EXECUTION
- ---------
- No execution has been attempted to date.
-
-
-
- COMMENT
- -------
-
- C2 Support Modules (AI) was developed as a precursor effort for the
- WWMCCS Information System (WIS). An executable version of the tool
- has been demonstrated. However, the tool has not been tested
- since delivery by the vendor.
-
-
-
- Comments On
- C2 Support Modules (AI)
- by Software Architecture & Engineering, Inc.
-
-
- by
- Mars Gralia
- The Johns Hopkins University
- Applied Physics Laboratory
- Laurel, Maryland
- (301) 953-5509
-
- March 26, 1987
-
-
- SUMMARY
- We have successfully compiled and executed the demonstration program of
- the 10 files named PD:<ADA.AI>ALSP*. We have thumbed through the User's
- Manual. We tried several of the examples; all those available via the
- demonstration program worked.
-
- The only problem encountered with this Ada/Lisp package is a minor typo
- in the Users Manual (file "alsp user.doc").
-
-
- REVIEW SCOPE
- We have done little more than mentioned above in the summary. Our
- machine is a VAX-8650 Cluster, running Ada version 1.3-34, and VMS
- version 4.5. We spent about a half hour examining the User's Manual
- (file name "alsp user.doc"). The "pager" file (name "alsp types.src")
- was readily fractionated by Pager, and compiled in the order specified
- (in file "alsp read.me"). Linking was uneventful after we discovered
- its file name. The demonstration program (file "ai types demo") behaved
- exactly as described.
-
- Even though the primary objective of this package is to provide
- Lisp-like functions to a user's Ada program, we have not tried that.
-
- We have examined none of the source code. The demonstration program
- appears complete and robust; perhaps this is indication of high code
- quality?
-
-
- BUG REPORT
- In the User's Manual (alsp user.doc), on page 52, it gives a
- demonstration sequence as:
- -> assert1 ((loves ?x1, ?y1), (friends, ?x1, ?y1))
- (((loves ?x1, ?y1), (friends, ?x1, ?y1)))
- -> assert1 ((wife ?x2, ?y2), (loves, ?x2, ?y2))
- (((loves ?x1, ?y1), (friends, ?x1, ?y1))
- ((wife ?x2, ?y2), (loves, ?x2, ?y2)))
- -> assert1 ((), (wife, Mary, John))
- (((loves ?x1, ?y1), (friends, ?x1, ?y1))
- ((wife ?x2, ?y2), (loves, ?x2, ?y2))
- ((), (wife, Mary, John)))
- and so on.
-
- The bug is a missing comma in the first two "assert1" lines. They
- should read:
- -> assert1 ((loves, ?x1, ?y1), (friends, ?x1, ?y1))
- ^
- -> assert1 ((wife, ?x2, ?y2), (loves, ?x2, ?y2))
- ^
-
-
- AVAILABLE FILES and Reasonable Reading Order
-
- I found this to be the right order to read the files. I have also taken
- the liberty to provide a tiny synopsis of each.
-
- 1.1 Alsp read . me
- A table of contents, with some annotation and the compilation order. It
- says there are generic packages and demonstration programs.
-
- 1.2 Alsp user . doc
- A pretty decent manual for the user. It explains how to run the
- demonstration programs and the background. Its formatted. (The demo
- program info is given in section 7.1. It also gives the compilation
- order.)
-
- 1.3 Alsp tech . doc
- The "top level design spec and final report". It looks pretty
- reasonable. It explains this was developed for Command and Control
- (C2) applications where LISP might be desired.
-
- 1.4 Alsp types . src
- The source code, in "Pager" archive format.
-
- 1.5 Alsp . abs
- This is the full management information ("abstract"): developer (in
- full), scheduled completion of 30 Jun 85, and so on.
-
- 1.6 `Comment' Reports
-
- 1.6.1 Alsp . cmm
- A pro forma remark from the archivist: compilation not attempted.
-
- 1.6.2 Alsp . cm2
- Another comment: "alsp design.doc.1 is a subset of alsp tech.doc.1; it
- should be deleted." They did not compile it either.
-
- 1.7 Garbage
-
- 1.7.1 Alsp ren . sub
- This looks like a VMS command file the developers accidentally sent to
- the archive. Useless to me.
-
- 1.7.2 Alsp src . dis
- It looks like a command file for collecting the pieces of this package
- for submission; useless to me.
-
-
- RECOMMENDED TESTS
-
- I wish I had time to run additional tests. (Instead, I'll assign it
- to my class!)
-
- 2.1 Memory Exhaustion
- A nasty characteristic of many Ada implementations is memory,
- obtained by the "new" construct, is never reclaimed. (And that's OK
- with the Language Reference Manual!) This phenomenon has limited
- the run-time endurance of several of my programs. Is it a limit
- here, too?
-
- A test program might create and delete large numbers of dynamic
- s-expressions, probably running for several hours. If it gets a
- "storage exception", we know this package suffers the same
- limitation. Unfortunately, there is no way to know it does *not*,
- perhaps very slowly, erode memory until there is nothing left.
-
- 2.2 Execution Speed
- It would be interesting to get a feel for the execution time of a
- program which uses these packages.
-
-
- OTHER COMMENTS
-
- It would be much more convenient if the authors would also provide
- the documentation in "source" form. That would allow me to
- re-format it for my size pages; it would look like a real manual.
- Furthermore, I don't care what the formatting language is; its
- fairly easy to convert from one to another with any decent editor.
- After all, I don't want a fancy technical report, just one whose
- page boundaries match the paper.
-
- (I would provide the "source" of this file, but there is none. Its
- done with a very primitive "what you see is what you get" formatter.
- (Where's old runoff when you need it?))
-