A-PIRARD@BLIULG11.BITNET (Andre PIRARD) (07/12/88)
I have always been amazed by glossaries being sorted alphabetically. This makes them a real pain to read. Isn't the place of LOOP just after DO, or even better, both described simultaneously with multiple headers. That way, a glossary is much more educational with everything in context. And alphabet is the realm of an index, isn't it? Now, the argument that some words don't easily fall in a category does not hold. More than 90% do and cross referencing will make for the rest. How would you have liked the words of this note sorted? Andr .
olorin@juniper.uucp (David Weinstein) (07/13/88)
In article <8807121636.AA12801@jade.berkeley.edu> A-PIRARD@BLIULG11.BITNET (Andre PIRARD) writes: >I have always been amazed by glossaries being sorted alphabetically. >This makes them a real pain to read. Isn't the place of LOOP just after DO, >or even better, both described simultaneously with multiple headers. >That way, a glossary is much more educational with everything in context. >And alphabet is the realm of an index, isn't it? Agreed. And my favorite, worn, and much battered copy refernce ("Starting Forth" by Leo Brodie [Forth 79 edition]) does just that in Appendix 4. Pity it isn't always done that way. -- Dave Weinstein Internet: olorin@walt.cc.utexas.edu UUCP: {ames,utah-cs,uunet}!ut-sally!ut-emx!{walt.cc.utexas.edu,juniper}!olorin
jb@rti.UUCP (Jeff Bartlett) (07/14/88)
In article olorin@juniper.uucp (David Weinstein) writes: > In article A-PIRARD@BLIULG11.BITNET (Andre PIRARD) writes: > >I have always been amazed by glossaries being sorted alphabetically. > >This makes them a real pain to read. Isn't the place of LOOP just after DO, > >or even better, both described simultaneously with multiple headers. > >That way, a glossary is much more educational with everything in context. > >And alphabet is the realm of an index, isn't it? > > Agreed. And my favorite, worn, and much battered copy refernce ("Starting > Forth" by Leo Brodie [Forth 79 edition]) does just that in Appendix 4. > > Pity it isn't always done that way. > An implementation of '79 that I once ported had the words in 'search order'. Since the dictionary was a linked list, the base vocabulary was ordered by frequency so that the more frequent words would be found earlier. This cut down on the average number of links to traverse during a LOAD. A version of forth for VM/CMS at NCSU had sixteen lists through the dictionary and the words were hashed to find which list to traverse. This was a port that started from an Rockwell AIM 65 evaluation kit and changed to 32 bits. It also had full access to the system. Later applications included a full screen editor and LISP. (Oooo, a prefix language implemented in a postfix language, brain aerobics :-) Jeff Bartlett Center for Digital System Research Research Triangle Institute jb@rti.rti.com ...mcnc!rti!jb
thomson@utah-cs.UUCP (Rich Thomson) (07/14/88)
In article <8807121636.AA12801@jade.berkeley.edu> A-PIRARD@BLIULG11.BITNET (Andre PIRARD) writes: >I have always been amazed by glossaries being sorted alphabetically. >This makes them a real pain to read. Isn't the place of LOOP just after DO, >or even better, both described simultaneously with multiple headers. > >Andr . I would like to see the FORTH standard words arranged as a hypertext structure so that I could view documentation for the words on-line and see related words easily. I wish I had a function oriented glossary for Multi-Forth on the Amiga. It would make it alot easier to remember the string word that does what I want, etc., etc. I suspect that the primary reason for alphabetic glossaries is simply laziness. -- Rich -- Rich Thomson, Oasis Technologies, 3190 MEB, U of U, Salt Lake City, Utah 84112 (801) 355-5146 thomson@cs.utah.edu {bellcore,ut-sally}!utah-cs!thomson Alcohol: the drug of availability