Conference rusure::math

188.0. "Math manual comments." by TURTLE::WIENER () Mon Nov 26 1984 13:22

+---------------+

! d i g i t a l !    I n t e r o f f i c e    M e m o r a n d u m

+---------------+





                                Date:  21 November 1984



                                From:  Jeffrey C. Wiener

                                Dept:  Commercial Languages and Tools

                                Ext.:  381-2199

                                Loc.  ZKO2-3/K06

                                File:  HARE::[WIENER.MTHNOTES]STATUS.RNO



Subj: The VAX math manual - This memo contains comments and questions posed

by the math manual reviewers. Also included is an updated distribution list.





The following is a list of comments and suggestions posed for the  new

math  manual.  Several of the questions posed remain open.  Your input

is welcome and desired.







1  COMMENTS AND SUGGESTIONS.





     1.  Add the calling sequence of the routine in the format defined

         in  the  GUIDE TO CREATING MODULAR LIBRARY PROCEDURES and the

         V3 documentation set.



             RESPONSE:  I can add the calling sequences.  To  use  the

         dot  notation,  I  will need to define it in the math manual;

         the dot notation is not being used in the  VMS  documentation

         set  anymore.   This  should not take up much space, the math

         routines use very little of the dot notation.



     2.  Will the math manual be part of the V5 documentation set?



             RESPONSE:  No.



     3.  Who is the writer for the math manual?



             RESPONSE:  At the present time I am the writer.



     4.  Have you looked at other vendor's math manuals?



             RESPONSE:  About two years ago I  looked  at  three  such

         manuals.   I  have  asked  our  library  to locate up to date

         copies of other vendors math manuals.


Comments and suggestions for the new math manual template.        Page 2

COMMENTS AND SUGGESTIONS.





     5.  When you refer to errors, such as an LSB error, refer  to  it

         as a relative error.



             RESPONSE:  This will be done.



     6.  When you report errors in the ACCURACY OF RESULT  section  of

         the  math  manual,  use LSB error distribution, interval LSB,

         maximum relative error, and root mean square errors.



             RESPONSE:  It depends upon my tools.   My  error  package

         gives  only LSB errors (for real valued functions of a single

         variable of type  F,  D,  or  G).   The  Hanek/LeBlanc/Wiener

         package might give me the rest of the information needed.



     7.  How  much/little  detail  will  go  into  the  IMPLEMENTATION

         section of the math manual?



             RESPONSE:    Some   people   thought   that   too    much

         implementation   information   was  presented  with  my  xLOG

         example.  Others felt that too little was  given.   I  prefer

         limiting  the material in this section.  The best way to make

         sure that I include  enough  detail  is  for  all  of  us  to

         carefully proof read the drafts of the math manual.



     8.  Add an algorithm  flowchart  with  to  each  routine.   These

         flowcharts   should   be   similar   to  those  presented  in

         Cody/Waite's "Software Manual for the Elementary  Functions",

         Prentice-Hall, Inc.  Implementation notes should refer to the

         flowchart (see Cody/Waite).



             RESPONSE:  Should flowcharts be  included?   If  so,  how

         much  detail  should  be  given?   Is there a tool that would

         permit me to generate the flowcharts interactively?



         I have flowcharts of each routine.  The flowcharts  are  very

         general, and no implementation notes are given.  If anyone is

         interested, I can  send  you  hard  copies  of  some  of  the

         flowcharts.  Please include your mail stop.



     9.  Is everybody pleased with my xLOG  example?   I  grouped  all

         four natural logarithm routines into one routine entry!  (I'm

         not.)



             RESPONSE:  It  will  take  more  space,  and  I  will  be

         duplicating  a  lot of information, but I think that the math

         manual will be more useful if each data type of each  routine

         is given a separate entry.  So, instead of an xLOG entry with

         ALOG, DLOG, GLOG, and  HLOG,  there  will  be  four  separate

         entries, ALOG, DLOG, GLOG, and HLOG.


Comments and suggestions for the new math manual template.        Page 3

COMMENTS AND SUGGESTIONS.





    10.  What is the overall format for the math manual?  Roughly  how

         many   pages   will   be  used?   How  many  pages  will  the

         implementation section be?



             RESPONSE:  I am working on  the  overall  format  of  the

         manual.



    11.  Why did you group the routines the way you did?



             RESPONSE:  Any suggestions?  More on this matter  when  I

         present the overall format for the math manual.









2  OPEN QUESTIONS THAT COULD USE REVIEWER RESPONSES.



Items 1.6, 1.7, 1.8, 1.9, and 1.11.



Please feel free to comment on any other items,  or  suggest  an  item

that needs to be addressed further.


Comments and suggestions for the new math manual template.        Page 4

CURRENT MATH MANUAL DISTRIBUTION LIST





3  CURRENT MATH MANUAL DISTRIBUTION LIST





! NOTES.DIS

! Math MANUAL distribution list

! Date last updated - 9-Nov-1984

METOO::BRADLEY          ! Chuck

HARE::BAILLIE           ! Linda Baillie         developer

HARE::BUDNIK            ! Ken Budnik            Supervisor

HARE::COWAN             ! Ken Cowan             developer

HARE::LAPINE            ! Matt LaPine           developer

HARE::LEVESQUE          ! Pam Levesque          developer

HARE::WIENER            ! Jeff Wiener           developer

HARE::SCARP             ! Tom Scarpelli         developer

HARE::STAN              ! Stan Rabinowitz

HARE::PETERSEN          ! Don

ORPHAN::MORSHED         ! Farokh Morshed        developer

ORPHAN::BENSON          ! Tom

ORPHAN::Brender         ! Ron

ORPHAN::HARRIS          ! Kevin

GAUSS::ANDERSON         ! Ned

GAUSS::PAYNE            ! Mary

EIFFEL::FALLET          ! Michael

NOVA::KLEIN             ! Steve DBMS_32

ELUDOM::FAIMAN          ! Neil

ELUDOM::NOLAN           ! Chris C-RTL

BACH::GREEK             ! Paul  LISP

NEWTON::EAST            ! Jeff

DELPHI::HAMMERSTROM     ! Paul                  O.S. doc manager

DELPHI::FOGAL           ! Lee                   supervisor

BABEL::RILEY            ! Cilla                 editing manager