Future at Berzerkeley

[Joe Dellinger]

In article <2344 at buengc.BU.EDU> bph at buengc.bu.edu (Blair P. Houghton) writes:
>
>Some sort of cross-referencing, indexing, dewey-decimal systematization,
>etc., would be exceptionally helpful.
>

	In our research group the problem was with tracking down
source code. IE, did someone ever write a Burg deconvolution program
I could use instead of having to write a new one from scratch? It used
to be the solution was to ask the one person who knew everything that
existed on the system. Then he graduated!

	Our current solution is to have everybody put a "keyword"
line into useful source code they might write, for example:

/* keywords: burg deconvolution */

	The first 100 lines of every .c, .f, .r, etc, file in the entire
file system is searched every Sunday morning for such keyword lines,
and a database is created. There is another database which gives synonyms.

	Then you can do "keyword decon", and get something like:

Synonyms for decon:  decon deconvolution whiten pef prediction weiner
/usr/src/our/cube/proc/pefsubs.f        prediction
/usr/src/our/cube/proc/RCtoIF.r         prediction reflection-coefficient 
/usr/src/our/cube/proc/Pef.c            prediction-error 
/usr/src/our/cube/proc/Whiten.c         whiten decon 
/users/joe/Work/Burg/Medc.c             burg deconvolution 

	Hope this idea is helpful.
\    /\    /\    /\/\/\/\/\/\/\.-.-.-.-.......___________
 \  /  \  /  \  /Dept of Geophysics, Stanford University \/\/\.-.-....___
  \/    \/    \/Joe Dellinger joe at hanauma.stanford.edu  apple!hanauma!joe\/\.-._