🚀 go-pugleaf

One problem which this discussion does not seem to have faced
squarely is the problem of identifying the use of a patented
technique in software.  One earlier article suggested that
an expert would have a huge data base containing all things
patented in software or in some other field applicable to
software.  How would this be useful?  It would take an
incredible program to analyse any given program text in any
programming language, highlight a few lines, and print in
the message buffer, "This portion of the program implements
the LZW compression algorithm".  Even RMS might have difficulty
writing a program to detect the use of any patented technique
in any program.  Fortunately, he is not motivated to do so.

If there is no way to spot patent violations mechanically or
in some way automatically, then we are back to the present
state of affairs -- only big companies will have the cash they
need to investigate possible patent violations on a large scale,
and only big companies will have the legal resources to force
compliance, or, as in many current cases, to blackmail possible
violators into paying protection money to avoid a court battle.

An unenforcible law is not worth having.  The present absurd
system wastes human resources and benefits no one.  And all
suggested alternatives seem to rely essentially on the knowledge
and good will of programmers.  Even if we can assume their good
will, it is impossible for programmers to acquire or apply an
effective knowledge of the huge number of software patents
already in existence.  The only possible solution is their
abolition.
			A. V. Le Blanc
			University of Manchester
			LeBlanc@uk.ac.mcc