🚀 go-pugleaf

In <4611@cluster.cs.su.oz.au> rpw3@rigden.wpd.sgi.com (Rob Warnock) writes:

>mab@wdl39.wdl.loral.com (Mark A Biggar) writes:
>+---------------
>| >I have a specific case to ask about.  This week the LPF has been posting
>| >messages about patent 4,783,761.  This covers a particular spelling check
>| >technique for a word processor.
>|
>| How are command completion systems like that built into emacs or tcsh
>| excluding from infringing on this patent claim?
>+---------------
>
>Prior art. BBN TENEX (later incorporated into DEC's TOPS-20) was doing
>command completion essentially identical to emacs or tcsh way back in the
>early 1970's. I'm not saying TENEX was first, understand, just that it
>predates patent 4,783,761 by more than a decade.


 Alas, "prior art" is a technical term under patent law.  Just
 as a "RISC" processor need not necessarily be as "reduced" (as
 the name might imply), "prior art" does not necessarily refer
 to "earlier equivalent work."   Although prior work *can* be
 "prior art," various other conditions must be met:


 First, the date:  The issue date of a patent is not the same
 as the legally provable date of invention; usually the date
 of invention is years earlier.  We would have to know the
 current inventor's date of invention, and this is not
 normally available short of a legal contest on precisely
 this issue.

 Even a legally provable earlier date for equivalent work
 would not show that the patent was invalid, but rather that
 the earlier inventor was entitled to the patent.


 Next, we normally think of "prior art" in terms of publications:
 books or magazine articles which clearly describe the invention.
 In general, "prior art," must have been "freely and openly
 available."  Even an actual working device may not constitute
 "prior art."

 We distinguish between a software product *containing* the
 invention, and the actual software invention itself.  In the
 case of software inventions, it is certainly possible for a
 product using an invention to be available, and for the average
 programmer ("practitioner") to still *not* know how the invention
 operates (at the level of detail protected by the patent).  If so,
 we may be dealing with a trade secret, which neither gets nor
 deserves patent protection, and which cannot interfere with
 a patent or an application.

 Even if the source code for such a system was "freely and openly
 available," if no documentation particularly described the
 invention, the average programmer *still* might not understand
 it.  In this case, even *published source code* might not be
 enough (by itself) to constitute "prior art."  In contrast, any
 descriptive publication, even in a minor journal, *would* be
 "prior art."  But no such publication was described.

 ---
 Terry Ritter    Ritter Software Engineering    (512) 892-0494