🚀 go-pugleaf

In article <1991Jul1.234602.13089@usenet.ins.cwru.edu> edguer@charlie.CES.CWRU.Edu (Aydin Edguer) writes:
>In article <1991Jul1.190608.24178@beaver.cs.washington.edu> pauld@stowe.cs.washington.edu (Paul Barton-Davis) writes:
>>And others say this:
>>
>>We would be much happier if the licence (and the LGPL) made some
>>explicit reference to people who develop specialized software systems
>>which are sufficiently specialized that only a handful of copies will
>>ever exist, and stopped assuming that everyone who writes software
>>works either for a computer vendor or general purpose software
>>company.
>
>Huh?!  I am really puzzled by this.
>
>If your market niche is _so_ specialized that you will only sell one copy
>then you will not lose any money by giving the source code along with
>the binaries (GPL or LGPL).
>
>Even when selling just a "handful" [five or less] you will not be losing
>revenue for yourself by releasing your object modules (LGPL).  No one will
>try to decompile your object modules unless there is a *big* gain for them.
>And there will not be such a gain for a program for which only a "handful"
>of copies exists.

This wasn't quite the point I was making. Its not clear to me that the
GPL would even allow me to use source code in this way (although it
may - I just am not sure). Note I said source code - not stuff from
the libraries, but perhaps something like taking gcc's parser, ripping
it out and using it for something else. If I charge time+materials to
a client for this, am I ripping of the original author of the parser,
or simply making good use of their work ? 

>If you are talking about selling software that includes large chunks of FSF
>code then, whether or not you are selling one or a thousand copies, you _OWE_
>the FSF and all the programmers who contribute to the FSF effort large
>chunks of money.  Instead of money, they only ask for the rights to use your
>code as your used theirs.  What could be more fair?

Suppose that the mythical company referred to as "we" (you'll note my
current address is in the .edu domain) did not charge any of its
clients for the code that came from FSF material. Does the 
GPL imply that use of copylefted material in a non-copylefted program
makes one financially indebted to its authors ? I thought that the
whole point of the GPL was to remove this kind of relationship between
software and its creators.

One of the problems with the whole GNU idea (as wonderful as it is) is
that it seems targetted at general purpose software (operating
systems, compilers, various tools). Neither it, nor RMS's general
ideas about software seem to address the issue of software that is
being written for very specific purposes. Case in point: writing
software to scan, halftone, compress, store and print images. Although
some large companies are attempting to offer "off-the-shelf" solutions
for this type of thing (Kodak & Agfa spring to mind), for many cases,
these generic systems do not have the performance needed to specific
applications. The software that is created for this type of task is a
very different animal from the stuff that the FSF has released to date
(please don't mention GhostScript - having an interpreter for a page
description language is about 5% of the problem).

What are companies that choose to make a living producing stuff like
this supposed to do with regards to software sharing ? I think that
the responses to Jim Showalter's questions provided some good insight
into how GNU is working: better to see it as a way for companies and
research organizations to collaborate in a loosely structured,
cooperative manner. GNU is not free, but is something whose cost has
been distributed across the many of us who choose to use, support and
develop parts of it. This is something ideally suited to general
purpose software. I'm not sure the same model can be extended to
special purpose systems created for 1 or 2 clients.

>I don't see how changing the GPL or LGPL will help you or the FSF.
>But, I am curious, how would you have the FSF word this *new* part of the GPL?
>
>	"Anyone may sell up to 5 copies of software.  After selling 5 or
>	 more copies of software, they must release the code to anyone,
>	 including former customers for up to 2 years from the date of
>	 sale of the fifth copy.  They must also make best effort to
>	 contact all customers to notify them of the availability of the
>	 source code."
>

I would guess something like:

	If this software is included in any form, modified
	or not, no charge will be made to anyone to whom
	you distribute it in relation to the provision of
	the functionality it provides. If you choose, you may charge
	for the modifications you perform, and for added
	functionality that you make available.

This directly violates the GPL. The only way I could see of
reconciling the two of them is to try and come with a way to specify
that you cannot charge anything for any modifications that leave the
overall functionality intact (i.e you can't charge for adding a new
keyword to gcc, but you can charge for taking its parser out and
modifying it to be a part of a C-based interpreter for image
manipulation). I don't believe this is possible to do.

So, seems to me we are stuck with a real division between those who
would like to use free software in a manner similar, but not identical
to that espoused by the GPL, and those who think that no-one should
ever be able to sell a software system that include GPL-covered code,
even if there is no charge for such code. Licensing from the FSF is a
possibility, but that heads back into the messy divide between
"general purpose" and "specialized" software.

Gone on too long. I'm not sure how much I believe all the above ....
-- 
"When in some obscure country town, the farmers come together to a special town
meeting, to express their opinion on some subject that is vexing the land, that
I think is the true Congress, and the most respectable one that is ever
assembled in the United States." - Thoreau