On Jan 1, 6:33 pm, tha...@xxxxxxxxxxxxxxxxxxxxxxxxx wrote:
> amicus_curious <A...@xxxxxxx> wrote:
>
> > What product or even product feature was actually innovated by OSS software
> > even in servers? All of Linux was a copy of functionality and format from
> > Unix.
Keep in mind that about 80% of what we now call "UNIX" was also OSS.
There was a huge body of libraries and applications published under
the BSD license. There were applications and "21st century office"
applications that were published as part of Project Athena under the
MIT license, this includes most of X11, most of the Andrew User
Interface Toolkit (the pattern for most modern GUI applications), the
EZ editor (most of the features of Microsoft Office, published while
Microsoft was still pushing Multiplan instead of Excel. There were
hundreds of other applications that were also created for UNIX under
open source licenses. Even the GNU applications were originally
written for UNIX but published under OSS licenses.
There are also many Public Domain applications which were republished
in both proprietary and OSS formats. UNIX administrators often had to
support multiple forks, multiple processor types, and multiple server
types. They often insisted on getting software in source code
format. UNIX administrators were also much more concerned about
security and were acutely aware of viruses being passed between MS-
DOS, most frequently through the boot sector and "hidden" tracks that
were often undetected while they spread viruses and worms to thousands
of computers.
UNIX administrators became even more insistent on source code and
adherence to public standards when the Morris Worm infected several
thousand UNIX systems in 1987. Many applications were tested against
"Reference" implementations that were published in source code
format. Even the proprietary implementations had to be compliant with
the OSS applications.
The term Open Source Software wasn't actually formalized until the mid
1990s, but the definition of Open Source Software included about 3
dozen licenses which were used to publish hundreds of applications
dating back almost as far back as 1977, the beginning of the modern
copyright law that covered software and software licensing.
AT&T licensed many of these OSS programs, and these licenses included
the right to republish derivative works using AT&T licenses and
copyright notices. One of the funniest instances was the "true"
application. The true application was an empty file that would return
successfully. The AT&T was the same "empty file" but included a 2
kilobyte copyright notice for an empty file.
Many of these same OSS applications were also also licensed under
other licenses, including the GNU public licenses, where derivative
products were implemented, enhanced, and expanded, including numerous
bug fixes and security fixes, under these GPL and related licenses.
Ironically, the proprietary implementations were not able to legally
implement these enhancements, because the enhancements were ONLY
released under the terms of the GPL licenses.
Even older Open Source applications dated back to NASA and the "Space
Race". Remember that NASA was not allowed to patent most of it's
technologies, and most of the technology was public domain because it
was funded by public tax funds and federal grants. Some of this
technology goes all the way back to the late 1950s, perhaps as far
back as 1957.
In addition, there was software that was declassified during the Ford
and Carter years, much of which dated back to World War II, and even
software dating back to 1947. This means that Open Source software,
including Public Domain Software republished under Open Source
licenses, may be as much as 60 years old.
> As has been explained to you before, most of the Internet stack
> was developed as OSS. Why should I list it all again if you insist
> on keeping your fingers in your ears? Sendmail, NAT, P2P (FreeNet),
> web servers/browsers, perl, fetchmail.... any of this ringing any
> bells?
Keep in mind that the DOD ARPA projects were largely funded by federal
research grants. Some of the technology dates back to the height of
the cold war, including the initial requirements to design a
communications system capable of surviving a nuclear holocaust.
Ironically, many of these design principles were incorporated into the
design of the TCP/IP stack, and the use of layered protocols, which
could be encrypted and scrambled at any layer.
It's ironic that we are seeing attempts to patent software that had
been developed by DOD/ARPA as far back as 1947, for exactly the same
purposes claimed in the current patent applications.
Keep in mind that many of the early computer pioneers, including
Sperry, Harris, Univac, and IBM, were also early UNIX adopters. Many
of these companies ported UNIX versions of software that dates all the
way back to the sorting algorythms used for sorting and counting
Hollerith cards during the Census of 1890. It's amazing that there
are actually people attempting to "patent" these algorythms even
though both the algorythm and the "claims" (intended use of the
algorythms) had been documented as much as 100 years ago.
> > Well, you admit that is what they are doing, even with the lame notion of
> > somehow making it sound normal by handwaving at some concept that others do
> > it too. This would be a good place to make mention of whatever you think
> > OSS developers have totally created on their own.
You seem to forget that OSS developers included thousands of
professionals who worked for insurance companies, banks, legal firms,
even manufacturers, who were NOT in the software business. They wrote
software as tools, to help their employers do various jobs more
efficiently.
A small portion of the code was strategic and proprietary, but often
the majority of the code, often as much as 95%, was more "generic" in
nature. Often, the algorithms were published in scientific journals,
IEEE journals, and later ACM journals. There were also a number of
special interest groups that supported generic functionality in
various programming languages, and introduced new paradigms (including
object oriented programming, multithreading, and most of what we
associate with Modern Computing.
These people weren't exactly "open source publishers", but they
published these snippets of code in a public domain way. When
copyright law changed in 1977, many companies created proprietary
licensed versions of the public domain software, but there were also
millions of lines of code which were also published under Open Source
Licenses such as the BSD, MIT, and GPL licenses. In some cases, the
same "root code" has been published under several different licenses
ranging from highly proprietary licenses, to the GNU Public License
with it's "mandatory source code disclosure" requirements.
Ironically, the Open Source Software was often better supported, and
provided proof of first publication. It's very hard to claim that you
"invented" something when you have kept your code under the protection
of proprietary "binary only" licenses and strict nondisclosure
agreements. It's even harder to prove that the publisher of the Open
Source version "stole" your idea, when the OSS version of the
information was widely available, especially when the OSS version is
published by a college student, or as a "simple little hack" or
"trivial patch" by the person disclosing the idea and it's source code
implementation.
> So nobody should ever attempt a competing version of software if
> something already exists? You have something against free markets?
Nothing wrong with coming up with a competing version, as long as
credit is given where credit is due, and licenses are complied with.
Ironically, it is often the monopolistic vendor of proprietary
software that is often the worst pirate of all.
> Was it wrong for MS to come up with their own browser just because
> one already existed?
The problem was that MS DIDN'T come up with their own browser. They
purchased "branding rights" to the Mosaic browser, which was generated
by OSS contributors under an NCSA license that was very similar to the
GNU public license.
Ironically, Marc Andreeson, the founder of Netscape, was also the lead
developer of Mosaic. Mosaic was published under a license that
prevented Andreeson from marketing a proprietary version, so he
completely reengineered the browser from scratch.
The problem was that at the very last minute, Microsoft's lawyers
added a few additional provisions, permitting them to market their own
proprietary derivative products, and told the vendor (Spyglass) to
either sign the contract, or the entire deal was off. There are a
number of different versions of the story, depending on who is telling
it, and when it is being told.
What is certain is that right after Spyglass made the deal, the NCSA
unilaterally rewrote the license under which Mosaic was released. The
contributors were not asked to accept the new license terms, or even
officially republish their code contributions under the new license.
Some say that it was this illegal seizure of intellectual property,
for sale to Microsoft, that ultimately lead to the court cases against
Microsoft.
Needless to say, after this change in policy, much of the support for
the NCSA vanished, and contributors rallied around new OSS licenses,
including the Artistic License (Apache), the GNU public license, the
Common Public License, and several other licenses which protected the
rights of the contributor, and allowed contributors to control the
publication of their applications or enhancements, especially if a
commercial publisher, like IBM, Sun, BEA, or Borland, wanted to
publish a proprietary commercial implementation.
> And I have listed the OSS innovations, you just continue to ignore
> them.
Seriously, there are entire web sites dedicated to the history of Open
Source software. Of course, Microsoft has it's own version of
history, which seems to change from year to year. I'm always amazed
at how Microsoft's own version of history has changed in various print
Microsoft publications over the last 20 years.
> > More handwaving rather than any specific examples. I suppose you might want
> > to mention the old hat GNU stuff that Stallman supposedly conjured up in the
> > early years, but weren't those "tools" themselves copies of what pay-for
> > software was doing in those days.
Actually, it was the other way around. Much of the pay-for software
was just publications of BSD software that was published under the
AT&T license. AT&T couldn't sell System III without the BSD
extensions, so they licensed the BSD extensions and included it in
their System V releases. The System V release 4 distribution included
ALL of the BSD software.
Many of those who wrote BSD software as students were not entirely
happy with AT&T getting their software, without paying them a royalty,
or even giving them a job, then charging as much as $700 per user for
licenses. In retailiation, many of them published these college class
projects under the GNU license. In some cases, the GNU versions were
"adopted" and then "enhanced" and "improved". Ironically, because the
GNU versions were released under the GNU public license, the
proprietary versions couldn't use these proprietary enhancements.
Some companies, like Sun and DEC and IBM, offered the GPL and other
Open Source software as "Complimentary" software (like getting a
complimentary steak with your salad), or "Unsupported Software" (often
better supported that the proprietary commercial versions whose
support was limited by tight staffing budgets.
Meanwhile, the OSS community was staffed by UNIX administrators,
college students, and others, many of whom contributed while working
from home. In many cases, they told their employers about their
accomplishments, and asked for permission to publish it under open
source licenses. In most cases, their employers were more concerned
about avoiding liability than about getting royalties. After all,
their business was insurance, banking, health care, or making cars,
not selling software.
> > It is hard to make much sense out of his
> > rantings, but maybe it takes one to know one.
Here are some good references:
http://www.opensource.org/
http://www.opensource.org/licenses/alphabetical
http://en.wikipedia.org/wiki/BSD_license
http://en.wikipedia.org/wiki/Hacker_ethic
http://sourceforge.net/softwaremap/trove_list.php?form_cat=187
http://www.darwinsys.com/history/hist.html
http://www.oreillynet.com/pub/a/network/2000/03/17/bsd.html
Note that even AT&T published an "Open Source" version of UNIX, in
1976, version 6 was published in source code form, to many colleges
and universities. Remember, at the time AT&T had a monopoly on the
telephone system and had been barred from producing their own
computers. AT&T accepted divestiture in 1983, partly so that they
could market UNIX and AT&T computers, including several PBX switches
as well as computers like the 3B2 and 3B20 and the #5ESS electronic
switch.
A professor at MIT even produced a fully annotated version of Version
6 UNIX that provided extensive documentation, it was two "pages" per
page, two on front, and two on the back, and filled a 2 inch notebook,
but for students who really wanted to understand the finer points of
UNIX, and the finer points of C programming, this was the best
tutorial of all.
The UNIX source code even featured object oriented programming,
inheritance, abstraction, and methods, back in 1976, when smalltalk
was still in it's infancy and Microsoft was still trying to get BASIC
to load from punched paper tape.
Since AT&T couldn't patent software at the time, and the 1976 version
predated modern copyright licenses, the code was essentially public
domain. The code was for "educational purposes" but there were not
many restrictions on creating proprietary versions such as Xenix,
Venix, OS/9, and of course BSD 2.0 through BSD 2.6.
Depending on who you talk to, many claim that it was the OSS community
who introduced demand paged virtual memory to UNIX, along with
networking, ethernet support, terminal independent support (termcap),
X11 graphics, wysiwyg editors, in addition to numerous compilers,
shells, scripting languages, and other common tools.
http://en.wikipedia.org/wiki/Unix
http://www.ee.ic.ac.uk/docs/software/unix/begin/appendix/history.html
When AT&T settled with BSD, about the only part of UNIX that was not
available under the 1992 BSD license was the regular expression
library. Much of the other AT&T specific code was easier to implement
using the BSD libraries and tools.
> Please refer back to the specific examples I mentioned earlier.
> You must remember... it was the stuff you rejected as not being
> 'products' (whatever that is supposed to mean).
Keep in mind that in "Microsoft World" the only "products" are huge
monolithic applications that have to be installed using Microsoft
Software Installer or Install-Shield, require numerous registry
modifications, and customizations are limited to whatever is included
in the GUI configuration menus. This monolithic application has to
include everything and do everything in a single executable, including
documentation, setup, configuration, error handling, and of course any
form of communication - usually in a totally mysterious, undocumented,
and proprietary format, or at least an undocumented "enhanced" version
of a public standard format.
Of course, in UNIX world, "applications" are highly specialized, and
they can be combined as building blocks like LEGO or Tinker-Toys,
using shells, python, PERL, Ruby, or other graphical scripting
languages. Furthermore, help, documentation, configuration, and
support are often accomplished using public format files. If you
really want, you can have a nice GUI interface configuration tool that
will generate or edit these file for you.
Of course, if you don't want to put the lego blocks together, there
are lots of other people who will put them together for you. Heck,
you might even think it's a "real application".
> Thad
> Yeah, I drank the Open Source cool-aid... Unlike the other brand, it had
> all the ingredients on the label.
I've been guzzling from the Open Source fountain since 1977.
|
|
