Re: Software, Copyrights, Patent & Open Source

On Sat, 21 Aug 1999, Mike Oliver <mikeoliver[_at_]home.com> wrote:
>
> ------------------------
> A bit of history . . .
> ------------------------
>
> The author of QDOS (quick and dirty operating system) -- Tim Patterson
> -- used the ideas in Digital Research CP/M -- itself a watered down
> version and derivative of some of the Unix command and I/O structure.

CP/M was not based on Unix, it was based more on the operating systems used by DEC on machines like the PDP8. DEC stands, as I recall, for Digital Electronics Corp., and CP/M was marketed by Digital Research, which I believed was named that because Gary Kidal (?) -- I'm writing all this from memory -- had worked for DEC and had actually developed CP/M at DEC, which mas not interested in marketing it.

> Why? Mostly because Unix was so prevalent at the time that to have
> third party development and support, it was easier to design something
> that functioned similarly to known main frame structure (there were
> probably a lot of technical reasons as well)

Unix was not prevalent at that time. In fact I am not sure that Unix even existed at the time that CP/M was released. Certainly it was not widely marketed then, being available only within AT&T and at Universities.

The DEC software that influenced the development of CP/M ran on minicomputers not on mainframes. Unix was designed to run on minis and mainframes and any other platform, but the traditional mainframe software was not suitable for use on micro, or even mini, computers and thus had no influence on the design of CP/M or the programs that copied CP/M.

At the time that CP/M came into existence copying some of the DEC features there were no cases holding that computer software was copyrightable, and I belive that remained true at the time that QDOS was written.

QDOS basically was nothing more than a subset of CP/M translated from the 8080/8085/Z80 instruction set to the 8086/8088 instruction set. Apparently Digital Research did not consider this to be a copyright violation, or at least did not consider suing. However, if they had and had gotten a court that adopted the same rationale that was adopted in Apple v. Franklin, I think it is quite clear that Digital Research could have established that it was an infringement and obtained an injunction.

QDOS was written, however, in the tradition that did not apply copyright restrictions to the copying and adaptation of computer programs.

> QDOS became MS-DOS, which was then copied on a functional level by
> Digital Research/Novell/Caldera in DR-DOS:

QDOS was then purchased by Microsoft and released with a few modifications by Microsoft as MSDOS/PCDOS. It was still a copy, on more than just a functional level, of CP/M. But the time that MSDOS was released CP/M had been adapted by Digital Research to work on 8086/8088 boxes, and had been greatly improved. (There was even a multi-user version of CP/M, called MP/M that was years ahead of anything that Microsoft had released.)

With version 2.0 of MSDOS Microsoft included some I/0 features copied from Unix and added support for hard disks. (Until then, my old North Star Horizion with its 10 megabyte ThinkerToy hard disk that had a Z80 chip and ran CP/M80, could run rings around an IBM PC with MSDOS.

Because of Microsoft's marketing success, CP/M was renamed DR DOS and it was designed to be able to run programs written for MSDSOS, which included copying the Unix features that Microsoft had included in MSDOS. There was, of course, no claim by Microsoft that this was a copyright violation, because Microsoft itself was the infringer, having purchased and then distributed a blatant copy of the CP/M code and then copying some of the functionality of Unix.

DR DOS was always several steps ahead of MSDOS in functionality. Digital Research (and Novell and Caldera) did not copy the functionality of MSDOS, rather MSDOS repeatedly copied the functionality of MSDOS.  
> (from paragraph 38 of Caldera's Amended Complaint Against Microsoft):
>
> "DRI designed DR DOS to be the functional equivalent of MS-DOS, i.e.,
> to support the same API set. DRI was readily able to accomplish this,
> largely because of its experience in the development of CP/M, from
> which QDOS and MS-DOS derived." [see
> http://www.calderathin.com/lawsuit/amendment.html]
>
> Copyright law did not prevent DRI's development of a competing O/S.
> Nor did it prevent inter operability, hierarchical command structure
> copying, or interface similarities.

Of course not. But if Apple v. Franklin is good law, a copyright action could have prevented Microsoft from selling MSDOS. (Which is, I suppose, a strong argument for the position that copyright of software is a good thing.)

> While these developments were taking place in the PC industry, Unix
> was splintering into a number of derivatives in the mainframe -
> network environment. See
> http://www.ehlis.com/adam/solaris/history.html (contains a fascinating
> chart showing the development). In many cases the splinters were
> copies or derivatives under the open license ATT was forced to issue
> under orders that prevented it from selling (licensing for a fee)
> software.
>
> ------------------------
> Open sources
> ------------------------
>
> The Berkeley Software Distribution (BSD version) of Unix eventually
> sprouted the OpenBSD version which is now open source. However,
> Linux developed separately and independently, again however, expressly
> copying the structure, command and functionality of much of what is
> Unix. Linux was not hindered in the least by copyright law. Linus
> Torvalds and his friends copied ideas and non-protectable expression.
> He essentially reverse engineered UNIX for the Intel386 (or now x86)
> processor - a practice long regarded as not violating the copyright
> of the original software owner (as long as the copy was legally
> obtained). And, while the DMCA and some whacky cases have somewhat
> restricted this reverse engineering right (typically, however, in
> cases where there is a larger contractual relationship), copyright
> law has NOT prevented (indeed, not even hindered):

Linus and the other authors of Linux wrote only the kernel of the Linux system, which was compatible with other Unix based systems. Most of the Linux ``operating system'' consists of programs written by the Free Software Foundation as part of the GNU project, and of other free/ open software like the X system and the Apache Web Server and sendmail, etc. Most of the GNU (Gnu's Not Unix) software have functionality similar to that supplied with proprietary systems, but in most cases the GNU software is better.  

> - an almost literal copy of Adobe's vaunted Photoshop product (ably
> named the GIMP)

I don't use GIMP much, and have never used Photoshop, but I don't think there is much similarity between the two programs. GIMP is largely composed of plugins that were not written by the GIMP's authors, and it is based on libraries that have no corresponding parts in Photoshop, but that are the basis for the entire GNOME desktop management system.

> - open source interfaces that emulate Win95 GUI, and the MAC
> - open source copies of other functional programs like sendmail,
> ftp, ssh, rsync, sftp etc.

Sendmail and ftp, and I believe the rest of these programs, were open source from the beginning.  

> Copyright law has likewise not prevented or impaired development of
> 'clones' of many popular games (pac man, space invaders), though I
> admit that newer computer games that feature 'characters' and plots
> are more likely protectable, at least in those elements.
>
> ------------------------
> Copyright law is a benefit
> ------------------------
>
> More importantly, however, copyright law has given open source
> developers *control* over rights to use their source code.
> Specifically, copyright law has permitted them to _claim ownership
> in the code_ and then to grant a _license_ to permit others to use
> the code but only certain conditions (predominantly, that all
> modifications be made publicly available in source, and no further
> restriction on a second user from making changes to the sources).
> See <http://www.gnu.org/copyleft/gpl.html>.

But without copyright law the economics of software publishing today would be so different that there is no way to say to what extent people could sell software that could be freely copied but without source code. We do know that people can make money distributing code under the GNU licenses. And if software can be copied without needing to pay for it, I don't see why anyone would pay for software that did not come with source code.

But in any case we are talking about a different universe. In the actual world, this history that you recite, when it is properly stated, describes a world in which MSDOS exists only because copyright law was not enforced.  

> In short, I see no evidence that copyright has hindered in any
> respect, the open source system of program development. Indeed,
> *because* of copyright, the movers behind the open source system have
> prevented other companies like Microsoft from acquiring sources as a
> free rider, closing development and making it proprietary, and then
> using their marketing might to quash competition.
>
> Finally, I continue to be surprised by the view of some people that if
> someone writes a software program, they should be *required* to donate
> it to the public. Copyright recognizes that the right to use the
> fruit of the author's creative output is in his or her control (at
> least where compulsory secondary licensing rules do not apply). Some
> authors choose a closed proprietary system. Some authors choose an
> open free-development system. Copyright law gives them this choice,
> it does not compel either of them (though the default rule is
> prevention of copying).

Copyright law by the terms of the Copyright Act has to do only with copying, not with use.

> I am just looking for people who criticize copyright law to be more
> precise about exactly what they are criticizing. I have always
> thought of copyright as rather limited in its scope, because to show
> infringement you must prove literal copying or access and substantial
> similarity, and because copyright is limited by idea-expression merger
> and process-expression-merger.

Fair enough.

> ------------------------
> Patent law, the bear
> ------------------------
>
> Contrast that with patent law. Patent law relating to utility
> patents expressly only protects novel, utilitarian, and non-obvious
> inventions. A patent can be infringed regardless of access, or
> copying. And at least in the U.S., there is no requirement to cite if
> not known (or even really investigate) non-patent prior art. Thus,
> having read at least 50 or so 'software' patents, and having been
> involved in a large software patent infringement case, I see the
> software patent as the main threat to open source development. Many
> of these patents seem outrageous on their face, because they build in
> an obvious or non-novel way, on the programming tricks used by hard
> core programmers since the 50's, 60's and 70's. Some are indeed
> novel, and deserve protection. But you are in a whole different ball
> game getting a C&D on a patent, than when you get one on a copyright
> claim. The sheer discovery burden of proving one of these patents
> invalid costs hundreds of thousands, if not millions of dollars. If
> the FTC had not sued Microsoft for antitrust violations, I'm betting
> that by now, Microsoft would have already sued some of the large Linux
> distro providers on some patent they hold in their arsenal. If not to
> achieve victory, at least to cause great financial pain. I thank my
> lucky stars every day that the FTC finally did something (even if I
> think this browser war integration thing is not the real case against
> Microsoft).

We are in agreement here. But, unfortunately, patent law does apply to use. Isn't the big question: how do we restrain the Federal Circuit?

> <references>
> The history of the computer is fascinating. I offer a few links here:
>
> http://www.icom.org/vlmp/computing.html
> (virtual museum of computing)
> http://www.physplt.wsu.edu/pub/people/BremnerFiles/MicroHistoryTOC.html
>
> (radio interview series)
> http://www.users.csbsju.edu/~jgramke/Help/unix/unix/data/history.html
> (unix history)
> </references>

Thanks. That is most helpful.

--
Peter D. Junger--Case Western Reserve University Law School--Cleveland, OH
 EMAIL: junger@samsara.law.cwru.edu    URL:  http://samsara.law.cwru.edu/
        NOTE: junger[_at_]pdj2-ra.f-remote.cwru.edu no longer exists