Table of Contents
The Push to Get Free Unix Licenses
by Warren Toomey
By the 1990s, Unix was no longer available through Bell Labs; it was sold as a commercial product either by AT&T or the Unix System Laboratories. All fine and good, except that I, like other people, were interested in the early versions of Unix and there was no legitimate way that we could obtain licenses for them. An e-mail dated October 1995 shows that I'd asked Dennis Ritchie about getting licenses. His reply, in part, said:
AT&T doesn't think it has the right to issue new licenses for any Unix editions, and won't do so. Several years ago, the Unix IP was transferred to Unix Systems Laboratories, at first as a mostly-owned subsidiary; USL was then sold to Novell, and recently Novell agreed to sell its Unix business to SCO (and HP is partially involved). As of the end of October 1995 this last sale had not been consummated.
In other words, AT&T is out of the Unix business, and has been for some time; any licenses will have to come from the current owner. But the “current owner” is itself in flux, and I doubt that even before the latest sale, Novell was quite geared up to issue Seventh Edition licenses; I have no useful suggestions about how to obtain one officially.
So, although we were trying to create an “old Unix” users group and repository, it would only be accessible by those with existing licenses for these systems. For some, this was difficult to prove as the original V5, V6 or V7 licenses were now 15 to 20 years old and many sites had simply lost or misplaced their licenses. Others simply wanted a way to buy licenses so that they could run the old Unix systems on the PDP-11s and Vaxen that they owned.
Approaching SCO for Help
In June 1996 I made an approach to Dion Johnson, who was Product Manager at (old) SCO who believed that they had acquired all the rights to the Unix source code. Dion told me in a private e-mail:
SCO owns the licensing rights to all versions of the UNIX system, or so our legal folks tell me. Now, of course there are many derivative, licensed versions, and some of the holders of those licenses have rights to sublicense.
As for your friends who have rescued ancient PDP machines … I am confident that SCO would cheerfully encourage them to run UNIX on these antiques without any payment to us. I can't quite officially give that permission myself, but I can speculate that SCO certainly would not mind.
While this was “ecstatic news” for me, it still did not make the use of Unix legal, and it did not clarify if people without source licenses could download old Unix source code from our nascent Unix Archive. I asked Dion if SCO's lawyers could “draft a letter … which sets out exactly what you said above” and if I could “legally distribute the source code to the PDP versions of UNIX, and to anybody?” (e-mail to Dion, 27 June 1996). He replied “If you want it to be an official letter, it's a whole different ballgame. I will see what I can find out.” (e-mail dated 27th June 1996).
I pressed Dion for something that was legal: “It's the people who have never had a license that are the problem. … Even a cheap ($100?) single-user license would probably be good enough. … We need a solution that's little or no work for SCO, reasonably cheap for the people currently without licenses, and legally above board.”
For the next few months, I heard a few times from Dion: “I will see if the licensing guys will do anything...”, “I am once again pushing on the bureaucracy to see if we can get this permission for you.”, “I am pestering the licensing powers again.”.
In the interim (July 1996), I reminded Dion that SCO had already agreed to release the 6th Edition Unix code so that John Lions' commentary could be published:
I've just had an interesting email from Peter Salus who has just negotiated the publication of John Lions' commentary on 6th Edition UNIX (with source code). He says:The Lions book (which should be out in a few weeks; it really *is* in production) was negotiated by me and then “agreed” to by Dan Doernberg, the publisher.
Dennis Ritchie and I each wrote to Doug Michels and Mike Tilson (both VPs at SCO and old UNIX hands) and they told their lawyers that it was ``OK.'' Thereafter, it was only a matter of drafting and faxing back and forth.
Why don't you refer Mr. Johnson to Michels and/or Tilson, referencing the V6 permission?
So it looks like SCO has agreed to the release of PDP UNIX source code, in some form. Perhaps we could use this agreement to work on a way to make PDP UNIX source code available to owners of PDP-11s?
Binary Licenses Only
Dion heard back from legal and wrote back in October:
From a legal standpoint, for the kind of license you want, there is still too much risk of exposure of parts of the UNIX source code that, even after all these years, is still under trade secret protection. Bear in mind that if we relinquish trade secret protection for certain techniques and processes that exist in the old version, it would also impact the claims of protection in current product.
We might consider offering a binary license for old versions of the UNIX System. This would be accomplished by modifying the form of the Digital license for the 0S/8. We might be able to offer something like that if you think it would help.
But source code looks like a no-go.
What a blow! I wasn't going to let this go, so I asked Dion if, as they already sell SysV source licenses, why couldn't they change the system name on the license and sell V6 and V7 licenses (e-mail to Dion, October 1996). He wrote back: “I appreciate your persistence. … We are willing to sell you a source license, provided we make enough money to cover the cost … But the issue is that this license does not give you the right to provide that source code to anyone else, which is what you want to do … to supply the source code easily and directly from your site.” (e-mail dated October 1996).
Surveying the Interested Parties
I continued agitating for early Unix source licenses by setting up a PDP-11 Source License Questionnaire which ran from November 1996 through to mid-1997. I thought that this would help inform SCO on the level of interest in source licenses. I e-mailed the link to Mike Tilson who had helped with the Lion's commentary. He wrote back (e-mail dated November 1996):
I personally think this would be a nice thing to do. As I explained in earlier e-mail, I think there are a number of issues that would require considerably more thought than the effort that went into releasing the Lions book, as this deals with a core of UNIX that is still of value even today (I guess the original authors did a pretty good job!) …
It would be nice for this to happen in some way, and when and if the right people in the development organization become engaged, I'd be happy to provide advice to them.
Meanwhile, Mike let me and Dion know (e-mail dated December 1996) that:
You may be interested to learn that SCO has licensed PDP-11 UNIX binary distributions for personal non-commercial education purposes, for Fifth, Sixth, and Seventh Edition versions of UNIX for the PDP-11. While this is not source code, it certainly seems of interest to PUPS members. Hope this helps.
This had been part of the release of version 2.2d of Bob Subnik's SIMH simulator: with the binary license Bob was able to distribute “UNIX V5, V6, and V7 for the PDP-11” as demonstration software with SIMH. It was a step in the right direction, but only a small step.
Petitioning SCO for Source Licenses
In July 1997, I decided to become more active and get as many people as I could find to petition SCO for early Unix source code licenses by setting up an on-line petition. I told Dion (e-mail dated July 1997):
I thought I should let you know that I'm organising a formal petition to SCO about PDP-11 UNIX source code licenses, rather than perpetually hassling you about it. Your comments on it are very much appreciated. We really want to work with SCO on this & appreciate SCO's efforts to date with the Lions Book and the binary licenses to 5th, 6th and 7th Edition.
At a later date, when the petition's “fully signed”, we will make a formal presentation of the petition to SCO.
Dion wrote back (e-mail July 1997): “An excellent idea! Thanks for letting me know. I will certainly be your advocate on this.”
It didn't take long for the electronic signatures to fill up on the petition; by the end of July the petition had over 300 signatures on it. Even before then, I'd sent the petition in to some of the VPs in SCO for their consideration. Dion wrote back: “That's great about the signatures. Yes, I perused the earlier list and it's really amazing that we have such famous support for this. I am sure it will be a great PR victory when we finally get this arranged.” (e-mail dated 23rd July 1997).
So, we had gone from binary licenses to what sounded like an arrangement on source licenses!
Drafting the License Terms
SCO were now seriously “looking into how we can provide this source code”, but there were “some intellectual property issues that must be resolved” (e-mail dated 23rd July 1997). The legal people inside SCO now had to “[protect] SCO's interests in the code in the right ways” (ibid.) and there would be “further internal iterations here as we craft a license that works for all parties.” (ibid.)
The minutiae of the dialogue on the license terms (before the first draft) between myself and SCO can be found in these e-mails:
Dion sent me the first draft of the proposed license in September 1997. Steven Schultz also received a copy, and he sent in some comments (e-mail dated 17 September 1997); I also had some concerns about the first draft, especially the country limitations, single CPU limitation and what Unix versions were included (e-mail dated 18 September 1997).
Most of these were “fixed” according to Dion's reply (dated 19 September), but there were still some issues to clarify, e.g. “personal use”, a one-time of an annual license fee (ibid.), “the issue of PWB/UNIX, Mini UNIX and 32V”.
In mid-October 1997, Dion wrote that “I am still dickering with the legal folks in NJ. …[T]hey are moving as slowly as possible. I fired another salvo last Friday. I may have to escalate back to the VP level to get them off the dime. Sorry it's taken so long”. Then, at the end of November 1997: “I have a new cut of the proposed license from our NJ Legal folks. I compare with past versions and see what it has. I know the 32V was added to the list of licensed versions (good). Not sure if the other problems were fixed”.
Hammering Away at the Lawyers
Obviously, the task of convincing the legal people was harder than I thought. Dion commented that “they know about the petition. It's against their religion to go along with stuff like this. These are old AT&T license lawyers. I just gotta keep hammering away on them and finally get the exec VP to weigh in on my side when I have exhausted all the discussion options. Gimme a couple more days.” (e-mail dated December 1997).
In February 1998, Dion was able to pass on the second draft of the proposed license. He noted that:
- Multiple CPUs are handled
- Multiple countries are handled by the device of filling in the “AUTHORIZED COUNTRY” blank with a list of countries.
- Fee is $100 one-time only
- Other variants on the early UNIX, such as MINI-UNIX could be viewed as successors to, say, 3rd edition, so they are covered. Only system V is excepted.
I had some small issues including the serial number of the CPU (for simulated CPUs, what is their serial number?) and the fact that 7th Edition Unix was missing from the license! (e-mail dated 23 February 1998). The next draft (dated 24 February 1998) fixed those issues.
The Last Few Details
After that, it was mostly working through the administrivia: “Who nominates the authorized countries and designated CPUs?” (24 February 1998), how SCO will inform me of license holders so I could give access to the Unix Archive (25 February 1998) and payment options (2 March 1998).
In early March, I send Dion what I proposed as a set of instructions for people to follow to buy a Unix source license from SCO and then get access to the Unix Archive. He replied (ibid.) with a few amendments.
SCO Provides Low-cost Source License for Ancient UNIX Versions
SCO, in cooperation with the PDP-11 UNIX Preservation Society, is happy to announce the availability of a special software license for UNIX Editions 1-7 and 32V.
This license permits hobbyists and enthusiasts to have access to the source code of these historic versions of UNIX, for personal and noncommercial use, and to share experiences and code updates with other licensees.
The license text may be found online at www.sco.com/offers, and the PUPS website. A $100 fee is charged to offset cost of legal and administrative handling of the license.
Dion felt that, even though this hadn't been officially blessed by legal, it was time to make it happen: “I suspect the legal/finance folks will holler foul when this hits them but hell that's their job. ” (ibid.). By the 20th of March, SCO already had “about a dozen licenses here, all paid up and signed off.” (e-mail from Dion).
So, finally, source licenses for early Unix were available, although they were not free. This would never have been possible without such a steadfast and stalwart advocate as Dion Johnson. At the end of March, Dion wrote: “The legal people are a little grumpy at me, but they also know my heart is pure and they sorta respect that. … I am very proud of doing this and grateful for your help and the help of the other PUPS and other participants in the Unix heritage. … I am known as a bit of a maverick, and customer advocate, so it's expected. Every company needs a few people like me”. I could not agree more!
Processing the New Unix Source Licenses
After that, it was a bit anticlimactic. People would purchase a Unix source license through SCO's website, Dion would e-mail me their name, I would provide access to the Unix Archive, and periodically Dion would snail-mail me photocopies of their signed licenses.
Then: Free Unix Source Licenses
On the 13th of May, 2000, SCO changed the price of their early Unix source licenses to $0, i.e. free. I don't seem to have an e-mail trail that describes the motivation behind this; however, by this time Dion had passed the early Unix reins over to David Eyes at SCO (e-mail dated November 1998).
SCO now provided a click-through page on their website where one could obtain the source license, and SCO still notified me of the new license holders. However, the click-through page (once clicked), gave the new license holder a link to the Unix Archive.
By March 2001, over 1,800 people had either bough the US$100 Unix source license or had obtained the free Unix source license (e-mail to David Eyes). At this stage, SCO were in the process of selling the Unix business over to Caldera. I had asked David: “What's happening with Caldera and the transfer of this IP, and is there anybody there that I can contact yet?” (ibid.). David wrote back: “The transaction still hasn't closed. I believe that the public word is that it will likely close in April. There was a last-minute renegotiation regarding the rights to OpenServer, which Caldera will now own completely as well.” (ibid.)
Finally: BSD-Style Unix Source Licenses
Once the Unix business had been sold to Caldera, there were more internal changes, and in January 2002 I received this e-mail from Dion:
I'm happy to let you know that Caldera International has placed the ancient UNIX releases (V1-7 and 32V) under a “BSD-style” license. I've attached a PDF of the license letter hereto. Feel free to propogate it as you see fit.
Many thanks to Warren Toomey, of PUPS, and to Caldera's Bill Broderick, director of licensing services here. Both of these gentlemen were instrumental in making this happen. And thanks to our CEO, Ransom Love, whose vision for Caldera International prescribes cooperation and mutual respect for the open source communities.
Wow! We had gone from no licenses to binary-only licenses, to paid then free licenses, and now BSD-style licenses. I could finally make the Unix Archive available by anonymous FTP and through the web.
(incomplete, more to come!)