From: Paul A Vixie (vixie@pa.dec.com) Subject: Open Letter to Lynne Jolitz Newsgroups: comp.unix.bsd View: Complete Thread (2 articles) | Original Format Date: 1992-03-24 17:30:51 PST Lynne, You sent your "response to Kolstad" to John Sokol, who posted it here under the subject "Re: netatalk-1.2 -> netatalk-1.2.1". I am replying, since after having heard from all of the principles (you and Bill, then Rob, Keith, and Mike), I think I finally mostly understand what's going on here and I have a few suggestions, a few questions, and of course, a plea for peace. I am making this an "open" letter because the plea for peace is really from a large number of "us" and is *for* a large number of "us". I'll start with a little bit of my own background. I was one of Symmetric Computer Systems' customers. In the process of buying my s/375 I visited your San Jose offices about eight times and spent quite a bit of time visiting with you, Bill, Bill's father, and your manufacturing technician. I have met your daughter (though she was much too young to remember me), and I remember thinking that Symmetric was a good company trying to do good things. I was happy to be a part of the project, even though my own s/375 never did work all that well and I ended up taking it up to the Los Altos Rod and Gun Club and blowing it to smithereens on the 25-yard tin-can range. (I have pictures.) When I ran into Bill in Colorado back in 1990, he remembered me and was proud to show off his 386 laptop running his BSD port. We talked about pregnancy and babies since you two had been through it and I was expecting my first son. My conversations with Bill were cordial and I like to think that they will be cordial next time we meet. Certainly when Bill was working for BSDI and I was setting up his 9600BPS Internet link through the Alternet POP here at DECWRL, my various conversations with Bill were all quite friendly. I am currently working to make 386BSD available via anonymous FTP from Gatekeeper.dec.com, and I have offerred to make mag tape copies for anyone who mails me a tape with return mailer and postage. 386BSD is an exciting piece of work and I want to help it succeed in any way I can. My bitstring macros are part of BSD, and my version of Cron will be there too, someday. I believe, from my experience as detailed above, that I am in a position to understand the context of the "386BSD vs. BSD/386" feud. I think there are a lot of bad feelings all around, that communication efficiency is at an all- time low, and that most of what you and Rob and Keith and Mike have been arguing about are misunderstandings rather than actual differences of opinion. I know the CSRG folks reasonably well and I know that they, like you and Bill, are very interested in a royalty-free BSD. We all believe that BSD is the One True OS, that it is better than SysV.n:(0>n>~), better than Mach, better than OSF/1, better than SunOS or Ultrix or UMIPS or anything else that either contains proprietary code or has no champions and no soul. Mike is working for BSDI now, but he said at the Usenix BOF that he plans to finish up some of his CSRG projects on BSDI's time, and contribute the work back to CSRG for inclusion in the upcoming hopefully-non-licensed 4.4BSD. Is this the mark of a man who wants to hoard his work or keep things proprietary? Rob Kolstad hired him even knowing that some of his initial efforts had to be released back to CSRG without any kind of direct monetary reward to BSDI. Is this the mark of a man who wants to hoard other people's work or keep things proprietary? Keith and the other CSRG folks are pounding keys at this very instant to replace more and more of BSD with royalty-free code. They are funded but they are not rolling in money and will probably never roll in money from selling BSD tapes at $850 a whack. They do it because they think it's important. Just as you and Bill do what you do because you think it's important. I know that if 386BSD has elements in it that will help 4.4BSD run well on 386/486 processorsm, and continues to have no licensing restrict- ions in it, that CSRG will cheerfully incorporate Bill's code into their upcoming release. Is that the mark of pride, or of expedience? Is CSRG seeming to attempt to prevent Bill's work from reaching the public? Somewhere I recall seeing Bill assert that some of the work he contributed to CSRG has yet to be released to the public -- that is, was not part of the NET2 release since it wasn't ready when NET2 shipped -- and yet that work was made available to BSDI for their licensed system. I am not sure how this can be verified since very few people have actually seen BSDI's system as it is still in alpha test. I tend to believe Keith Bostic's claim that BSDI has no "head start", that no vendor has copies of CSRG's unreleased code yet, and that BSDI will get it whenever CSRG makes their next release (hopefully 4.4BSD but we'll see how it goes.) I know that BSDI has contributed code to CSRG since I have logged into Okeeffe and seen it in the CSRG source pool. This is analagous to Chris Torek's libc submissions, or my bitstring or cron submissions. Any of us are welcome to submit our code to comp.sources.unix in addition to sending it to CSRG; CSRG is, however, under no obligation to release anything to anybody until they are comfortable that they have something worth shipping. Likewise, if Bill submitted a bunch of his work to CSRG and then did not keep a copy, I'm not sure it falls to CSRG to unravel Bill's submission from the source pool it was integrated into just so they can give it back to Bill. Rest assured that it WILL be made available whenever CSRG makes its next release. If Bill did some work for BSDI that has not been given back to CSRG, I'm afraid he'll have to stand by his employment contract which presumably gives BSDI full rights to whatever work they paid Bill for. I have a sneaking suspicion that Bill's BSDI work was early enough to have been included in BSDI's unlicensed early submission to CSRG, but I don't know the details on this and I'll have to go with what Keith said on this point since he knows CSRG's source pool a lot better than I do. You mentioned that Trent Hein wrote an init for BSDI which was given to CSRG but not made available to your 386BSD effort. That's true. The code is sitting on Okeeffe right now, waiting for the day when it can be tested and integrated and -- someday -- shipped out with 4.4BSD. It seems that your complaint on this issue is that BSDI gave it to CSRG but not to 386BSD, and that CSRG hasn't given it to 386BSD either. How can I convince you that CSRG has nothing to gain from making interrim unintegrated releases available, and that it likewise has nothing to gain from "hoarding" the code that people donate to it. The implication throughout Bill's "Road Not Taken" article and your response to Kolstad is that CSRG is unfairly aiding BSDI. Given that some code has been shared between them but not yet with 386BSD or comp.sources.unix, and that Mike recently left CSRG to take a job at BSDI, I can see why this implication looks plausible. But I know Keith and Rob, and you know them, and I think we both know them well enough to know that this sort of crap is totally beneath them. Please experiment with another possibility: that things are as they seem, and that we all -- all of us, including you and Bill and me and Rob and Keith and Mike and Kirk and Marc and the "cast of thousands" -- are each in our own way trying to save the universe from the looming, lumbering death star. You and Bill believe that you can best serve the cause by making 386BSD avail- able as early as possible and then sending out lots of upgrades and patches. The BSDI folks believe that they can best serve the cause by coming out with a high-quality, supported system with a team of paid programmers to constantly enhance it and fix bugs, and they have put their livelihood on the line such that if they don't make a profit they'll all have to eat stale breadcrumbs. The CSRG folks believe that they can best serve the cause by encouraging the "cast of thousands" to submit lots of useful source code which CSRG can then pick through and build a coherent, integrated system that most people will be able to use "right off the tape". I personally believe that I can best serve the cause by running the world's third-largest anonymous FTP server, sending 386BSD tapes to people who don't have Internet access, writing and running an "ftp by mail" server for UUCP-limited users, writing lots of royalty-free code and publishing it on comp.sources.unix and submitting it to CSRG when it's good enough, moderating comp.sources.unix, and playing mediator in disputes like this one where all the principals know me and are willing to at least listen to what I have to say. We all have different ways of going about it. But it's the same cause, and we are moved by substantially the same spirit. I don't always see the reason behind Bill's actions -- such as his outburst at the Usenix BSD BOF -- but I am still very glad that he's on this ethereal "team" and that he and his code are out there in the world, making it a better place. Can we stop this infighting and get back to the important part of this whole business -- which is writing code? The rubber is meeting the road. None of us have time for this metadiscussion. I hope that you and Bill will someday realize that you don't have any enemies out here, that noone is out to get you or prevent your work from being used, and that noone has knowingly or intent- ionally cheated you or the community at large. I really do believe that we all have the same goal; our approaches and "what we think will work" are the only difference between us. We can do a lot more good by keeping our noses (fingers) to the grindstone (keyboard) and just getting the work done. Cheers, -- Paul Vixie, DEC Network Systems Lab Palo Alto, California, USA "Ready, Fire, Aim" decwrl!vixie vixie!paul