On Wed, Jul 14, 2021 at 11:01:58AM -0400, Clem Cole wrote: > By formal definition, the tarball and the rest of UNIX from Research is and > always has been, '*Open Source*' in the sources were available. *But they > were licensed*. This was fairly typical of much early software BTW. The > binary nature only came about with the minicomputers. It may have been "Open Source" by your definition, but there is a very specific definition of "Open Source(tm)" and it has always been, from the beginning, defined to mean code licensed under terms which meet the Open Source Definition[1] (OSD). The AT&T license, for better or for worse does not mean the terms of the OSD. [1] https://opensource.org/osd > The sources in the tarball were not '*Free and Open Source*' -- which > becomes the crux of the issue. [Sadly the OSS folks have confused this > over the years and that important detail is lost]. Hardly. "Free and Open Source" (FOSS) is a term which developed *after* the the term "Open Source" was coined and trademarked. That term was not created by the "OSS folks", but by people who were trying the solve a political problem. The GPL meets the definition of the Open Source Definition, so GPL-licensed software is "Open Source(tm)". But Stallman objected to that usage, preferring his terminology "Free Software" on the grounds that it came first. So FOSS was a compromise to keep the FSF partisan happy. But to take this back to TUHS, sorry, no code which falls under AT&T License can be called "Open Source(tm)". If AT&T were still trying to sell Unix under its original terms including the AT&T Unpublished Trade Secret "all your student's minds belong to us" license, and tried to claim that Unix was "Open Source", the Open Source Initiative could sue AT&T for trademark infringement. If you must, you could try to claim that AT&T was "Source Available" --- which is a terminology I've seen some used. But I think your assumptions of how easily the AT&T License could be obtained, and how "anyone who wanted it could get it" may be looking at the past with rose-colored classes. Cheers, - Ted
On Wed, Jul 14, 2021 at 01:40:53PM -0400, Theodore Y. Ts'o wrote:
> If you must, you could try to claim that AT&T was "Source Available"
> --- which is a terminology I've seen some used. But I think your
> assumptions of how easily the AT&T License could be obtained, and how
> "anyone who wanted it could get it" may be looking at the past with
> rose-colored classes.
Clem was in "the club". I do remember those times, barely, I was a
bit too young to have a clear view of things. But it certainly
seemed like some Universities made the source pretty available.
UW Madison was not one of those, I had to beg and plead to get
access to the source.
So Clem's memory is fine, his experience was you could get the source.
But that wasn't the universal experience at all, and I agree with
Ted that just getting access to the source doesn't make it remotely
open source.
[-- Attachment #1: Type: text/plain, Size: 3133 bytes --] On Wed, Jul 14, 2021 at 1:40 PM Theodore Y. Ts'o <tytso@mit.edu> wrote: > On Wed, Jul 14, 2021 at 11:01:58AM -0400, Clem Cole wrote: > > By formal definition, the tarball and the rest of UNIX from Research is > and > > always has been, '*Open Source*' in the sources were available. *But > they > > were licensed*. This was fairly typical of much early software BTW. The > > binary nature only came about with the minicomputers. > Please don't go here (again). Yes, it has been trademarked, but the official trademarked term is different from reality --> just like the guy that got a copyright for email and claims to have invented it. People were 'open sourcing' software before you and I were born. They just did not have a name for it - thank you. The real 'father' of Open Source as we think of it today was Prof Don Pederson and his Industrial Liaison Office (ILO) of the EE Dept of UCB in the late 1960s -- long before rms, et al. As 'dop' used to say, I give everything away because then I go in the back door, not the front door like a salesman. MIT/CMU/Stanford et al we often licensing their work. In many ways, CMU and Stanford were two of the worst. The ILO gave away all its products. We would not have the current electronics industry without the work dop and his students produced. As I have also pointed in other email tapes like the original, '1BSD' was managed and distributed by the ILO because dop had set of the infrastructure 10-15 years earlier to send out mag tapes and other IP to 'interested parties.' Yes, computer networks changed the distribution and access medium, but please refrain from trying to rewrite history. The GNU project and FOSS movement that was created took the idea and advanced it, making use of better ways of communicating the ideas, removing the academic clubiness as Larry suggested. Larry is right, if you were a peer organization or maybe a patron of same, getting source was possible. As rms noted, at some point the sources to things go harder and harder to get access. ITS, WAITS, and even CTSS were all written at a time when you go from IBM and DEC their sources - typically on 7 or 9 track mag-tape and were usually available on microfiche. You also got the circuit schematics too. Local modifications to both HW and SW were normal. But starting with the Minis this began to change and it started to get harder and harder. SW started being a revenue source for those companies -- DEC in particular, so they started to be hold back the sources. The rest is history... Folks like rms objected because the behavior they were used to had changed and he and people like him, could do nothing about it. So he created the Gnu project to compete with those commercial products. But just like have been getting 'email' since the late 1960s/early 1970s on my computers, it was not named. Someone body claimed the name later. But the function was old. The same is for sharing software written and given away, now we have a name and a way to describe the behavior. Cheers Clem ᐧ [-- Attachment #2: Type: text/html, Size: 5319 bytes --]
> Arguably ancient PDP-10 operating systems like ITS, WAITS, TENEX were
> somewhat "open" and "free", but it's not a clear cut case.
The open source movement was a revival of the old days of SHARE and other
user groups.
SAP, the SHARE assembly program for the IBM 704, was freely available--with
source code--to all members of the SHARE user group. I am not aware of any
restrictions on redistribution.
Other more specialized programs were also freely available through SHARE. In
particular, Fortran formatted IO was adopted directly from a SHARE program
written by Roy Nutt (who also wrote SAP and helped write Fortran I).
Bell Labs freely distributed the BESYS operating system for the IBM 704.
At the time (1958) no operating system was available from IBM.
IBM provided source code for the Fortran II compiler. In the
fashion of the time, I spent a memorable all-night session with
that code at hand, finding and fixing a bizarre bug (a computed GOTO
bombed if the number of branches was 74 mod 75) with a bizarre cause
(the code changed the index-register field in certain instructions on the
fly--inconsistently). And there was no operating system to help, because
BESYS swapped itself out to make room for the compiler.
Doug
[-- Attachment #1: Type: text/plain, Size: 1481 bytes --] "SHARE. It's not an acronym. It's what we do." *Still* a bad-ass slogan after all these years. Adam On Wed, Jul 14, 2021 at 7:22 PM Douglas McIlroy < douglas.mcilroy@dartmouth.edu> wrote: > > Arguably ancient PDP-10 operating systems like ITS, WAITS, TENEX were > > somewhat "open" and "free", but it's not a clear cut case. > > The open source movement was a revival of the old days of SHARE and other > user groups. > > SAP, the SHARE assembly program for the IBM 704, was freely available--with > source code--to all members of the SHARE user group. I am not aware of any > restrictions on redistribution. > > Other more specialized programs were also freely available through SHARE. > In > particular, Fortran formatted IO was adopted directly from a SHARE program > written by Roy Nutt (who also wrote SAP and helped write Fortran I). > > Bell Labs freely distributed the BESYS operating system for the IBM 704. > At the time (1958) no operating system was available from IBM. > > IBM provided source code for the Fortran II compiler. In the > fashion of the time, I spent a memorable all-night session with > that code at hand, finding and fixing a bizarre bug (a computed GOTO > bombed if the number of branches was 74 mod 75) with a bizarre cause > (the code changed the index-register field in certain instructions on the > fly--inconsistently). And there was no operating system to help, because > BESYS swapped itself out to make room for the compiler. > > Doug > [-- Attachment #2: Type: text/html, Size: 1962 bytes --]
[-- Attachment #1: Type: text/plain, Size: 3413 bytes --] Thank you, Doug. On Wed, Jul 14, 2021 at 10:22 PM Douglas McIlroy < douglas.mcilroy@dartmouth.edu> wrote: > The open source movement was a revival of the old days of SHARE and other > user groups. > Amen, my basic point, although I was also trying to pointing at that these user groups got started b*ecause the vendors gave the sources to their products out.* We SHARED patches and features. DECUS started out the same way. For instance, many/most PDP-10 OS's used the DEC compilers and often even found a way to run TOPS-10 binaries by emulating the UUOs. The IBM/360 world worked pretty much the same way. My own experience was that the compilers (e.g WATFIV-FTNG-ALGOLW-PL/1) and language interpreters (APL-Snolbol) for the TSS and MTS had been 'ported' from the IBM-supplied OS [my own first job was doing just that]. The same story was true for the PDP-8 with DOS-8/TSS-8 and the like. By the time of the PDP-11, while some of the DEC source code was available (such as the Fortran-IV for RT-11/RSX), since it took at PDP-10/BLISS to support it, DEC had it its protection - so moving it/stealing it - would have been harder. By the time of the VAX, DEC was charging a lot of money of SW and it was actually a revenue stream, so they keep a lot more locked up and had started to do the same with PDP-10 world. So, the available/unavailable source issue came when things started to get closed up, which really started with the rise of the SW industry and making revenue with the use of your SW. OEMs and IVSs started to be a lot less willing to reveal what they thought was their 'special sauce.' Some/many end-users started to balk. RMS just took it to a new level - just look at how he reacted to Symbolics being closed source :-) The question that used to come up (and still does not an extent) is how are the engineers and teams of people that developed the SW going to be paid/renumerated for their work? The RMS/GNU answer had been service revenue [and living like a student in a rent-controlled APT in Central Sq]. What has happened for most of the biggest FOSS projects, the salaries are paid for firms like my own that pay developers to work on the SW and most FOSS projects die when the developer/maintainer is unable to continue (if not just gets bored). In fact, [I can not say I personally know this - but have read internal memos that make the claim], Intel pays for more Linux developers and now LLVM developers than any firm. What's interesting is that Intel does not really directly sell its HW product to end-users. We sell to others than use our chips to make their products. We have finally moved to the support model for the compilers (I've personally been fighting that battle for 15 years). So back to my basic point ... while giving the *behavior* a name, the *idea *of "Open Source" is really not anything new. While it may be new in their lifetime/experience, it is frankly at minimum a sad, if not outright disingenuous, statement for the people to try to imply otherwise because they are unwilling to look back into history and understand, much less accept it as a fact. Trying to rewrite history is just not pretty to witness. And I am pleased to see that a few folks (like Larry) that have lived a little both times have tried to pass the torch with more complete history. Clem. ᐧ [-- Attachment #2: Type: text/html, Size: 6380 bytes --]
On Thu, Jul 15, 2021 at 11:07:10AM -0400, Clem Cole wrote: > In fact, [I can not say I personally know this - but have read internal > memos that make the claim], Intel pays for more Linux developers and now > LLVM developers than any firm. What's interesting is that Intel does not > really directly sell its HW product to end-users. We sell to others than > use our chips to make their products. We have finally moved to the > support model for the compilers (I've personally been fighting that battle > for 15 years). That claim is probably from the data collected from the Linux Foundation, which publishes these stats every year or two. The most recent one is here: https://www.linuxfoundation.org/wp-content/uploads/2020_kernel_history_report_082720.pdf The top ten organizations responsible for commits from 2007 -- 2019: (None) 11.95% Intel 10.01% Red Hat 8.90% (Unknown) 4.09% IBM 3.79% SuSE 3.49% Linaro 3.17% (Consultant) 2.96% Google 2.79% Samsung 2.58% "None" means no organizational affiliation (e.g., hobbyists, students, etc.) "Unknown" means the organization affiliation couldn't be determined. For more recent data, if you look at the commits for the 5.10 release (end of 2020), the top ten list by organizations looks like this: Huawei 8.9% Intel 8.0% (Unknown) 6.6% (None) 4.9% Red Hat 5.7% Google 5.2% AMD 4.3% Linaro 4.1% Samsung 3.5% IBM 3.2% For the full list and more stats, see: https://lwn.net/Articles/839772/ > So back to my basic point ... while giving the *behavior* a name, the *idea > *of "Open Source" is really not anything new. I do think there is something which is radically new --- which is that it's not a single company publishing all of the source code for a particular OS, whether it's System/360 or the PDP-8 Disk Operating System, or whatever. In other words, it's the shared nature of the collaboration, which partially solves the question of "who pays" --- the answer is, "lots of companies, and they do so when it makes business sense for them to do so". Intel may have had the largest number of contributions to Linux historically --- but that was still 10%, and it was eclipsed by people with no organizational affliation, and in the 5.10 kernel Huawei slightly edged out Intel with 8.9% vs 8.0% contributions. I completely agree with you that one of the key questions is the business case issue. Not only who pays, but how do they justify the software investment to the bean counters? Of course, the "Stone Soup" story predates computers, so this certainly isn't a new business model. And arguably the X Window Systems and the Open Software Foundation also had a similar model where multiple companies contributed to a common codebase, with perhaps mixed levels of success. The thing which Linux has managed to achieve, however, is the fact that there is a large and diverse base of corporate contributions. That to me is what makes the Linux model so interesting, and has been a reason for its long-term sustainability. Other companies may have been making their source code availble, but the underlying business model behind their "source available" practices was quite different. Cheers, - Ted
[-- Attachment #1: Type: text/plain, Size: 1317 bytes --] The thing which Linux has managed to achieve, however, is the fact > that there is a large and diverse base of corporate contributions. > That to me is what makes the Linux model so interesting, and has been > a reason for its long-term sustainability. > > Although from a somewhat different perspective, it's also why the Linux kernel syscall interface is so unruly, right? You've got your...some number in the small dozens of common syscalls, which are already present for the most part in v6 or v7. These are the ones I, at least, think of when I think of the Unix manual, section 2. And then you've got all the other calls added in by (usually) this database vendor or that storage vendor or the other display adapter vendor to make their stuff work more efficiently. And obviously there's a tradeoff there. Elegance departs, and you've probably introduced some security risk because these syscalls are not nearly as well-exercised as the common ones, but on the other hand you have these large companies paying to work on the kernel, and you have them supporting their product on Linux systems because the system can be bent into accommodating them more easily, and it will run better there than on OSes where they don't get to introduce features that benefit their products, which further drives adoption. [-- Attachment #2: Type: text/html, Size: 1681 bytes --]
On 7/15/21, Adam Thornton <athornton@gmail.com> wrote:
> The thing which Linux has managed to achieve, however, is the fact
>> that there is a large and diverse base of corporate contributions.
>> That to me is what makes the Linux model so interesting, and has been
>> a reason for its long-term sustainability.
>>
>>
> Although from a somewhat different perspective, it's also why the Linux
> kernel syscall interface is so unruly, right?
>
> You've got your...some number in the small dozens of common syscalls, which
> are already present for the most part in v6 or v7. These are the ones I,
> at least, think of when I think of the Unix manual, section 2.
>
> And then you've got all the other calls added in by (usually) this database
> vendor or that storage vendor or the other display adapter vendor to make
> their stuff work more efficiently.
>
> And obviously there's a tradeoff there. Elegance departs, and you've
> probably introduced some security risk because these syscalls are not
> nearly as well-exercised as the common ones, but on the other hand you have
> these large companies paying to work on the kernel, and you have them
> supporting their product on Linux systems because the system can be bent
> into accommodating them more easily, and it will run better there than on
> OSes where they don't get to introduce features that benefit their
> products, which further drives adoption.
The last time I looked it was actually FreeBSD that had the most
system calls (more than 500). Linux had more or less around the same
number as OpenBSD (more than 300).
UNIX V7 had around 50 -- this is still the golden standard, but
obviously a lot has changed since then...
--Andy
[-- Attachment #1: Type: text/plain, Size: 2617 bytes --] On Thu, Jul 15, 2021 at 3:33 PM Theodore Y. Ts'o <tytso@mit.edu> wrote: > > So back to my basic point ... while giving the *behavior* a name, the > *idea > > *of "Open Source" is really not anything new. > > I do think there is something which is radically new --- which is that > it's not a single company publishing all of the source code for a > particular OS, whether it's System/360 or the PDP-8 Disk Operating > System, or whatever. Ted - that *is what* Doug pointed out!!! They did not create anything that was new. SHARED / DECUS / USENIX and the like were providing that exact same function starting in the late 1950s!!! Companies and Universities all pooled their resources to make things better and to get new and improved solutions. Sometimes they started with things that come from the original OEM. Also often they created their own technology and made it available to everyone. Sometime they combine both. And it was a 'bazaar where everyone had access and you chose to use it to not. Sounds pretty familiar, BTW. What >>has<< changed (dramatically) was the *method* and *ability* to *distribute* your work and/or the manner you *obtained* someone else's efforts. Today we download via the Web (much less ftp from a public area), which is much more convenient than becoming a member of an organization and having to obtain (typically for a small $50-$100 trape copying fee) a 9-track distribution tape. But even the concept of 'free' is really not new as I said. Things like UCB's ILO used that model for a long time. MIT, CMU, Stanford, Univerity of Waterloo, Cambridge, et al, just made their work to any interested parties. But due to the new way of being *interconnected *and a *much better distribution scheme* that indeed is a huge feature. But please understand 'open source and collective sharing/working together is not a new thing that just appeared with the Gnu project and was accelerated and taken to a new level with the Linux work. I personally blame esr's book for that beginning of the rewriting of history/kicking the previous generations in the shins, as readers of it, or worse readers of summations of it, miss the big picture instead of the reality of standing on other shoulders. I do want to give create for the cool and important things that have come. I just want to make sure we don't forget the success of the modern world is 100% dependent on two important things: moore's law to make things more economic and the hard work of a lot of people that came before (now and before me for that matter). ᐧ [-- Attachment #2: Type: text/html, Size: 4662 bytes --]
[-- Attachment #1: Type: text/plain, Size: 3956 bytes --] I remember going to one of those cattle-call hiring events. I wanted to speak with the Intel compiler guy and when I got up to him, all he said was “Ganapathi”. I actually knew who/what hw was talking about. So, has Intel killed their own compiler toolset? Joe McGuckin ViaNet Communications joe@via.net 650-207-0372 cell 650-213-1302 office 650-969-2124 fax > On Jul 15, 2021, at 12:33 PM, Theodore Y. Ts'o <tytso@mit.edu> wrote: > > On Thu, Jul 15, 2021 at 11:07:10AM -0400, Clem Cole wrote: >> In fact, [I can not say I personally know this - but have read internal >> memos that make the claim], Intel pays for more Linux developers and now >> LLVM developers than any firm. What's interesting is that Intel does not >> really directly sell its HW product to end-users. We sell to others than >> use our chips to make their products. We have finally moved to the >> support model for the compilers (I've personally been fighting that battle >> for 15 years). > > That claim is probably from the data collected from the Linux > Foundation, which publishes these stats every year or two. The most > recent one is here: > > https://www.linuxfoundation.org/wp-content/uploads/2020_kernel_history_report_082720.pdf > > The top ten organizations responsible for commits from 2007 -- 2019: > > (None) 11.95% > Intel 10.01% > Red Hat 8.90% > (Unknown) 4.09% > IBM 3.79% > SuSE 3.49% > Linaro 3.17% > (Consultant) 2.96% > Google 2.79% > Samsung 2.58% > > "None" means no organizational affiliation (e.g., hobbyists, students, > etc.) "Unknown" means the organization affiliation couldn't be > determined. > > For more recent data, if you look at the commits for the 5.10 release > (end of 2020), the top ten list by organizations looks like this: > > Huawei 8.9% > Intel 8.0% > (Unknown) 6.6% > (None) 4.9% > Red Hat 5.7% > Google 5.2% > AMD 4.3% > Linaro 4.1% > Samsung 3.5% > IBM 3.2% > > For the full list and more stats, see: https://lwn.net/Articles/839772/ > >> So back to my basic point ... while giving the *behavior* a name, the *idea >> *of "Open Source" is really not anything new. > > I do think there is something which is radically new --- which is that > it's not a single company publishing all of the source code for a > particular OS, whether it's System/360 or the PDP-8 Disk Operating > System, or whatever. > > In other words, it's the shared nature of the collaboration, which > partially solves the question of "who pays" --- the answer is, "lots > of companies, and they do so when it makes business sense for them to > do so". Intel may have had the largest number of contributions to > Linux historically --- but that was still 10%, and it was eclipsed by > people with no organizational affliation, and in the 5.10 kernel > Huawei slightly edged out Intel with 8.9% vs 8.0% contributions. > > I completely agree with you that one of the key questions is the > business case issue. Not only who pays, but how do they justify the > software investment to the bean counters? Of course, the "Stone Soup" > story predates computers, so this certainly isn't a new business > model. And arguably the X Window Systems and the Open Software > Foundation also had a similar model where multiple companies > contributed to a common codebase, with perhaps mixed levels of > success. > > The thing which Linux has managed to achieve, however, is the fact > that there is a large and diverse base of corporate contributions. > That to me is what makes the Linux model so interesting, and has been > a reason for its long-term sustainability. > > Other companies may have been making their source code availble, but > the underlying business model behind their "source available" practices > was quite different. > > Cheers, > > - Ted [-- Attachment #2: Type: text/html, Size: 8933 bytes --]
On Thu, Jul 15, 2021 at 04:30:15PM -0400, Clem Cole wrote:
>
> Ted - that *is what* Doug pointed out!!! They did not create anything that
> was new. SHARED / DECUS / USENIX and the like were providing that exact
> same function starting in the late 1950s!!! Companies and Universities all
> pooled their resources to make things better and to get new and improved
> solutions. Sometimes they started with things that come from the
> original OEM. Also often they created their own technology and made it
> available to everyone. Sometime they combine both. And it was a
> 'bazaar where everyone had access and you chose to use it to not. Sounds
> pretty familiar, BTW.
I remember looking at the DECUS program catalog for the PDP-8, and I
seem to recall that for the most part, individuals were sharing their
programs with others. In that way, it wasn't all that different from
say, CPM/UG, and HUG (Heathkit Users Group). But the thing is, for
the most part, it was a single author sharing individual programs, and
often changes were not accepted back.
Consider the history of Bill Jolitz and 386BSD, and the collection of
patches that eventuallya became NetBSD and FreeBSD, which was formed
because they were frustrated that they couldn't get their patch sets
back into Jolitz's code base. Technology plays a part, in that it
enables the change. But it's not just about technology. There is
also a very strong social component. Even when you were richly
interconnected at the network level, this does not guarantee that will
be willing to be richly interconnected in terms of accepting patch
sets from people who you may not know across the Internet, into *your*
program, for which you are the author and high priest.
I don't remember the exact date, but it would have been in the early
90's, when at the time I was already contributing patches to Linux,
and where ftp and e-mail and applying patches via context diffs was
very much available. At that time, we were interested in getting
support for MIT Project Athena's Hesiod extenstions into the BIND
distributions (we had just been carrying patches against BIND for many
years).
In order to get those patches integrated, Paul Vixie invited me to his
house in Redwood City, and so I flew from Boston to San Francisco,
carrying my Linux laptop with the BIND patches, and we got the patches
integrated into master BIND sources. Paul was a gracious host, and it
was lovely that I got to spend some time with him. But it was
interesting that my physical presence was needed, or at least highly
useful, in terms of getting those patches into BIND. Requiring
physical presence to get patches integrated.... doesn't scale.
And so it wasn't a matter of technology, since the technology for
Linus, who didn't know me from Adam in 1991, to accept patches from me
implementing BSD Job Control, was certainly available when I was
working with Paul to get the Hesiod changes integrated into BIND. But
like with Jolitz and 386BSD, it's a mindset thing, not just technology.
I also want to emphasize again, the question of business model is also
something which I think is different, and *important*. It's one thing
for Academics and Researchers to be willing to give changes away to
anyone who wants. It's quite another for a company to give away their
intellectual property in such a way that it can actually be used by
their competitors, either because that's the social convention, or
because it's enforced by the license. Was the practices we use today
for Linux built on the traditions of comp.sources.unix, and BSD, and
AT&T Research, and IBM making sources available for System/360, yadda,
yadda, yadda? Of course! I'm not denying that.
But at the same time, to claim that nothing is new under the Sun, and
*all* of this had been done decades earlier, is also not the whole
story. And to call IBM releasing System/360, when they retained
control of the license, and wasn't accepting any changes back, and
*darned* well would have sued anyone trying to use that code on
non-IBM computers into a smoking crater, as "Open Source" can be
highly misleading, because that is not what most people associate with
the term "Open Source" today.
And if we take a look at what AT&T Lawyers did with the Unix source
code, at some point, it most *defintely* was the antithesis of "Open
Source". Which would lead me to assert that Unix was never really
released under what today we would call, "Open Source".
Cheers,
- Ted
I was part of a discussion about a bug in the DECUS tape in Leeds uni,
in '82-84 window. I was a very small part I might add, not the
principal. I can't remember the package. It was probably trivia, like
walking a specific SYS$SYSTEM value in a way which was dangerous or
encoded assumptions about device:directory:user models in VMS.
The feedback I got from this process was "thanks, we'll think about
it" was closure, for those days. We'd been pretty specific about a
fix. I got the sense the tape was an annual affair. And the likelihood
of our "patch" being both accepted, and added to the next round of the
tape was low-to-zero because everyone wanted "moar" and so people
focussed on adding things, not fixing things.
The exception here was compilers: people always want bugs fixed in a
compiler. Or the NAG library, but both compilers (language spec) and
NAG (strict maths formalisms about correctness) had policed mechanisms
to accept user input, validate, run through a remorselessly tight
compliance check, and emit, if it survived.
A bug in the implementation of MUD for dec-10? ok, so the word
"potato" is misspelled on one screen. Move on.
On Fri, Jul 16, 2021 at 11:59 AM Theodore Y. Ts'o <tytso@mit.edu> wrote:
>
> On Thu, Jul 15, 2021 at 04:30:15PM -0400, Clem Cole wrote:
> >
> > Ted - that *is what* Doug pointed out!!! They did not create anything that
> > was new. SHARED / DECUS / USENIX and the like were providing that exact
> > same function starting in the late 1950s!!! Companies and Universities all
> > pooled their resources to make things better and to get new and improved
> > solutions. Sometimes they started with things that come from the
> > original OEM. Also often they created their own technology and made it
> > available to everyone. Sometime they combine both. And it was a
> > 'bazaar where everyone had access and you chose to use it to not. Sounds
> > pretty familiar, BTW.
>
> I remember looking at the DECUS program catalog for the PDP-8, and I
> seem to recall that for the most part, individuals were sharing their
> programs with others. In that way, it wasn't all that different from
> say, CPM/UG, and HUG (Heathkit Users Group). But the thing is, for
> the most part, it was a single author sharing individual programs, and
> often changes were not accepted back.
>
> Consider the history of Bill Jolitz and 386BSD, and the collection of
> patches that eventuallya became NetBSD and FreeBSD, which was formed
> because they were frustrated that they couldn't get their patch sets
> back into Jolitz's code base. Technology plays a part, in that it
> enables the change. But it's not just about technology. There is
> also a very strong social component. Even when you were richly
> interconnected at the network level, this does not guarantee that will
> be willing to be richly interconnected in terms of accepting patch
> sets from people who you may not know across the Internet, into *your*
> program, for which you are the author and high priest.
>
> I don't remember the exact date, but it would have been in the early
> 90's, when at the time I was already contributing patches to Linux,
> and where ftp and e-mail and applying patches via context diffs was
> very much available. At that time, we were interested in getting
> support for MIT Project Athena's Hesiod extenstions into the BIND
> distributions (we had just been carrying patches against BIND for many
> years).
>
> In order to get those patches integrated, Paul Vixie invited me to his
> house in Redwood City, and so I flew from Boston to San Francisco,
> carrying my Linux laptop with the BIND patches, and we got the patches
> integrated into master BIND sources. Paul was a gracious host, and it
> was lovely that I got to spend some time with him. But it was
> interesting that my physical presence was needed, or at least highly
> useful, in terms of getting those patches into BIND. Requiring
> physical presence to get patches integrated.... doesn't scale.
>
> And so it wasn't a matter of technology, since the technology for
> Linus, who didn't know me from Adam in 1991, to accept patches from me
> implementing BSD Job Control, was certainly available when I was
> working with Paul to get the Hesiod changes integrated into BIND. But
> like with Jolitz and 386BSD, it's a mindset thing, not just technology.
>
> I also want to emphasize again, the question of business model is also
> something which I think is different, and *important*. It's one thing
> for Academics and Researchers to be willing to give changes away to
> anyone who wants. It's quite another for a company to give away their
> intellectual property in such a way that it can actually be used by
> their competitors, either because that's the social convention, or
> because it's enforced by the license. Was the practices we use today
> for Linux built on the traditions of comp.sources.unix, and BSD, and
> AT&T Research, and IBM making sources available for System/360, yadda,
> yadda, yadda? Of course! I'm not denying that.
>
> But at the same time, to claim that nothing is new under the Sun, and
> *all* of this had been done decades earlier, is also not the whole
> story. And to call IBM releasing System/360, when they retained
> control of the license, and wasn't accepting any changes back, and
> *darned* well would have sued anyone trying to use that code on
> non-IBM computers into a smoking crater, as "Open Source" can be
> highly misleading, because that is not what most people associate with
> the term "Open Source" today.
>
> And if we take a look at what AT&T Lawyers did with the Unix source
> code, at some point, it most *defintely* was the antithesis of "Open
> Source". Which would lead me to assert that Unix was never really
> released under what today we would call, "Open Source".
>
> Cheers,
>
> - Ted
On Thu, Jul 15, 2021 at 12:49:07PM -0700, Adam Thornton wrote: > Although from a somewhat different perspective, it's also why the Linux > kernel syscall interface is so unruly, right? > > You've got your...some number in the small dozens of common syscalls, which > are already present for the most part in v6 or v7. These are the ones I, > at least, think of when I think of the Unix manual, section 2. Actually, a more reason why we have so many system calls is that there is a strong belief that backwards compatibility is important. So much so that a Linux, or even Minix binary, from the early 1990's, will still run on a modern Linux kernel today. So that means we have separate system calls for wait, wait3, wait4, etc. Support for openat(2), linkat(2), etc., which got added to Posix by way of Solaris, need to be implemented as a separate system calls, since we need to keep the original open(2) system calls. Etc. > And obviously there's a tradeoff there. Elegance departs, and you've > probably introduced some security risk because these syscalls are not > nearly as well-exercised as the common ones, but on the other hand you have > these large companies paying to work on the kernel, and you have them > supporting their product on Linux systems because the system can be bent > into accommodating them more easily, and it will run better there than on > OSes where they don't get to introduce features that benefit their > products, which further drives adoption. In many cases, inside the kernel, system calls like wait(2) and wait3(2) will be implemeanted in terms of wait4(2), the implementation of open(2) and openat(2) share a common codepath, and so on. So although the *number* of system calls may look big, in many cases there are shared code paths. This is especially true for the system calls that implement 64-bit support, where the old legacy 32-bit system calls are just wrappers that call the 64-bit implementations of those system calls. There are also some architectures such as Alpha, where some of Linux's system calls used the Ultrix system call numbering scheme, and other Ultrix system calls were emulated, so that the Netscape browser, which was provided in binary form only for Ultrix on Alpha, would run on Linux Alpha. Being able to run a subset of Ultrix binaries on Linux Alpha was certainly a hack, but the ability to run a web browser (there were no open source web browsers at that time) certainly drove adoption for at least some users. Was there perhaps some security risk in doing that? Probably, although people cared a lot less about that in the early 90's. And these days Linux support for architectures such as Alpha and HP's PA-RISC are done only by folks who do for fun. :-) - Ted P.S. At one point Linux x86 could also run SCO Unix binaries. Which led to an amusing situation where MIT had purchased a site license for a proprietary spreadsheet program that ran on SCO Unix, for use by students who would be running Linux. I worked with someone at that company (who eventually became one of the founders of Red Hat) and he gave me a custom application binary that checked for MIT's IP network address prefix, which was how the site license enforcement was implemented. Turns out his development environment was Linux, cross-compiling to create a SCO Unix binary, because the Linux development environment was more developer friendly than SCO's. And so here he was, building a SCO Unix application on a Linux development machine, and then handing it to me so that our students could run that SCO Unix application on Linux systems at MIT. The joys of syscall/OS emulation. :-)