primary platform and the future direction

[Mark Tarver]

I like the appeal but I was hoping for some information on a primary platform.  I know, I must be getting annoying with my singular focus and constant questioning about the primary target of choice.

I put this in a separate thread from the appeal.  Let me draw the projected line of work.

The answer to the identity of the primary platform is determined by many things.  First, whether I build it or another.  Second what people actually want.  In this matter the Shen license does not offer any decision - and rightly so.  Its job is simply to protect the kernel. Therefore my position in this matter is, at best, primus inter pares.

As regards whether I build it or another; this depends partly on funding.  Let us suppose funding is found then this is what is likely to happen.

1.  I would encourage those working on the JVM/Java port derived from Hakan's work to develop this port as much as possible.  Likewise for Ramil's port to C and in fact for anybody wishing to develop any of the platforms.  For myself I would not contribute to the JVM/Java code - not because it is unworthy, but because I do not know enough Java to do so.  Java is not my language.  Therefore I would remain in an advisory capacity here, as with other ports like Python and C etc.  

2.  I've discussed this idea with Greg and he agrees; I would put together an exhaustive set of standard licensing tests for any Shen port which would provide a computable test for passing something as Shen.  This would remove any lingering fear of me exercising some arbitrary droit de seigneur over other people's work.  If it passes the standard tests, it is legal and that is the end of the matter.   This should satisfy all but the religious.

3.  TBoS 2nd edition comes out.

4.  Re .primary platform; my personal choice would be to develop the CL platform for Shen, basing my choice on SBCL for several reasons.  

  a. I know CL well. 

  b. SBCL has become more stable with recent releases and we've ironed out the annoying niggles in the Shen port (like warnings etc).  

  c. SBCL has an excellent performance.  

  d. Recent releases have threads allowing multicore processing in Shen.  

  e. We can interface SBCL to GTk+.  

  f.  Last it is possible to produce stand alone executables in SBCL/GTk+; this is established technology. I would hope to port much of the technology of Qi/tk over to this platform.  

However I emphasise this is my personal path; and others may want to go in a different direction and are free to do so.

5.  We implement the Ring to allow free sharing of our work under Shen/BSD.

The way ahead is pretty clear and our technology is set to achieve it.  It is the open source funding engine that simply refuses to deliver the power and this has been the major weak point.

Mark    

[h...@ancell-ent.com]

From: Mark Tarver <dr.mt...@gmail.com>
Date: Sat, 17 Aug 2013 23:51:51 -0700 (PDT)

[ On a primary platform, getting funding, etc. ]

2. I've discussed this idea with Greg and he agrees; I would put
together an exhaustive set of standard licensing tests for any Shen
port which would provide a computable test for passing something as
Shen. This would remove any lingering fear of me exercising some
arbitrary droit de seigneur over other people's work. If it passes
the standard tests, it is legal and that is the end of the matter.
This should satisfy all but the religious.

This would be a great opportunity to radically simplify the current
license, with its inherently ambiguous and therefore legally dangerous
*form* of saying the same thing as many 5 different ways.

It could be reduced to three essentials, a "Thou shalt not break the
spec" preamble, the detailed terms and conditions "You demonstrate you
haven't broken the spec by passing the license test suite, which will be
freely available and usable" (the later needed to avoid the Sun Java
Technology Compatibility Kit (TCK) trap that Apache tripped over) and
"Use of a Shen implementation is free [as detailed here].", and an
explanation of how to mechanically apply and refer to the license. I.e.
following the *form* of the GPL 2.0; I myself also don't like it, *but*
it's passed the tests of legal clarity, enforceability, etc. in US and
German courts: http://www.gnu.org/licenses/old-licenses/gpl-2.0.html

Every bit that's added to that needs to be carefully considered to avoid
making the job of legal analysis too difficult. See for example how the
GPL 2.0 breaks out a separate less or not legally binding FAQ.

- Harold

[Mark Tarver]

I don't have any problem with this analysis at all.  

Mark

[h...@ancell-ent.com]

I'm very glad to hear that; many people including myself have not been able to get past the license, leaving Shen a language for self-education that otherwise *cannot* change the world.

 

While IANAL, in the absence of a pro bono lawyer volunteer, I volunteer to draft a new license in form, call it the Shen License 2.0, given a pledge it'll be received in the intended spirit and if satisfactory replace "1.0" for future releases of Shen once there is a formal compatibility suite.

 

Don't know if it needs to be in the license or not (better safe than sorry?), but as we previously agreed implementations must pass with their default flags, and non-default flags are allowed to change behavior.  (In the area of my interest a flag for more purely functional behavior is needed for best SMP performance in the style of Clojure, although the default wouldn't be a second class citizen, just slower.)

 

- Harold

[Jacob]

"However I emphasize this is my personal path; and others may want to go in a different direction and are free to do so."

I do think that using the JVM as a primary platform is a difficult choice for a target; the sheer amount of crap you would need to learn is staggering...

However, I think the leader of Shen, the head of its development, should not choose esoteric, unfamiliar targets(LISP anything will instantly reduce adoption, I guarantee it). Since Java would be an annoying target for you, how about targeting the DartVM and building the IDE in/on the web?  This would not require some of the annoying low level mucking around that Java byte code entails and it would not be as bad(both architecturally and performance wise) as targeting javascript.  Dart code could be web bytecode(like Seth Ladd once said, Dart code is a great web "bytecode") that Shen spits out!  Let the DVM handle ubiquity and performance, you could handle abstraction and sugar! I really think that we(as in you) could gain great momentum by being an early adopter of their platform.  The JVM is a very competitive place to host a language but the DVM is not, not yet at least.

In Summary...

1. If not JVM then DVM

2. DVM is multi-platform

3. DVM is fast    http://bonovox.be/blog/dart-vs-java-the-deltablue-benchmark/, https://www.dartlang.org/performance/, https://www.youtube.com/watch?v=huawCRlo9H4 

4. Dart "bytecode"(Dart code) is abstract compared to Java bytecode, making the target easier.

5. DVM takes advantage of things like polymer making GUI treatment a piece of cake.  By targeting Dart you get html5+ access for free!  https://www.youtube.com/watch?v=fqULJBBEVQE

6. DVM has a good concurrency model(like Erlang), making it superior to javascript only targets.

7. DVM compiles to javascript anyway, it has good performance when it does so.

8. DVM does tree shaking!

9. We could be one of the first languages to target this exciting new platform that has deep pockets supporting it, a potential boon for Shen!

10. I need more points dangit!

[Mark Tarver]

That's fine - email it, but note that 

a.  The license is being clarified wrt a key term rather than substantively changed.

b.   The test suite will take a time to assemble - it needs to include all the code in TBoS 2nd edition and more.

c.   We're busy right now;  the time frame is before the end of this year.

Mark

[Mark Tarver]

As I've said this is one path; not the only one.  But it has several advantages.

1.  Provided funding is secured, it will be delivered because I will have the time to do it.   Rather than waiting for vapourware or being held to ransom by somebody with a private agenda.

2.  It can be delivered in a short time frame.

3.  Ignoring the emotional associations of Lisp (we are a Lisp anyway, albeit a powerful one, so this is like being allergic to oneself) it ticks all the boxes I'd expect of a platform.

4.  It can exist as a placeholder to inspire other designs while people perfect their ideas.  It can be hollowed out and replaced by other platforms. Qi served as you remember as a blueprint for Shen though it was heavily embedded in Lisp; it was hollowed out and made portable.

There is also a 5.

If you have the Ring, then the identity of the primary platform is less significant because wherever you sit, you can generate the code for all the other platforms.   

So in all its not a bad default strategy.  

9. We could be one of the first languages to target this exciting new platform that has deep pockets supporting it, a potential boon for Shen!

Aha, now there you are talking sunshine :).  This is why I said it is a default.  Because if a major sponsor puts his mouth and his money behind a platform of his choice then the direction could change.  And you are back to the appeal again, where the action is right now. 

Mark 

[Marko Kocić]

Hi Mark,

I'm glad that Shen license is being reconsidered again, since I agree that it is right now preventing Shen adoption in some circles. I'm just not sure about the direction of change regarding the new license.

So far, main complain that I heard about licensing is not that its content is wrong, but that it's custom. Almost every developer, manager, company know about major license types like GPL, BSD, MIT, LGPL and Apache. Most large companies have already policies regarding software / library licensing in place (MIT good, GPL bad, LGPL need approval). That means in any company, developer that wants to introduce Shen will have to not only show its technical superiority, but also go through legal department maze to get it approved, or it will be automatically rejected since some companies don't have a specialized legal department and just use common knowledge about licensing. Also, the custom licenses means that Shen will never become a part of Debian or any other Linux distribution, which would also affect adoption.

One possible approach to mitigate this is to just use one of the already existing opens source licenses, possibly with extensions. For example GCC is GPL, with extension that states that binary product of compilation doesn't fall into GPL although it contains portions of GPL licensed C runtime. Similar approach has Java, which is also GPL, meaning that anyone can fork it and update it, but cannot call it Java unless it passes a full test suite. Something similar might also work for Shen, but IANAL.

Cheers,
Marko

[Mark Tarver]

Stallman is devious and socially disgusting, so GPL in any flavour is out.   There is no way I would want any association with the FSF.   

Lest you think this is harsh; a war broke out some time back between the BSD people running OpenBSD and Stallman.  The resulting thread had over 900 messages in it and I read just about every one. The BSD people challenged RMS very strongly in that thread. I suggest anybody having lingering elements of respect for Stallman read it closely.  

http://openbsd.7691.n7.nabble.com/Real-men-don-t-attack-straw-men-td55042i60.html

In it he says

I don't think it is wrong in general to relicense code from BSD to GPL.  However, in some cases I think it is more useful not to do so, in order to contribute changes back to the original BSD-licensed project.

In fact there is a history here; and you should read Theo de Raadt's remarks on BSDed code which was taken over wholesale and relicensed to GPL w.o. any significant change.     .   

http://undeadly.org/cgi?action=article&sid=20070913014315

Apparently this came close to a legal case with that one.   The FSF was rightly described by the OpenBSD as a Mafia and Stallman was derided as a hypocrite and a liar and I agree with their assessment.  I'd advise people to steer well clear of the GPL, which is a commerce killer.

Desktop Linux is a victim of its own following and has failed even according to the people actively involved in it.  See Miguel de Icaza, one of the founders of the GNOME project, 

http://www.omgubuntu.co.uk/2012/08/gnome-founder-says-desktop-linux-is-dead

He cites rapid development ‘breaking both open source and proprietary software alike’, as well as the ‘incompatibility across Linux distributions’ as being key reasons for Linux’s failure to take hold.

“This killed the ecosystem for third party developers trying to target Linux on the desktop. You would try once, do your best effort to support the “top” distro or if you were feeling generous “the top three” distros. Only to find out that your software no longer worked six months later.”

Exactly what I said in the license page would kill desktop Linux.  The OS design methodology sucks. Billions of man hours, millions of pounds and they end up with 1.25% market share.  Pathetic.  Not even Dell could move Linux and they discounted £50 off Linux machines.

So even if Windows 8 is crap - and it seems to be - Linux is not there to pick up the ball because it still sucks and Lunduke has made an annual lecture out of it sucking.   And even if somebody does come up with a good desktop Linux, the years of suckage will put people off trying it and the philosophy of 'it has to be free' will suck the guts out of the funding so it will too eventually suck.  Ubuntu has been a long term loss-maker for Shuttleworth and if wasn't for his deep pockets, it would have long since died.  He can blow $20M on a flight into space, so for him its a hobby.

The other aspect is that ranting on about the evils of proprietary software and saying 'we can't use Shen because it is not BSD/GPL' has not helped Linux one bit.  People want to run proprietary stuff on their machine, to play games and so forth and making your OS unfriendly and advertising your hostility does not pay.   OS fanaticism killed off that possibility.  If Shen puts them off because it does not match their standards of purity, I can safely say that those standards have and will cost them dearly.

Regarding the corporations, that is different.  But dealing with corporations requires sagacity.   That's all I want to say on that right now.

Mark

[Mark Tarver]

I said it is rational to avoid placing the Shen project under the auspices of a license that is written by an organisation using sharp practice.  I think that is true.  We do not want to be ruled by Stallman or his lawyer Eben as Depper found out.

http://www.redhat.com/archives/redhat-install-list/2001-August/msg00346.html

But let's put the nasty goings-on around the FSF aside and just put the GPL license next to the Shen license.    What does the Shen license forbid - essentially tampering with the kernel code and issuing a faulty distro.  Who is put out by that?  Only somebody with very detailed knowledge and some agenda of his own.  If somebody did distribute a faulty kernel, people would probably just think they were crazy.  Making (+ 7 8) come out 12 is silly. There may be people like that, and they may appear but certainly not in number.   The macro system is a much easier way of configuring Shen. So number of people inconvenienced = k where k is very small.

The GPL on the other hand prevents any programmer using Shen to make code which he can sell as code.  So that means that the jobbing coder who wants to sell a Shen based program cannot do so.  Number of people inconvenienced = n where n > k.

So if we measure the two licenses by the number of people prevented from doing something they actually have a need to do then the Shen license wins hands down.  In that sense it can be said to be freer than the GPL.   For that reason I'm not much worried about the GPL.  Nobody with sense would go there.

IMO Stallman has few redeeming qualities; but he does have one and that is he found a cause and stuck to it.  By repeating himself for 30 years, he illustrated the power of persuasion.   There is a lesson to be learnt there.  The Shen license is a good one; it simply needs to be worded so that the fear factor is removed.

Mark

[Jacob]

I know you said it was ONE path, I am just trying to get you to take the scenic route instead of the class 5 route(hiking reference).  People tend to like benevolent dictators(in the programming world at least) just look at python, clojure, etc.  I think if YOU, in particular, symbiotically attach Shen to a trendy platform it will gain some traction.  People want to see the leader pave the way, people seem more hesitant when it's a random contributor "supporting" a platform instead of working on it full time!  I also think that hoping someone will come along and spearhead a more popular route is not a good idea either.    

--Bottom Line--
Very few will care about SBCL being the platform(it is sad, but true); I want Shen to acquire much higher traffic/adoption.

[Mark Tarver]

It's a possibility, no more.  Some backer might change that.

The advantage is that people have been hanging about for months now waiting for fast concurrent Shen with graphics and this delivers it to them because I will do it quickly.  It means we can implement stuff, try out ideas and share apps while longer-term platforms mature.  When platform holders on other platforms arrive at this stage, they can copy the ideas in Shen/SBCL.  If these ideas are written in Shen, it should be even easier.

Mark

[h...@ancell-ent.com]

On Monday, 19 August 2013 08:16:11 UTC+1, Marko Kocić wrote:

Hi Mark,

I'm glad that Shen license is being reconsidered again

Errrm, I've proposed more of a redrafting/rearranging of the current
license and/or its intent, than something that could be reasonably
called "reconsidered."

since I agree that it is right now preventing Shen adoption in
some circles. I'm just not sure about the direction of change
regarding the new license. So far, main complain that I heard
about licensing is not that its content is wrong, but that it's
custom. Almost every developer, manager, company know about major

license types like GPL....

That's certainly an issue, but I don't know of any widely accepted
license that will address Dr. Tarver's fragmentation concern, although
I will examine Sun's additions to the GPL. But note those are
enforcing a trademark, not a copyright.

That means in any company, developer that wants to introduce Shen
will have to not only show its technical superiority, but also go
through legal department maze to get it approved, or it will be
automatically rejected since some companies don't have a
specialized legal department and just use common knowledge about
licensing.

Something that I'm going to look into is that we are addressing two
sets of people, those who simply want to use Shen to write their own
programs, and those who want to work on an implementation of it. I
may propose splitting the license into two, if the one for the former
can be made "so simple it's obviously correct".

E.g. as I understand it none of the "Don't break the spec", you're
checking out a library book concerns touch on the programmers simply
using Shen to get their job done.

Also, the custom licenses means that Shen will never become a part
of Debian or any other Linux distribution, which would also affect
adoption.

That's not always a good thing, Debian in particular is notorious for
having ancient versions in their official repositories, and after a
release they only fix security bugs. A lot of work is also required,
following each organization's particular rules of purity et. al. to
keep something in the official repositories. There are ways to
address this, and all these distros had to find some way to make peace
with Sun/Oracle's Java distribution so they'd be competitive.

One possible approach to mitigate this is to just use one of the
already existing opens source licenses, possibly with extensions. For
example GCC is GPL, with extension that states that binary product of
compilation doesn't fall into GPL although it contains portions of GPL
licensed C runtime. Similar approach has Java, which is also GPL,
meaning that anyone can fork it and update it, but cannot call it Java
unless it passes a full test suite. Something similar might also work
for Shen, but IANAL.

Here's the GCC Runtime Library Exception:

http://www.gnu.org/licenses/gcc-exception.html

After definitions it's pretty simple:

1. Grant of Additional Permission.

You have permission to propagate a work of Target Code formed by
combining the Runtime Library with Independent Modules, even if such
propagation would otherwise violate the terms of GPLv3, provided
that all Target Code was generated by Eligible Compilation
Processes. You may then convey such a combination under terms of
your choice, consistent with the licensing of the Independent
Modules.

2. No Weakening of GCC Copyleft.

The availability of this Exception does not imply any general
presumption that third-party software is unaffected by the copyleft
requirements of the license of GCC.

On the other hand, we can't just copy it, since the FSF copyrights all
their licenses:

Copyright © 2009 Free Software Foundation, Inc. <http://fsf.org/>

Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.

On Monday 19 Aug 2013 02:10:46 -0700 (PDT) Mark Tarver said:

Stallman is devious and socially disgusting, so GPL in any flavour
is out. There is no way I would want any association with the FSF.

Richard M. Stallman (RMS) and I were roommates when he started the
GNU project, after we'd both given up on the MIT/LMI Lisp Machine
effort. This was years prior to the formal GPL 1.0, but not its
progenitor licenses on e.g. GNU Emacs.

Which was written by an obscure guy named James Gosling and stolen by
RMS; contrary to e.g. the propaganda reflected in Wikipeida:

GNU Emacs was initially based on Gosling Emacs, but Stallman's
replacement of its Mocklisp interpreter with a true Lisp interpreter
required that nearly all of its code be rewritten.

Foundations like the C code for the buffergap editing buffers and
redisplay were fine, and the variable backup-by-copying-when-linked to
this day remains in the code; it is legally a derivative work.

I later worked on Emacs at UniPress, the commercial rightsholder, upon
which I became a "Software Hoarder", and a couple of my coworkers were
accused to their face of setting his apartment building on fire (it
was some kids with kerosene and matches). The owner/operators of the
company were old UNIX hands and wise enough to not get into a pissing
match with RMS and his disciples....

However this shortcut by him was *remarkably* unwise for a flagship
GNU product and one so intimately associated with him, he was the
major developer and maintainer of the original EMACS after Guy Steele
created it by unifying 3 original versions (it should be self-evident
that RMS didn't have the people skills or capital to get everyone to
agree on one set of key bindings). E.g. a major reason Linux beat BSD
was the 1992 lawsuit by AT&T's Unix System Laboratories.

We indeed want no association with the FSF whatsoever. That said, to
my knowledge no FOSS license has been more tested in court than the
GPL 2.0, but the most important thing is that a properly drafted FOSS
license is enforceable, see Jacobsen v. Katzer, or SCO v. the world.

[ The Linux adverse possession appropriation of OpenBSD drivers,
to which Stallman lent rhetorical help. ]

[ Some very good points about how an utter lack of discipline by
developers and distros strangled the Linux desktop market. ]

I myself am thinking very hard about buying a copy of Windows 7 before
it's too late, Debian Linux on the desktop has gotten significantly
worse between Lenny and Squeeze, e.g. in the kernel USB 1.x for sound
output was broken. Linux takes a GCC internals approach to drivers by
not having a stable kernel application binary interface (ABI), but in
turn they can't keep all of their drivers working, and GCC has become
well nigh unmaintainable.

In short, we care about quality, RMS/GNU/FSF care about their curious
definition of freedom, Linus/Linux uses the GPL in a competition
limiting way that damages quality.

So even if Windows 8 is crap - and it seems to be - Linux is not there
to pick up the ball because it still sucks and Lunduke has made an
annual lecture out of it sucking. And even if somebody does come up
with a good desktop Linux, the years of suckage will put people off
trying it and the philosophy of 'it has to be free' will suck the guts
out of the funding so it will too eventually suck. Ubuntu has been a
long term loss-maker for Shuttleworth and if wasn't for his deep
pockets, it would have long since died. He can blow $20M on a flight
into space, so for him its a hobby.

The other aspect is that ranting on about the evils of proprietary
software and saying 'we can't use Shen because it is not BSD/GPL' has
not helped Linux one bit. People want to run proprietary stuff on
their machine, to play games and so forth and making your OS unfriendly
and advertising your hostility does not pay. OS fanaticism killed off
that possibility. If Shen puts them off because it does not match
their standards of purity, I can safely say that those standards have
and will cost them dearly.

I think there's more of a balance here. However much the Linux
desktop sucks, many people including myself find it a better server
and software development platform than Windows (who needs more of a
GUI than EMACS? :-). If you aren't writing a GUI program much of the
above many not apply.

[ On the delicacy of dealing with corporations. ]

On Monday, 19 Aug 2013 06:04:13 -0700 (PDT) Mark Tarver said:

I said it is rational to avoid placing the Shen project under the
auspices of a license that is written by an organisation using sharp

practice....

The FSF took possession of your favorite Common Lisp implementation; as
Wikipedia puts it (http://en.wikipedia.org/wiki/CLISP):

Haible did not originally intend to distribute CLISP under the GPL,
but in a well-publicised email exchange with Richard Stallman, he
eventually agreed to do so. The issue at stake was whether CLISP was
a derivative work of the GNU readline library.[1]

Back to you:

But let's put the nasty goings-on around the FSF aside and just put the
GPL license next to the Shen license. What does the Shen license
forbid - essentially tampering with the kernel code and issuing a

faulty distro. Who is put out by that? [...]

Well, there are people who have their own ideas of what a future Lisp
should be, or have a different focus than you, I want to make SMP/
ccNUMA systems sing, you like the Actors model. But I'll first run
any proposed "improvements" in my implementation by the list to see if
there's a pure Shen way of doing them, and of course any changes will
be compile or runtime options you will have to ask for with a flag or
the like.

The GPL on the other hand prevents any programmer using Shen to make

code which he can sell as code....

This is why the the GCC Runtime Library Exception exists, however much
it frosts RMS that it's used for so much proprietary stuff.

[...] In that sense it can be said to be freer than the GPL.

That's a really low floor ^_^. Only the GNU Affero General Public
License is worse, it "is a free, copyleft license for software and
other kinds of works, specifically designed to ensure cooperation with
the community in the case of network server software." I.e. it
attempts to be viral if you use it to provide an Internet service.

It's being used by some significant systems:

http://en.wikipedia.org/wiki/List_of_AGPL_web_applications

Then again MIT et. al. don't care if others try to create their own
versions of EdX, the really significant value is the material which is
cloned from courses they've previously developed. At no small cost,
e.g. 6.001/SICP cost 19 million USD in 2013 dollars.

For that reason I'm not much worried about the GPL. Nobody with
sense would go there.

It's noteworthy that it's not popular for Lisp implementations.

IMO Stallman has few redeeming qualities; but he does have one and that
is he found a cause and stuck to it. By repeating himself for 30
years, he illustrated the power of persuasion. There is a lesson to
be learnt there. The Shen license is a good one; it simply needs to be
worded so that the fear factor is removed.

If we can recast it into a form people are familiar with and have
already accepted, i.e. like the GCC Runtime Library Exception and the
Sun Java trademark license (modified to preclude an Apache Harmony
problem by making the Shen version of the "Technology Compatibility
Kit" free), we'll be in lots better shape.

- Harold

[Mark Tarver]

Richard M. Stallman (RMS) and I were roommates when he started the
GNU project, after we'd both given up on the MIT/LMI Lisp Machine
effort.  This was years prior to the formal GPL 1.0, but not its
progenitor licenses on e.g. GNU Emacs.

Which was written by an obscure guy named James Gosling and stolen by
RMS; contrary to e.g. the propaganda reflected in Wikipeida:

Do you mean that the GPL was stolen or Emacs?  It sounds like you mean the former.  I know about what happened with Emacs. For all Shenturions who have an interest in Emacs and the history of Lisp have a read of Dan Weinreb.   He used to have a blog, but its gone and so has the original page.  He died of cancer last year.  However I've resurrected his words through the Internet Archive and you can read them.

http://web.archive.org/web/20090101103828/http://danweinreb.org/blog/rebuttal-to-stallmans-story-about-the-formation-of-symbolics-and-lmi

 Mark

[Greg Baryza]

On Sunday, August 18, 2013 11:16:30 AM UTC-4, Harold Ancell wrote:

 From: Mark Tarver <dr.mt...@gmail.com>
  Date: Sat, 17 Aug 2013 23:51:51 -0700 (PDT)

  I've discussed this idea with Greg and he agrees; I would put
  together an exhaustive set of standard licensing tests for any Shen
  port which would provide a computable test for passing something as
  Shen.  This would remove any lingering fear of me exercising some
  arbitrary droit de seigneur over other people's work.  If it passes
  the standard tests, it is legal and that is the end of the matter.
  This should satisfy all but the religious.

 

I believe this will help, but in the same way that Shen itself has versions, so will the test suite. And there is nothing you can do to make the paranoids believe you won't invalidate their work with a newer test suite.

 

That said, this will eliminate lots of trivial disputes; or perhaps just move them to a discussion of whether the tests are "fair" (in some sense of the word).

[Mark Tarver]

I want to make clear that the license is not being reconsidered; its being redrafted to make it clearer and fixed to an objective and computable standard free of the suspicion of tampering.   For all the reasons stated, there is no way we would go to GPL.   I wanted people to know why I feel that way and so I put up these links.  Stallman's history is catching up on him.  You can't pull this bullshit in the age of the Internet and get away with it.  'Information wants to be free' applies without exception, and you have to be straight with people.  His actions have simply given reason for people to avoid the GPL in all its manifestations and will ultimately 

hurt the FSF.  However this is a headache for people like Lefty Schlesinger who front the OS movement (http://opensourcetogo.blogspot.co.uk/2009/07/emailing-richard-stallman.html) who have to deal with him.  It should not be a headache for us and I won't let it be.

For the record I do not believe in The One True License.  I believe in project-driven licensing; meaning that you look at what the project is supposed to achieve and you choose the license to fit.  In the case of Shen, this is such an unusual project, both in its goals and methodology, that some license was needed that reflected both the need to freely share source and the need to maintain strong standards for the kernel.  I could not find that license and so the Shen license was written.

I spoke in the license of the need to avoid forking and fragmentation and how this killed desktop Linux and de Icaza echoed the latter view.   But there is another aspect which is touched on in the video.  I said that the Ring could provide new levels of software integrity for safety-critical applications.  That is true.  However all the math'l logic and assurance of type security become voided if the basic kernel is corrupt.   

Regarding BSD, I actually believe for this application the Shen license is superior.  From the point of view of a corporation or any end user, a platform that contains the assurance of great reliability, but gives them freedom to produce proprietary code and borrow kernel code to do it, is actually more valuable than a pile of BSD code which carries no assurance whatsoever and which could cost reputation, money or worse lives if you used it in a serious app.

Hence though I read Slashdot and Reddit, I don't take everything on board and I don't take at face value the karma points.    'Microsoft is evil' will get you modded up on Slashdot even though it is silly.   The main problem here is removing the fear factor from the license and the suggestions made here will do that.

I've said that desktop Linux is nowhere now.  Maybe if they'd adopted a Shen style licensing approach and spliced the open source methodology with the discipline enforced by having to meet a legal standard for their work, I would not be walking into PC World with a choice of buying either Windows 8 or an overpriced Mac.   The Shen project is actually one of the most successful open source projects around and has done far better than desktop Linux in relation to its resources.  We have something important to teach here and we should not be bullied.

Mark

--
You received this message because you are subscribed to the Google Groups "Shen" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qilang+un...@googlegroups.com.
To post to this group, send an email to qil...@googlegroups.com.
Visit this group at http://groups.google.com/group/qilang.
For more options, visit https://groups.google.com/groups/opt_out.

[h...@ancell-ent.com]

From: Greg Baryza <bar...@intersystems.com>
Date: Tue, 20 Aug 2013 07:32:12 -0700 (PDT)

I believe [an exhaustive set of standard licensing tests] will help,

but in the same way that Shen itself has versions, so will the test
suite. And there is nothing you can do to make the paranoids believe
you won't invalidate their work with a newer test suite.

I haven't come across anyone who's "paranoid", e.g. doubting the good
will of Team Shen. The doubts have been over an inherently ambiguous
and very hard to analyze license (which we propose to fix by changing
the form), a custom license (which we can't fix, but can analogize to
two license additions that were widely accepted), and the potential for
subjective judgements of "don't break the spec" (which this will
perforce fix).

That said, this will eliminate lots of trivial disputes; or perhaps
just move them to a discussion of whether the tests are "fair" (in
some sense of the word).

Indeed, and it'll be hard to create one that doesn't "break the spec".

Which gets to the issue of motivations. Team Shen says this is all to
achieve Java(TM)'s "write once, run anywhere" (US Registered Trademark
#2807022) and I see no reason to doubt them. Sun's bad faith towards
the Apache Harmony project had obvious commercial motivations, ones I
believe we saw played out in Oracle's lawsuit against Google's Android
project (Sun and Oracle wanted you to use their limited, frozen in time,
not at all free Java implementation for mobile phones instead of rolling
your own).

By then Java was wildly popular (a safe and less twisted C++). Team
Shen want the widest possible adaptation of Shen and that's not served
by playing stupid games with the community, and for the foreseeable
future, at the first real hint of that, Shen will almost certainly die.
Moving to "an exhaustive set of standard licensing tests" instead of
subject judgement will in fact help avoid the appearance of that.

- Harold

[h...@ancell-ent.com]

Note, if you're not interested in old Lisp or EMACS history or the
what's about the finish of this list's Hate Week towards Richard M.
Stallman, you can probably skip this (I'm in part writing it to get some
oral history on record). I'm drafting a following message from one of
the comments that's more on topic about the limitations of "Free"
software as that has all played out.

From: Mark Tarver <dr.mt...@gmail.com>

Date: Tue, 20 Aug 2013 02:57:58 -0700 (PDT)

Richard M. Stallman (RMS) and I were roommates when he started the
GNU project, after we'd both given up on the MIT/LMI Lisp Machine
effort. This was years prior to the formal GPL 1.0, but not its
progenitor licenses on e.g. GNU Emacs.

Which was written by an obscure guy named James Gosling and stolen
by RMS; contrary to e.g. the propaganda reflected in Wikipeida:

Do you mean that the GPL was stolen or Emacs? It sounds like you mean
the former.

The latter. GNU Emacs started from a version of (James) Gosling Emacs,
which was the first UNIX(TM) non-ersatz EMACS, in large part because it
was about the first written for the VAX instead of a split I/D PDP-11
with 64KB code, 56KB data and an 8KB stack.

I know about what happened with Emacs. For all Shenturions who have an
interest in Emacs and the history of Lisp have a read of Dan Weinreb.
He used to have a blog, but its gone and so has the original page. He
died of cancer last year. However I've resurrected his words through
the Internet Archive and you can read them.

http://web.archive.org/web/20090101103828/http://danweinreb.org/blog/rebuttal-to-stallmans-story-about-the-formation-of-symbolics-and-lmi

I can add to this history, I showed up on the scene when the wire
wrapping for CADR #9 was being tested, (by another Lispm of course). At
that point Dan had decamped for California to work on software for the
LLNL S-1 project (ECL supercomputer for the USNavy with some ambitious
software goals), and there was a telling story about that:

The word was that he and RMS had a technical disagreement which RMS
resolved by deleting all the alternative sources, including from
backup tapes.

Now, Dan later denied this, and thinking about it a quarter century
later I suspect for the AI Lab this would be a transgression too far,
e.g. sometime later someone accused of messing with backup tapes was
declared persona non grata.

But it's very telling that *no one* expressed any doubt RMS was capable
of this, and I can assure you his fanaticism and bad manners did not
start with the GNU project (the FSF and GPL came later).

A while later I went to work for LMI. I was recruited because I was the
only person in the community with a foot in both the Lispm and UNIX
camps, was hired in part because I had a record of getting things done,
and I timed out on them 8-9 months later because I ran out of things I
could do, besides learning how to do digital circuit design.

And therefore can confirm a number of things Dan said, directly
contradicting RMS. E.g. I was one of the last people in the community
willing to share a meal with him, i has been said the gravamen of RMS's
dispute was that all the people he used to have lunch or dinner with
decamped to Symbolics.

Echoing your link to another blog posting about RMS's issues with the
opposite sex, having dinner with him could be dangerous, a very good
friend mentioned a very late night/early morning dinner at Moon Villa,
the good Cantonese place in Chinatown we'd go to after everything else
but IHOP was closed, where RMS hit on a gangster's moll (girlfriend). My
friend was frightened that RMS would get all of them killed. I can
personally attest that he would hit on any woman he found attractive who
came "into range", no matter how attached they were.

So, from Dan's posting:

"[...] Greenblatt himself would be the CEO. The other members of the
Lisp machine project were extremely dubious of Greenblatt’s ability to
run a company."

And that was a very accurate perception. No other experienced hardware
designer was willing to work with him until I recruited a graduating
classmate who for example build a baby Lisp Machine in his digital
design lab. Said friend then put in many 100 hours weeks (sic) and made
LMI's new processor, an evolution of the CADR, work. LMI eventually
went bankrupt, was bought by its Canadian distributor, and that was
killed by internal Canadian politics (really; the company was shut down
by law enforcement agents who expected to find a shell company with
perhaps a front desk secretary, not a bunch of people building machines
and shipping them out the door...).

In the comments David Chapman mentions LMI bootstrapping through sales;
well, there was some family and the like seed investment, and then the
company ran out of money. For a couple of months we were told "We've
cut your paychecks but they're in the safe; if we get the investment
from Texas Instruments (TI) we'll pay you. Working till then is of
course entirely optional."

Stallman’s characterization of Symbolics as “looking for ways to
destroy” LMI is pure fantasy.

Indeed. LMI was never a serious threat, and certainly not perceived as
one during this critical period when Greenblatt was the only experienced
guy working on the project. In microcode as well as hardware, for that
matter; new microcode was required to e.g. take advantage of the
machine's added 16 bit TRW multiplier chip (these were the bad old days
when we used the fastest family of TTL available, some years before you
could fit such a system on a chip you could afford to design and get
manufactured, something TI in conjunction with LMI started doing then).
I in fact became the LMI Head of Intelligence ^_^ and carefully traded
information with my friends who worked for Symbolics so we each had a
limited window into what the other was doing.

For emptying out the AI Lab of hackers, I can't quite say, but see my
above about eating out with other people. Note the first comment, by
Dave Moon (microcode and many other things wizard), "But I think some
other people left the AI lab for Symbolics later. Was that Symbolics
pull or Stallman push?"

To finish, Symbolics salesmen were pretty notorious, the company was
utterly unethical with Macsyma (one of the things to prompt Danny
"Connection Machine" Hillis to ask me if it was vaguely possible for LMI
to supply Thinking Machine's Lisp Machines), but the technical staff was
nothing but ethical (of both companies, I'd say), and RMS's morality
tale of the "end of hacking" at MIT/the AI Lab etc. is way beyond self-
serving. And telling, that he's willing to slander a lot of very good
people to promote his vision of "free" computing. Which is partly bunk
as I will develop in my following message.

- Harold

[Mark Tarver]

On Tuesday, August 20, 2013 3:32:12 PM UTC+1, Greg Baryza wrote:

On Sunday, August 18, 2013 11:16:30 AM UTC-4, Harold Ancell wrote:

 From: Mark Tarver <dr.mt...@gmail.com>
  Date: Sat, 17 Aug 2013 23:51:51 -0700 (PDT)

  I've discussed this idea with Greg and he agrees; I would put
  together an exhaustive set of standard licensing tests for any Shen
  port which would provide a computable test for passing something as
  Shen.  This would remove any lingering fear of me exercising some
  arbitrary droit de seigneur over other people's work.  If it passes
  the standard tests, it is legal and that is the end of the matter.
  This should satisfy all but the religious.

 

I believe this will help, but in the same way that Shen itself has versions, so will the test suite. And there is nothing you can do to make the paranoids believe you won't invalidate their work with a newer test suite.

I've thought of that and it can be dealt with.    . However this is not the place to discuss the rewording of a license.   The wording of the license was one of a list of items in the OP.

 

That said, this will eliminate lots of trivial disputes; or perhaps just move them to a discussion of whether the tests are "fair" (in some sense of the word).

These people  - paranoids - are not our concern.  You will never eliminate them or gossip based on it.  Resentment, paranoia and conspiracy mentality are all over the internet.  The OS community is riven by suspicion and factionalism as my links show.   I'm here to deal with the reasonable fear, not the 'lying awake at 3.00am' stuff.

Mark

 

[Mark Tarver]

OK; I propose to move the detailed discussion to email at this point and endeavour to hammer out a form of wording acceptable to both of us and then circulate it through the 2011 committee.

Mark