From: ········@acm.org
Subject: LispOS, LispVM projects?
Date: 
Message-ID: <71vfvh$gd0$1@nnrp1.dejanews.com>
What happened?

--
Fernando D. Mato Mira
Real-Time SW Eng & Networking
CSEM
CH-2007 Neuchatel
Switzerland

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    

From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <71vl0v$gpt$3@spitting-spider.aracnet.com>
In article <············@nnrp1.dejanews.com>,
	········@acm.org writes:
> What happened?

  All talk, no work.

  Mike McDonald
  ·······@mikemac.com
From: ···@rebol.com
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <m2lnlojot0.fsf@ivantheterrible.rebol.net>
·······@mikemac.com (Mike McDonald) writes:

> In article <············@nnrp1.dejanews.com>,
> 	········@acm.org writes:
> > What happened?
> 
>   All talk, no work.
> 

Not true!  John Morrison at MAK put a lot of work into a kernel
that was designed to run a VM that could support lisp.

What really happened is that the project petered out at the
beginning of the semester.

~jrm
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <7207on$9pr$1@spitting-spider.aracnet.com>
In article <··············@ivantheterrible.rebol.net>,
	···@rebol.com writes:
> ·······@mikemac.com (Mike McDonald) writes:
> 
>> In article <············@nnrp1.dejanews.com>,
>> 	········@acm.org writes:
>> > What happened?
>> 
>>   All talk, no work.
>> 
> 
> Not true!  John Morrison at MAK put a lot of work into a kernel
> that was designed to run a VM that could support lisp.

  John Morrison is working on JJOS, a Java OS project. (Unfortunately,
etherboot doesn't like any of my ethernet cards (even the NE2000 I bought
specificly for it!) so I haven't been able to play with it much.)

> What really happened is that the project petered out at the
> beginning of the semester.
> 
> ~jrm

  LispOS petered out long before this semester started.

  Mike McDonald
  ·······@mikemac.com
From: ···@rebol.com
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <m2k918jn0y.fsf@ivantheterrible.rebol.net>
·······@mikemac.com (Mike McDonald) writes:

> In article <··············@ivantheterrible.rebol.net>,
> 	···@rebol.com writes:
> > ·······@mikemac.com (Mike McDonald) writes:
> > 
> >> In article <············@nnrp1.dejanews.com>,
> >> 	········@acm.org writes:
> >> > What happened?
> >> 
> >>   All talk, no work.
> >> 
> > 
> > Not true!  John Morrison at MAK put a lot of work into a kernel
> > that was designed to run a VM that could support lisp.
> 
>   John Morrison is working on JJOS, a Java OS project. (Unfortunately,
> etherboot doesn't like any of my ethernet cards (even the NE2000 I bought
> specificly for it!) so I haven't been able to play with it much.)

Yes, but he believes this offers a good oportunity to slip lisp in via
the back door.  To quote him, `What I really want is a LispM'.

> > What really happened is that the project petered out at the
> > beginning of the semester.
> > 
> > ~jrm
> 
>   LispOS petered out long before this semester started.

The traffic surrounding it dropped dramatically in September of
1997, if I remember correctly.
From: John Morrison
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <364c68d7.395292330@news.tiac.net>
Hi,

Geez, so *this* is where all you guys hang out.  Shoulda known.

···@rebol.com wrote:

>·······@mikemac.com (Mike McDonald) writes:
>
>> In article <··············@ivantheterrible.rebol.net>,
>> 	···@rebol.com writes:
>> > ·······@mikemac.com (Mike McDonald) writes:
>> > 
>> >> In article <············@nnrp1.dejanews.com>,
>> >> 	········@acm.org writes:
>> >> > What happened?
>> >> 
>> >>   All talk, no work.
>> >> 
>> > 
>> > Not true!  John Morrison at MAK put a lot of work into a kernel
>> > that was designed to run a VM that could support lisp.
>> 
>>   John Morrison is working on JJOS, a Java OS project. (Unfortunately,
>> etherboot doesn't like any of my ethernet cards (even the NE2000 I bought
>> specificly for it!) so I haven't been able to play with it much.)
>
>Yes, but he believes this offers a good oportunity to slip lisp in via
>the back door.  To quote him, `What I really want is a LispM'.

You betcha.  I'm thinking, "Trojan Horse."

>
>> > What really happened is that the project petered out at the
>> > beginning of the semester.

Well, er, somewhat embarrassingly, I'm kind of *OLD* for school.
Actually, however, the JVM *is* being written by a student.  I'm just
writing the native code which supports it.  However, I expect I'll
pitch in to help with the JVM's memory management and bootstrapping
code, and probably also with the first drivers (VGA, keyboard, and IDE
disk).

The reason I haven't released another snapshot of the sources is that
a Day Job Crisis reared its head, and it took about 2 weeks to kill
it.  I hope to do more this weekend.

>> > 
>> > ~jrm
>> 
>>   LispOS petered out long before this semester started.
>
>The traffic surrounding it dropped dramatically in September of
>1997, if I remember correctly.

I am somewhat surprised (shocked, really), that so few people have
volunteered (from either the JOS or LispOS communities) to pitch in.
If every line of every email and every web site were code, we'd be
(more or less) done by now.  I am hoping that it's nothing I've/we've
said or written.  I am hoping that will change once we start managing
the sources via networked CVS instead of monolithic tar/zip files.  I
have called for volunteers on both the jos-kernel mailing list (the
archives for which have not been updated in months) and the LispOS
mailing list.  If I appealed any harder for volunteers, my dignity
would suffer (and it needs all the help it can get in the best of
circumstances).

Actually, most of the people who responded at all and have been kind
enough to offer advice seem to be participating in this particular
USENET thread.

I hope somebody manages to pull this off even if I don't have s**t to
do with it.  I'm getting kind of worried -- most of our newer
programmers seem to think that the only programming environment that
ever existed is the Visual C++ IDE, and a distinct and very vocal
minority are of the Linux Rules School.  When discussions of the
relative merits of Windows and UNIX are discussed, their eyes kind of
glaze over when LispMs are mentioned -- do you know how hard it is to
explain Dynamic Windows to either a Microsoft or Linux fanatic?  They
don't even know what I'm talking about, and have no idea that LispMs
ever existed.

Oh well.  On to more mundane considerations:  I'm trying to score a
working LispM with documentation, because I figure this will be much
more efficient (in terms of my effort).  Maybe you guys who are going
to the upcoming user's group conference (I'm not going) can prevail
upon SOMEBODY to OK the release of either a band (or something) to go
along with the emulator.  Whatever we can get permission to release is
what I should probably be working on.

(If you want to reply to me, please remove the "nospam" from my email
address in the header.  Damned spammers.)

-jm


==== John Morrison ==== ··@mak.com == http://www.mak.com/welcome.html
==== MaK Technologies Inc., 185 Alewife Brook Parkway, Cambridge, MA 02138
==== vox:617-876-8085 x115
==== fax:617-876-9208
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72i153$3cl$1@spitting-spider.aracnet.com>
In article <··················@news.tiac.net>,
	··@nospam.mak.com (John Morrison) writes:

>>> > What really happened is that the project petered out at the
>>> > beginning of the semester.
> 
> Well, er, somewhat embarrassingly, I'm kind of *OLD* for school.
> Actually, however, the JVM *is* being written by a student.  I'm just
> writing the native code which supports it.  However, I expect I'll
> pitch in to help with the JVM's memory management and bootstrapping
> code, and probably also with the first drivers (VGA, keyboard, and IDE
> disk).
> 
> The reason I haven't released another snapshot of the sources is that
> a Day Job Crisis reared its head, and it took about 2 weeks to kill
> it.  I hope to do more this weekend.

  I'm looking forward to the patch! :-) (John's code doesn't like my graphics
card for some reason.)

> Actually, most of the people who responded at all and have been kind
> enough to offer advice seem to be participating in this particular
> USENET thread.

  I think there's actually only three people on the LispOS mailing list. The
rest are mailbots!

> I hope somebody manages to pull this off even if I don't have s**t to
> do with it.  I'm getting kind of worried -- most of our newer
> programmers seem to think that the only programming environment that
> ever existed is the Visual C++ IDE, and a distinct and very vocal
> minority are of the Linux Rules School.  When discussions of the
> relative merits of Windows and UNIX are discussed, their eyes kind of
> glaze over when LispMs are mentioned -- do you know how hard it is to
> explain Dynamic Windows to either a Microsoft or Linux fanatic?  They
> don't even know what I'm talking about, and have no idea that LispMs
> ever existed.

  I'm glad I'm not the only one having this nightmare!

> Oh well.  On to more mundane considerations:  I'm trying to score a
> working LispM with documentation, because I figure this will be much
> more efficient (in terms of my effort).  

  This might be a bit misleading to others. From our private discussions, what
John is really looking for isn't a working machine so much as a working world
load with doecs and source code so that an emulator can be written to run the
world load in. Any of the of MIT CADR, TI Explorer, LMIs, ... would be ideal.
(MIT has agreed in principle to the CADR code but they can't find a copy of
the sources. TI also seems to have misplaced theirs, although they haven't
said whether they'd be willing to release it anyway. LMI has completely
disappeared. Only Symbolics is still around and they're determined to take
their code to the grave with them. Which is a shame even if it's their right
to do so.) So, if anyone has or know someone who has the sources to any of the
older LispMs, please, PLEASE, speek up!

> Maybe you guys who are going
> to the upcoming user's group conference (I'm not going) can prevail
> upon SOMEBODY to OK the release of either a band (or something) to go
> along with the emulator.  Whatever we can get permission to release is
> what I should probably be working on.



  Mike McDonald
  ·······@mikemac.com
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1311982054270001@194.163.195.67>
In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
wrote:

> disappeared. Only Symbolics is still around and they're determined to take
> their code to the grave with them.

Symbolics is still alive and I wouldn't dig a grave.
Open Genera 2.0 is still in the works.

-- 
http://www.lavielle.com/~joswig
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72ie70$gp6$1@spitting-spider.aracnet.com>
In article <·······················@194.163.195.67>,
	······@lavielle.com (Rainer Joswig) writes:
> In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
> wrote:
> 
>> disappeared. Only Symbolics is still around and they're determined to take
>> their code to the grave with them.
> 
> Symbolics is still alive and I wouldn't dig a grave.

  I think they are terminally ill. Granted, it takes a long time for a company
to die (unless they do something stupid to accelerate it!), so we have some
time to dig the grave. I wish it wasn't so.

> Open Genera 2.0 is still in the works.

  Another funny! It's only a little over a year late and counting.

  Mike McDonald
  ·······@mikemac.com
From: Lyman S. Taylor
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72ig8n$60c@pravda.cc.gatech.edu>
In article <············@spitting-spider.aracnet.com>,
Mike McDonald <·······@mikemac.com> wrote:
>In article <·······················@194.163.195.67>,
>	······@lavielle.com (Rainer Joswig) writes:
>> In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
>> wrote:

>> Symbolics is still alive and I wouldn't dig a grave.
>
>  I think they are terminally ill. 

   I didn't save the message from info.slug, but I thought they were in 
   yet another "Chap. 11" phase.   Or was it the slug gateway that
   was dying?   Or was that just the non open Genera stuff? 

>> Open Genera 2.0 is still in the works.

   Lisp machine layered on Unix.  Hmmm, I wonder if it could run on 
   Alpha Linux.  A "demo version" on a relatively inexpensive alpha linux 
   box might make so that a least some of those who have only every seen 
   the Visual C++ IDE would could see something different. :-) 

  
-- 
					
Lyman S. Taylor                "Twinkie Cream; food of the Gods" 
(·····@cc.gatech.edu)                     Jarod, "The Pretender" 
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72imk7$osp$2@spitting-spider.aracnet.com>
  This is all based upon my understanding. Take it all with a grain of salt.

In article <··········@pravda.cc.gatech.edu>,
	·····@cc.gatech.edu (Lyman S. Taylor) writes:
> In article <············@spitting-spider.aracnet.com>,
> Mike McDonald <·······@mikemac.com> wrote:
>>In article <·······················@194.163.195.67>,
>>	······@lavielle.com (Rainer Joswig) writes:
>>> In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
>>> wrote:
> 
>>> Symbolics is still alive and I wouldn't dig a grave.
>>
>>  I think they are terminally ill. 
> 
>    I didn't save the message from info.slug, but I thought they were in 
>    yet another "Chap. 11" phase.   Or was it the slug gateway that
>    was dying?   Or was that just the non open Genera stuff? 

  Actually, they're no longer in Chap. 11. They were liquidated last Feb. (??)
Two guys who's names I can never remember bought the remains and are pushing
forward with 8.5 for Alphas and XLs.

>>> Open Genera 2.0 is still in the works.
> 
>    Lisp machine layered on Unix.  Hmmm, I wonder if it could run on 
>    Alpha Linux.  A "demo version" on a relatively inexpensive alpha linux 
>    box might make so that a least some of those who have only every seen 
>    the Visual C++ IDE would could see something different. :-) 

  They were going to look into Alpha Linux when they got a chance. The box
maybe cheap but Open Genera 8.3 was expensive.

  Mike McDonald
  ·······@mikemac.com
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1411980834160001@194.163.195.67>
In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
wrote:

>   Actually, they're no longer in Chap. 11. They were liquidated last Feb. (??)
> Two guys who's names I can never remember bought the remains and are pushing
> forward with 8.5 for Alphas and XLs.
> 
> >>> Open Genera 2.0 is still in the works.
> > 
> >    Lisp machine layered on Unix.  Hmmm, I wonder if it could run on 
> >    Alpha Linux.  A "demo version" on a relatively inexpensive alpha linux 
> >    box might make so that a least some of those who have only every seen 
> >    the Visual C++ IDE would could see something different. :-) 
> 
>   They were going to look into Alpha Linux when they got a chance. The box
> maybe cheap but Open Genera 8.3 was expensive.

It runs under DIGITAL Unix. It would be nice to get rid of
DIGITAL Unix either by replacing it with Linux or
(even better ;-) ) with Open Genera directly.
Unfortunately SPARCs are more popular here.

-- 
http://www.lavielle.com/~joswig
From: Christopher Stacy
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <ug1bjayg2.fsf@pilgrim.com>
I think you have to have a 64 bit machine, or it's hopeless.
From: David Gadbois
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <w0k4sruc6jx.fsf@lagavulin.cyc.com>
Christopher Stacy <······@pilgrim.com> writes:
> I think you have to have a 64 bit machine, or it's hopeless.

People keep making this point, but I just don't understand it.  Last I
heard, the Ivory emulator did some tricks like holding the 8-bit tags
along with the 32-bit pointer value in a single 64-bit register, but
what does that really buy?  Is there enough state that register
pressure is an issue?  Does that alone give more than, say, a factor
of two in performance?

--David Gadbois
From: Lyman S. Taylor
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <737ohc$a31@pravda.cc.gatech.edu>
In article <···············@lagavulin.cyc.com>,
David Gadbois  <·······@lagavulin.cyc.com> wrote:
>Christopher Stacy <······@pilgrim.com> writes:
>> I think you have to have a 64 bit machine, or it's hopeless.
>
>People keep making this point, but I just don't understand it.  Last I
>heard, the Ivory emulator did some tricks like holding the 8-bit tags
>along with the 32-bit pointer value in a single 64-bit register, but
>what does that really buy? 

   In short,  speed. 

   [ On a 32 bit machine it  isn't "hopeless" , just  "slow as molasses" ] 


   Given that you are trying to emulate a machine with registers larger
   than 32 bits.  It means not haveing to use 2 registers to "model"  1. 

   For one this aviods having to do 2 loads get the data and the tag. 
   As opposed to one load to get both. 

   two registers                                one register

   load data                                    load data/tag
   load tag                                     mask tag into another register

   That mask isn't going to have to reach out into the memory hierarchy 
   to get the data.  Load once, use often. :-) 


> Is there enough state that register
>pressure is an issue? 

    Your running an emulator.  Register pressure is an issue from the
    simply just because of that.   You need registers for the emulator
    and you need registers for the code being executed. 

    Secondly, the Symbolics archictecture had a stack cache which also needs
    to be emulated.  Preferable by registers.  So that is consumer also. 

    I'd be very surprised with there were any underutilized registers
    by the Symbolics emulator.  It was written in hand tuned assembler.

    I'm not sure how they delt with the issues of the symbolics being
    Word addresses versus the byte address of the Alpha. [ that part of 
    one paper isn't on the web. ]   I imagine this may be a register 
    consumer too. 
   
> Does that alone give more than, say, a factor
>of two in performance?

    Pushing losts of tuff out to the high levels of the memory hierarchy 
    typically has much more than a factor of two performance on high
    clock rate processors like the Alpha. 

-- 
					
Lyman S. Taylor           "Computers are too reliable to replace
(·····@cc.gatech.edu)     	 humans effectively."
			        Commander Nathan Spring, "Starcops"
From: Tim Bradshaw
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <ey3yap23949.fsf@todday.aiai.ed.ac.uk>
* Lyman S Taylor wrote:

>    For one this aviods having to do 2 loads get the data and the tag. 
>    As opposed to one load to get both. 

>    two registers                                one register

>    load data                                    load data/tag
>    load tag                                     mask tag into another register

>    That mask isn't going to have to reach out into the memory hierarchy 
>    to get the data.  Load once, use often. :-) 

Actually, I think this may not be as much of an issue as all that.  If
the data & tag are next to each other in address space, then those two
loads probably come down to a single memory fetch and then a cache
hit.  If the cache hit is single-cycle then this may actually execute
as fast as the other case, perhaps depending on pipeline issues.

Of course this assumes a 64-bit+ path to main memory, but I think
everyone has that now (?).

Of course it's still a pain because it's a completely different
implementation, and because of register pressure.

--tim
From: David Gadbois
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <w0k3e7acncb.fsf@lagavulin.cyc.com>
·····@cc.gatech.edu (Lyman S. Taylor) writes:
> David Gadbois  <·······@lagavulin.cyc.com> wrote:
> >
> >Last I heard, the Ivory emulator did some tricks like holding the
> >8-bit tags along with the 32-bit pointer value in a single 64-bit
> >register, but what does that really buy?
> 
> In short,  speed. 
> 
> [On a 32 bit machine it isn't "hopeless" , just "slow as molasses"]

Well, Open Genera 1.0 on a 190MHz 21064 was perfectly usable for
development.  More recent 32-bit machines are roughly 4-8 times faster
than that old part, and, more importantly, have bigger caches with
more associativity.  So I would expect a 32-bit version of the
emulator to be even more usable, even with a factor of two hit for a
32-bit emulator.

I still don't understand why folks are obsessing over some notion of
absolute performance of the emulator.  I mean, jeez, if performance is
an issue, get rid of the emulation first.

> Given that you are trying to emulate a machine with registers larger
> than 32 bits.  It means not haveing to use 2 registers to "model" 1.
> 
> For one this aviods having to do 2 loads get the data and the tag.
> As opposed to one load to get both.

I gather that the current emulator uses two loads on the Alpha anyway:
The tag bits are stored separately from the contents.  The emulation
of the SYS:MEMORY-READ instruction would go something like:

   Mask the tag out of the register holding the argument address
   Load the contents of the address
   Compute the address of the tag
   Load the tag
   Mask in the tag to the contents

The rational was not wasting 24 bits for every (emulated) word of
memory and to allow for easier (unboxed) communication of values to
the run-time system.

--David Gadbois
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <73cerm$48b$2@spitting-spider.aracnet.com>
In article <···············@lagavulin.cyc.com>,
	David Gadbois <·······@lagavulin.cyc.com> writes:
> ·····@cc.gatech.edu (Lyman S. Taylor) writes:
>> David Gadbois  <·······@lagavulin.cyc.com> wrote:
>> >
>> >Last I heard, the Ivory emulator did some tricks like holding the
>> >8-bit tags along with the 32-bit pointer value in a single 64-bit
>> >register, but what does that really buy?
>> 
>> In short,  speed. 
>> 
>> [On a 32 bit machine it isn't "hopeless" , just "slow as molasses"]
> 
> Well, Open Genera 1.0 on a 190MHz 21064 was perfectly usable for
> development.  More recent 32-bit machines are roughly 4-8 times faster
> than that old part, and, more importantly, have bigger caches with
> more associativity.  So I would expect a 32-bit version of the
> emulator to be even more usable, even with a factor of two hit for a
> 32-bit emulator.

  And I have to believe it'd beat the pants off of a 36X0 series.

> I still don't understand why folks are obsessing over some notion of
> absolute performance of the emulator.  I mean, jeez, if performance is
> an issue, get rid of the emulation first.

  Amen!!!! I don't use Genera for speed. (I'm still happy with the preformance
of a 3620 or XL1201. If a 32bit version of OpenGenera could give me that level
of performance, I'd be more than happy!) I use it for all of the functionality
it brings with it.

  Mike McDonald
  ·······@mikemac.com
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1411980050250001@194.163.195.67>
In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
wrote:

> >> disappeared. Only Symbolics is still around and they're determined to take
> >> their code to the grave with them.
> > 
> > Symbolics is still alive and I wouldn't dig a grave.
> 
>   I think they are terminally ill. Granted, it takes a long time for a company
> to die (unless they do something stupid to accelerate it!), so we have some
> time to dig the grave. I wish it wasn't so.

They are still there and that is enough for me.

> > Open Genera 2.0 is still in the works.
> 
>   Another funny! It's only a little over a year late and counting.

Progress has been slow but its really getting nearer to being
released. I have seen it running and it's not that bad.

Actually I'm starting to think about DEC Alpha prices. ;-)

-- 
http://www.lavielle.com/~joswig
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72imo9$osp$3@spitting-spider.aracnet.com>
In article <·······················@194.163.195.67>,
	······@lavielle.com (Rainer Joswig) writes:

>> > Open Genera 2.0 is still in the works.
>> 
>>   Another funny! It's only a little over a year late and counting.
> 
> Progress has been slow but its really getting nearer to being
> released. I have seen it running and it's not that bad.
> 
> Actually I'm starting to think about DEC Alpha prices. ;-)

  You come into an inheritence recently? OpenGenera used to be quite pricey
for an individual. Or do you know of a new price? :-)

  Mike McDonald
  ·······@mikemac.com
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1411980827480001@194.163.195.67>
In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
wrote:

> In article <·······················@194.163.195.67>,
>         ······@lavielle.com (Rainer Joswig) writes:
> 
> >> > Open Genera 2.0 is still in the works.
> >> 
> >>   Another funny! It's only a little over a year late and counting.
> > 
> > Progress has been slow but its really getting nearer to being
> > released. I have seen it running and it's not that bad.
> > 
> > Actually I'm starting to think about DEC Alpha prices. ;-)
> 
>   You come into an inheritence recently? OpenGenera used to be quite pricey
> for an individual.

I will not pay myself. ;-)

> Or do you know of a new price? :-)

It's $5000 for a single machine, multiple emulations at the same
time,  lots of source, lots of software, ...

Sadly I can tell you that this is ******really******* cheap.
Ask the other vendors, who are offering similar Lisp environments
for Unix. Seems like you have to be a big telecom company or in the
military to be able to afford Lisp development under Unix
with a commercial system.

-- 
http://www.lavielle.com/~joswig
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1411982021110001@194.163.195.67>
In article <············@spitting-spider.aracnet.com>, ·······@mikemac.com
wrote:

> > Sadly I can tell you that this is ******really******* cheap.
> > Ask the other vendors, who are offering similar Lisp environments
> > for Unix. 
> 
>   Yup, compared to $4000 for CL and CLIM that some vendors charge, it is
a bargain, if
> you have an Alpha!

$4000 for CL ***and*** CLIM? Which vendor is that?

-- 
http://www.lavielle.com/~joswig
From: Steven Perryman
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72s60j$gsk@bmdhh222.europe.nortel.com>
In article <·······················@194.163.195.67> ······@lavielle.com (Rainer Joswig) writes:

>It's $5000 for a single machine, multiple emulations at the same
>time,  lots of source, lots of software, ...

>Sadly I can tell you that this is ******really******* cheap.
>Ask the other vendors, who are offering similar Lisp environments
>for Unix. Seems like you have to be a big telecom company or in the
>military to be able to afford Lisp development under Unix
>with a commercial system.

Sadly so.
I have a Sun Ultra at home, but you try getting a Solaris eval copy of
something like Lispworks under the same terms that they entertain for stuff
like PCs or Linux ( $50 or free for the latter) . If they just laugh in your
face, I consider that being as 'polite' ...


Regards,
Steven Perryman
·······@nortel.co.uk
From: Pierre Mai
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <87pvarck8e.fsf@dent.isdn.cs.tu-berlin.de>
·······@mikemac.com (Mike McDonald) writes:

> to do so.) So, if anyone has or know someone who has the sources to any of the
> older LispMs, please, PLEASE, speek up!

IIRC the sources for the BBN Lisp Machines' implementation of CL is
available on the CMU Lisp Repository, a file called bbn_cl.tgz.  But
this is mostly Scheme Code IIRRC, since IIRRRC the BBN machines were
Scheme based (somebody who used them might correct me here...).

Regs, Pierre.

-- 
Pierre Mai <····@acm.org>               http://home.pages.de/~trillian/
  "One smaller motivation which, in part, stems from altruism is Microsoft-
   bashing." [Microsoft memo, see http://www.opensource.org/halloween1.html]
From: Keith M. Corbett
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <5C742.1$nI4.1087@brnws01.ne.mediaone.net>
Mike McDonald wrote in message <············@spitting-spider.aracnet.com>...

>(MIT has agreed in principle to the CADR code but they can't find a copy of
>the sources.

Have they contacted the usual suspects? the original spin-off founders?

>TI also seems to have misplaced theirs, although they haven't


I doubt they have the rights to redistribute the MIT and LMI code upon which
they built the Explorer OS.

>said whether they'd be willing to release it anyway. LMI has completely
>disappeared. <snip>

I lost my copies (masters) of the final LMI tapes some time ago. (I turned
out the lights at LMI - twice.) IIRC, when GigaMos folded, there was serious
controversy over who held the assets. Claimants included the founders, a
Japanese investor, the government of Saskatchewan, and employees in
Cambridge and Canada. (I am not making this up.)

>Only Symbolics is still around and they're determined to take

>their code to the grave with them. Which is a shame even if it's their
right
>to do so.) So, if anyone has or know someone who has the sources to any of
the
>older LispMs, please, PLEASE, speek up!


No, no, no, let's not start another thread about copyrights! :>

/kmc
From: Lars Lundback
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <3646D0A2.843DBF25@eralslk.ericsson.se>
········@acm.org wrote:
> 
> What happened?
> 

Reading the LispOS mailing list, and some of the _very_ good remarks on
large software development, in current postings to this newsgroup, tells
you what happened: 

    Most people forgot to put up their basic goals before discussing
activities.

No one was able to put up a convincing goal that related to the question
Marcus G. Daniels put to the LispOS discussion:

   "How is a LispOS supposed to be better than just writing lots
    of programs in a Unix/Win95 Lisp, and using libraries like ILU, 
    etc. to talk to the world? "

And last, no target group for a Lisp OS was identified. But I found this
goal, submitted by Mike McDonald (who could, and should elaborate a bit
more on Why Nothing Happened):

   "Personally, I'd like to have a PC LispM because of
    sentimental reasons and it fits my personal needs.
    But I don't for one second believe there's any chance
    of it being a commercial success."

That description fits my own view of a Lisp OS exactly.

Lars Lundback
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-0911981243180001@pbg3.lavielle.com>
In article <·················@eralslk.ericsson.se>, Lars Lundback
<·······@eralslk.ericsson.se> wrote:

>    "Personally, I'd like to have a PC LispM because of
>     sentimental reasons and it fits my personal needs.

I would have use for such a system (which would be with some work,
some years away).

>     But I don't for one second believe there's any chance
>     of it being a commercial success."

I do.

-- 
http://www.lavielle.com/~joswig
From: Lars Lundback
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <364AFC82.34415DF3@eralslk.ericsson.se>
········@acm.org wrote:
> 
> What happened?
> 
> --
> Fernando D. Mato Mira

Well Fernando,

It seems that we have a priority problem in dealing with a Lisp OS. It's
a bit like the hen - egg situation. Which comes first: The applications
that would really benefit, or the OS illuminating a new scene where
applications can thrive?

I share the vision of a personal LispM that my co-responders have.
Having been into programming for the last 30 years, and sniffing around
Lisp for at least 20, I finally decided a year ago to draw my own map of
a personal pure-Lisp platform. The pages you guys you put on the
Internet have been invaluable. 

My first motive is simple. I think it's a lot of fun. There are some
others: performance and clarity. To quote a novelist:

".. such was the nature of compromise: A condition that satisfied no
one, but left each with the comforting feeling that others had been done
in too."

I suspect that I have been done in too much. I sense that a Lisp OS
could help in bringing back clarity and performance. Then we'll see.

Regards, Lars
From: ········@acm.org
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72fe8e$epk$1@nnrp1.dejanews.com>
In article <·················@eralslk.ericsson.se>,
  Lars Lundback <·······@eralslk.ericsson.se> wrote:

> It seems that we have a priority problem in dealing with a Lisp OS. It's
> a bit like the hen - egg situation. Which comes first: The applications
> that would really benefit, or the OS illuminating a new scene where
> applications can thrive?

Well, there were FORTRAN and C compilers for Symbolics, so I guess
it would not be totally heretic to suggest that a C compiler is
needed. Now, a lot of programs that one might want to run (but _not_
maintain!) require more than C, but also a Unix personality, so maybe
it's necessary to look at the thing from a microkernel perspective.
(Who would care about writing a Unix personality on top of Lisp??)

Do any of the available ukerns look good for a lispm? Does it _have_
or is it better for it to be in Lisp, too? If starting from an available
one, what would need to be added (probably nothing removed, if Unix
is to be supported)?

Eventually, if one does not miss anything from Unix, one should
be able to configure the ukern to remove everything that is not
needed by Lisp, if that makes it more efficient, or avoid it altogether.

--
Fernando D. Mato Mira
Real-Time SW Eng & Networking
CSEM
CH-2007 Neuchatel
Switzerland

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    
From: Kelly Murray
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <364CD04F.72213A81@IntelliMarket.Com>
Back when I worked at the University of Florida, where I could
could actually be paid to waste time working on non-profitable stuff,
I proposed a "New Lisp Machine" to this newgroup.  
This was right when Linux started to come alive 
(I still have floppies with v97 kernels, and I still have
a v98 or 99 Linux running on one of my machines at home.)
I was amazed at the enthusiasm for a clone of 30 year old technology.
There was not much interest in the LispM project back then, so I
wasted my time on other stuff.

A couple years ago, the LispOS came up again, which I believe
was started by a big TCL vs Lisp, bytes vs objects discussion.
The good news is that this time there was more interest,
though the result has been about the same.
Let's hope we get a third concerted effort that makes more progress.

You can ready my view of the issues by reading a collected
archive of my own postings to the mailing list which I 
threw together about a year ago.

http://www.intellimarket.com/silkos.html

I think my ideas and approach are sound, and have trouble understanding
why others don't want to pursue it.  If you think you have
any insight into the issue, please let me know.

In any case, I continue to pursue my own personal effort,
that perhaps by the time I retire when I'm 65,
might actually be "finished", since that is about 25 years from 
now, and Windows2002 written in Java+++ will probably still
be the operating system in use...

-Kelly Murray
From: Mike McDonald
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72im9v$osp$1@spitting-spider.aracnet.com>
In article <·················@intellimarket.com>,
	Kelly Murray <···@IntelliMarket.Com> writes:
> Back when I worked at the University of Florida, where I could
> could actually be paid to waste time working on non-profitable stuff,
> I proposed a "New Lisp Machine" to this newgroup. 

  I remember that! I still hold the same position that I did then. (It should
be X based and run on top of some other OS (use the underlying OS as the machine
dependant layer so we don't have to write hideous device drivers.).
 
> This was right when Linux started to come alive 
> (I still have floppies with v97 kernels, and I still have
> a v98 or 99 Linux running on one of my machines at home.)
> I was amazed at the enthusiasm for a clone of 30 year old technology.
> There was not much interest in the LispM project back then, so I
> wasted my time on other stuff.
> 
> A couple years ago, the LispOS came up again, which I believe
> was started by a big TCL vs Lisp, bytes vs objects discussion.
> The good news is that this time there was more interest,
> though the result has been about the same.
> Let's hope we get a third concerted effort that makes more progress.
> 
> You can ready my view of the issues by reading a collected
> archive of my own postings to the mailing list which I 
> threw together about a year ago.
> 
> http://www.intellimarket.com/silkos.html
> 
> I think my ideas and approach are sound, and have trouble understanding
> why others don't want to pursue it.  If you think you have
> any insight into the issue, please let me know.

  At the time, you were only letting out precompiled binaries for the Solaris
version of ACL. (I hope I remember the version correctly.) Since most of us
aren't Franz customers (anymore) and we don't have Sun boxes, it had limited
appeal to me. I, personally, am not interested in a HTML based interface. If I
wanted that, I'd join the JOS group. At the minimum, I'd like an X based
equivilant of a Symbolics environment. But that's me.

> In any case, I continue to pursue my own personal effort,
> that perhaps by the time I retire when I'm 65,
> might actually be "finished", since that is about 25 years from 
> now, and Windows2002 written in Java+++ will probably still
> be the operating system in use...
> 
> -Kelly Murray

  I'm still working on my part of the LispOS. Since it's such a bugger, I
might be done in time for your retirement!

  Mike McDonald
  ·······@mikemac.com
From: ········@acm.org
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72ov7j$3jk$1@nnrp1.dejanews.com>
If on Linux, I would go with MkLinux for several reasons:

- Evolutionary path to Mach
- I don't care about learning x86 asm
- I'd rather buy the prettiest and most compatible laptop
  available : MacOS [X]; *Linux; Windows* on VPC
  (Linux and Solaris on VPC, too?)

The only problem is that PPC is still a 32-bit arch
only.

Maybe there could be some marketing interest for
Apple to provide some kind of sponsoring? There are
also lisp-friendly insiders, and the company likes
well-designed things..

Lisp @ Apple - Chapter I: TI Explorer
Lisp @ Apple - Chapter II: Dylan
Lisp @ Apple - Chapter III: Lisp OS?

--
Fernando D. Mato Mira
Real-Time SW Eng & Networking
CSEM
CH-2007 Neuchatel
Switzerland

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1611981531460001@pbg3.lavielle.com>
In article <············@nnrp1.dejanews.com>, ········@acm.org wrote:

> If on Linux, I would go with MkLinux for several reasons:
> 
> - Evolutionary path to Mach
> - I don't care about learning x86 asm
> - I'd rather buy the prettiest and most compatible laptop
>   available : MacOS [X]; *Linux; Windows* on VPC
>   (Linux and Solaris on VPC, too?)
> 
> The only problem is that PPC is still a 32-bit arch
> only.
> 
> Maybe there could be some marketing interest for
> Apple to provide some kind of sponsoring? There are
> also lisp-friendly insiders, and the company likes
> well-designed things..
> 
> Lisp @ Apple - Chapter I: TI Explorer
> Lisp @ Apple - Chapter II: Dylan
> Lisp @ Apple - Chapter III: Lisp OS?
> 
> --
> Fernando D. Mato Mira
> Real-Time SW Eng & Networking
> CSEM
> CH-2007 Neuchatel
> Switzerland
> 
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own    


My extension of your list:

Lisp @ Apple - Chapter 0: Coral Lisp
Lisp @ Apple - Chapter Ia: TI Explorer
Lisp @ Apple - Chapter Ib: Symbolics MacIvory
Lisp @ Apple - Chapter II: Macintosh Common Lisp (God's own development
enviroment)
Lisp @ Apple - Chapter III: AppleScript (looks a lot like Lisp
semantically, see SK8Script)
Lisp @ Apple - Chapter IV: Dylan
Lisp @ Apple - Chapter V: Newton (borrows a lot from Lisp)
Lisp @ Apple - Chapter VI: SK8
Lisp @ Apple - Chapter VII: MCL rescued
Lisp @ Apple - Chapter VIII: No more Lisp inside Apple: Cambridge Lab,
Newton Inc. and ATG closed
Lisp @ Apple - Chapter IX: MCL screaming on G3 machines

But I doubt there will be a Chapter X: Lisp OS.
Reason: Apple is very closed. No more Clones. No technical infos for Be.
Atleast you can get some info from the Linux for Mac sources.

If you want to have a Lisp machine on a Mac now: Replace the Finder with MCL.
Should be possible with some additional work (changing the name, setting
some bits).
Then boot. I haven't tried it - yet. ;-)

-- 
http://www.lavielle.com/~joswig
From: Fernando D. Mato Mira
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <36506A22.6263A317@acm.org>
Rainer Joswig wrote:

> If you want to have a Lisp machine on a Mac now: Replace the Finder with MCL.

The idea is to run MacOS, Linux and the LispOS at the same time.

--
Fernando D. Mato Mira
Real-Time SW Eng & Networking
Advanced Systems Engineering Division
CSEM                                   HTTP:     www.csemne.ch
Jaquet-Droz 1                         email:  ········@acm.org
CH-2007 Neuchatel                       tel: +41 (32) 720-5157
Switzerland                             FAX: +41 (32) 720-5720
From: Rainer Joswig
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <joswig-1611981957020001@pbg3.lavielle.com>
In article <·················@acm.org>, "Fernando D. Mato Mira"
<········@acm.org> wrote:

> Rainer Joswig wrote:
> 
> > If you want to have a Lisp machine on a Mac now: Replace the Finder
with MCL.
> 
> The idea is to run MacOS, Linux and the LispOS at the same time.

How does this work? AFAIK not even MacOS and Linux are able
to run at the same time...

-- 
http://www.lavielle.com/~joswig
From: Fernando D. Mato Mira
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <365152CB.6F6BE495@acm.org>
Rainer Joswig wrote:

> How does this work? AFAIK not even MacOS and Linux are able
> to run at the same time...

But it's still not MacOS X.
MKLinux runs as a user-mode mach thread, so theoretically..


--
Fernando D. Mato Mira
Real-Time SW Eng & Networking
Advanced Systems Engineering Division
CSEM                                   HTTP:       www.csem.ch
Jaquet-Droz 1                         email:  ········@acm.org
CH-2007 Neuchatel                       tel: +41 (32) 720-5157
Switzerland                             FAX: +41 (32) 720-5720
From: [Invalid-From-Line]
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <slrn75dd5m.203.gb@jill.westfalen.de>
On Mon, 16 Nov 1998 19:57:02 +0100, Rainer Joswig <······@lavielle.com> wrote:
>How does this work? AFAIK not even MacOS and Linux are able
>to run at the same time...

There might be some solution to that, as soon as Sheepshaver for LinuxPPC
exists :-)

bye, Georg
From: Andi Kleen
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <m3g1bjitkz.fsf@fred.muc.de>
In article <············@nnrp1.dejanews.com>,
········@acm.org writes:
> If on Linux, I would go with MkLinux for several reasons:
> - Evolutionary path to Mach

Mach seems to be a complete dead end. 

-Andi
From: CsO
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <72qism$cm8$1@phys-ma.sol.co.uk>
Andi Kleen wrote in message ...
>In article <············@nnrp1.dejanews.com>,
>········@acm.org writes:
>> If on Linux, I would go with MkLinux for several reasons:
>> - Evolutionary path to Mach
>Mach seems to be a complete dead end.


hi
whatever happened to the gnu hurd?
From: Christopher B. Browne
Subject: Re: LispOS, LispVM projects?
Date: 
Message-ID: <slrn7522ci.ajh.cbbrowne@godel.brownes.org>
On Tue, 17 Nov 1998 01:18:06 -0000, CsO <···@mindless.com> posted:
>Andi Kleen wrote in message ...
>>In article <············@nnrp1.dejanews.com>,
>>········@acm.org writes:
>>> If on Linux, I would go with MkLinux for several reasons:
>>> - Evolutionary path to Mach
>>Mach seems to be a complete dead end.
>whatever happened to the gnu hurd?

Development continues, probably most quickly in the form of the Debian Hurd
project.  See <http://www.debian.org>, and look for mailing lists...

It has definitely not progressed *real* fast...

-- 
Those who do not understand Unix are condemned to reinvent it, poorly. 	
-- Henry Spencer          <http://www.hex.net/~cbbrowne/lsf.html>
········@hex.net - "What have you contributed to Linux today?..."