Driverless alternatives

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Driverless alternatives

KenDickey
No good graphics system, but the Scheme language is _much_ better & safer than C++ !

Scheme48 has been used to program mobile robots.  A formal proof system verified the implementation http://mumble.net/~jar/pubs/lsc.pdf  Scheme48 uses a bytecode interpreter but has a PreScheme compiler which converts a pure Scheme subset to C.  This was the work that inspired Slang (IMHO).  

        http://s48.org

Gambit Scheme compiles to native code and is roughtly compatable with C in speed.  had been used to program modest hardware and they did a real-time GC implementation.

        http://dynamo.iro.umontreal.ca/wiki/index.php/Main_Page

Among other things, Scheme is a great system in which to write very efficient state machines.

$0.02
--
Ken [dot] Dickey [at] whidbey [dot] com

_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org
-KenD
Reply | Threaded
Open this post in threaded view
|

Re: Driverless alternatives

baulamon
Hello,
Ambit Scheme compiles to native code and is roughtly compatable with C in speed.  had been used to program modest hardware and they did a real-time GC implementation.
Reply | Threaded
Open this post in threaded view
|

Re: Driverless alternatives

Casey Ransberger-2
In reply to this post by KenDickey
Not sure about the inspiration for Slang bit. For some reason I think the idea came from from experience that Dan Ingalls had with translating something-or-other to BCPL. But I can't back that up because I can't remember where I read it off the top of my head… so I'll shut up:)


On Wed, Dec 18, 2013 at 8:47 AM, Ken Dickey <[hidden email]> wrote:
No good graphics system, but the Scheme language is _much_ better & safer than C++ !

Scheme48 has been used to program mobile robots.  A formal proof system verified the implementation http://mumble.net/~jar/pubs/lsc.pdf  Scheme48 uses a bytecode interpreter but has a PreScheme compiler which converts a pure Scheme subset to C.  This was the work that inspired Slang (IMHO).

        http://s48.org

Gambit Scheme compiles to native code and is roughtly compatable with C in speed.  had been used to program modest hardware and they did a real-time GC implementation.

        http://dynamo.iro.umontreal.ca/wiki/index.php/Main_Page

Among other things, Scheme is a great system in which to write very efficient state machines.

$0.02
--
Ken [dot] Dickey [at] whidbey [dot] com

_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org


_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org
Reply | Threaded
Open this post in threaded view
|

Re: Driverless alternatives

KenDickey
On Wed, 18 Dec 2013 21:23:04 -0800
Casey Ransberger <[hidden email]> wrote:

> Not sure about the inspiration for Slang bit. For some reason I think the
> idea came from from experience that Dan Ingalls had with translating
> something-or-other to BCPL. But I can't back that up because I can't
> remember where I read it off the top of my head… so I'll shut up:)

Bitrot here as well.  I do remember talking with someone on the team about PreScheme, perhaps John Maloney.  I did some small bytecode benchmarks for him to compare against the original Squeak bytecode VM when they were first getting up (before 1.0).  I think I did code for MacScheme, Scheme48, possibly Smalltalk Agents.  But again my memory is no longer to be trusted.

Bytecode VMs were not a big world then, so I think most everyone knew everything that was going on.   Ideas are pretty viral.  I think that perhaps ideas share people rather than the other way around.  The idea of "being first" to think of something is rather suspect.  PreScheme does significantly predate Slang, however.  

Kali (distributed Scheme48 derivative) was squirting bytecodes around the net significantly before Java as well -- including full closures.  

The Java bytecode VM was put together by a couple of guys who were not well versed in VMs by the way.  An experiment that got out of control so to speak, which is why the original Java (which did at least fit on a 'floppy' disk) was always very slow compared to the VMs of the day.  The bytecode set got frozen before they had a chance to make things fast.  After taking a look, I did suggest that they reserve one bytecode to be able to swap/transition VMs to a different bytecode set, which would have allowed them to run Smalltalk, Scheme, whatever bytecodes mixed in the same environment.  But they used bytecodes for "add 1 byte", "add 1 short", "add 1 long", rather than "add 1" and just burnt up all the opcodes, sigh.  Opportunities lost.  

Can't keep losers from losing.  Better to focus on opportunities to help winners win big.

Cheers,
-KenD

> > On Wed, Dec 18, 2013 at 8:47 AM, Ken Dickey <[hidden email]> wrote:
> > .. Scheme48
> > uses a bytecode interpreter but has a PreScheme compiler which converts a
> > pure Scheme subset to C.  This was the work that inspired Slang (IMHO).
..

_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org
-KenD