News for hackers. Stuff that matters.
Re:A Story about Symbolics Lisp Machines by RockyMcNuts on Monday September 21, 2015 @08:02PM (#10255921)
I don't think Lisp ran well if at all on early microprocessors like the 8-bit 8086. The Lisp community was, then, as now, committed to Lisp being the One True Way, and centered around the MIT AI lab. The PC revolution was happening, so naturally they built their own CPU, OS and workstations around Lisp. Half the MIT AI lab ended up going to Symbolics, and half to LMI. The West Coast hackers out of Berkeley and Stanford were more into C and Unix. Once the 68000 and the first Sun (especially), also Apollo, SGI workstations came out, they performed well at a broad range of tasks including even Lisp. C compiled software often outperformed similar software written in Lisp. And when volume ramped up (for 68000-based Macintosh as well) they had much better cost/performance compared to relatively low-volume Lisp machines. (like $30,000 for a decent Sun workstation vs. something like twice as much for a Symbolics)

See Hackers, by Steven Levy, also some stuff in Wikipedia

https://en.wikipedia.org/wiki/Lisp_Machines

https://en.wikipedia.org/wiki/Symbolics

Had a summer job doing UI work on a Symbolics machine at MITRE, which was kind of like working in the future in summer of 1985... object oriented development, megapixel WIMP interface. I don't think Sun came close to that until UIM/X, Motif, PowerBuilder in early 90s, Visual C++, VB on PCs...like a 10-year head start.

But python especially feels like the revenge of Scheme and Lisp in the sense of being the same sort of dynamic interpreted language and environment where code is data, all the data and code (including libraries) are very accessible and amenable to inspection and tweaking.

Re:A Story about Symbolics Lisp Machines by eggy on Tuesday September 22, 2015 @07:18AM (#10257850)
Python is no where near Lisp in code as data. Also, I think what's important is to marry the Lisp programming language and the Lisp Machines. The hardware was built with Lisp in mind. I am a fifth of the way through a 1991 book called "The Architecture of Symbolic Computers" by Peter M. Kogge, and it is like my eyes have opened wide. No more splitting hairs over Ruby, Python, Clojure, Linux, OSX and Windows: Give me a machine tailored for the software running on it. the Von Neumann register-based machine cannot efficiently serve all, it bottlenecks. People talk about Linux as the break away from the Windows and Mac OS crowd, but let's face it, it is all the same thing running on Von Neuman architecture hardware. There are people trying to build machines from scratch to run their own OS on. I hope somebody succeeds. I don't think we need to go back to Lisp Machines, but there's a lot of gold nuggets there to mine and re-use. Technology doesn't age linearly. Sometimes things are forgotten, and need to be rediscovered for progress to jump again. I work in Common Lisp (SBCL, LispWorks), Racket (Scheme), and I play with PicoLisp (incredible for its size). Shen is something I am trying to wrap my head around (optional typing, and optional laziness, functional, pattern matching, and there implementations of the small instruction set it runs on for many common languages - CL (SBCL), Java, Ruby, Python, JavaScript, Haskell, Clojure...

Here's to the coming hardware revolution!

Re:A Story about Symbolics Lisp Machines by Taniwha on Tuesday September 22, 2015 @01:04AM (#10256710)
Sun was just spinning up - but remember that first Mac (3/4M pixel WIMP interface) had become available the year before for $2400 (and I was making a living in '84/85 porting Unix to 68ks as fast as we could, every man and their dog had built a workstation in their garage)
Re:A Story about Symbolics Lisp Machines by pdw on Tuesday September 22, 2015 @04:21AM (#10257324)
The early Macs had tiny displays, something like 512x342. The first Macs with a resolution of 1024x768 (¾ megapixel) wouldn't appear until the 90s.
Re:A Story about Symbolics Lisp Machines by draven on Tuesday September 22, 2015 @04:02AM (#10257261)
How is python homoiconic?
Re:A Story about Symbolics Lisp Machines by Volt on Tuesday September 22, 2015 @04:51AM (#10257432)
He didn't say it was.
Re:A Story about Symbolics Lisp Machines by Guthur on Tuesday September 22, 2015 @06:10AM (#10257651)
> But python especially feels like the revenge of Scheme and Lisp in the sense of being the same sort of dynamic interpreted language and environment where code is data

He implied it by saying that code is data in Python. Being able to inspect code is not the equivalent.

It is also incorrect to refer to Scheme and Lisp as just dynamic interpreted languages; while they are dynamic and tend to have an interpreter they have also had compilers for quite some time if not from near the very beginning (1957 according to this http://www.softwarepreservation.org/projects/LISP/ibm/Blair-...).

Re:A Story about Symbolics Lisp Machines by TeMPOraL on Tuesday September 22, 2015 @06:15AM (#10257668)
He didn't, but it was implied in "But python especially feels like the revenge of Scheme and Lisp in the sense of being the same sort of dynamic interpreted language and environment where code is data".

Python doesn't feel to me like revenge of Lisp. It's more like having a shard of a powerful magic artefact; it has some power, but a whole artefact has qualitatively more.

Re:A Story about Symbolics Lisp Machines by protomyth on Tuesday September 22, 2015 @07:36AM (#10257899)
"I don't think Lisp ran well if at all on early microprocessors like the 8-bit 8086"

The 8086 was a 16-bit CPU, but you are correct that Lisp didn't run well on it.

Re:A Story about Symbolics Lisp Machines by RockyMcNuts on Tuesday September 22, 2015 @10:07AM (#10258633)
my bad lol ... was thinking 8088, Z80
Re:A Story about Symbolics Lisp Machines by protomyth on Tuesday September 22, 2015 @10:34AM (#10258806)
Yep, the Z80 was an 8-bit CPU, but the 8088 was regarded as a 16-bit CPU. It did have a 8-bit external data bus (the 8086 had 16-bit) to make it cheaper, but the register sizes and instructions were the same as the 8086.
Re:A Story about Symbolics Lisp Machines by dTal on Wednesday September 23, 2015 @11:33AM (#10265644)
It's not quite as straightforward as that with the Z80 either - it has 16 bit addressing, a few 16-bit registers (which can also be treated as two separate 8-bit registers), and instructions for manipulating 16-bit numbers. The 16-bit operations feel a bit like second-class citizens, which I suppose is why it's still regarded as an 8-bit CPU, but they are there.
Re:A Story about Symbolics Lisp Machines by JoeAltmaier on Tuesday September 22, 2015 @10:19AM (#10258706)
Hey, the 80186 was 8-bit, so maybe you were thinking of that?
Re:A Story about Symbolics Lisp Machines by protomyth on Tuesday September 22, 2015 @10:32AM (#10258792)
The 80186 was also a 16-bit chip with a 20-bit address (like the 8086).
Re:A Story about Symbolics Lisp Machines by JoeAltmaier on Tuesday September 22, 2015 @10:35AM (#10258819)
8-bit bus interface?
Re:A Story about Symbolics Lisp Machines by protomyth on Tuesday September 22, 2015 @10:44AM (#10258886)
The 8088, 80188 had 8-bit buses, but I haven't met anyone who claimed that made them 8-bit chips. Neither really affected programmers and it was a maneuver to make the chip a bit cheaper. The 68008 was the same way, and used in some Sinclair designs.
Re:A Story about Symbolics Lisp Machines by JoeAltmaier on Tuesday September 22, 2015 @10:49AM (#10258921)
Right, that makes sense.

Programmatically they were still 16-bit. But as an OS engineer, the 8-bit bus was very much evident when writing drivers, moving data, initializing memory etc.

Re:A Story about Symbolics Lisp Machines by protomyth on Tuesday September 22, 2015 @11:12AM (#10259079)
How would the code be different between the 8088 and 8086? The execution unit and internal bus are exactly the same between the chips. The only thing I can think of is that it took twice as long to write data out, but it was the exact same instructions. The prefetch queues were different sizes also.
Re:A Story about Symbolics Lisp Machines by JoeAltmaier on Tuesday September 22, 2015 @11:51AM (#10259370)
Driver writers care about latency, program DMA devices, set up memory factors (RAS/CAS, bank addresses etc) and bus width is visible there. Also atomic operations were done with bus-lock - a 16-bit update took two bus cycles! So the high and low parts could be non-atomically updated.
Re:A Story about Symbolics Lisp Machines by protomyth on Tuesday September 22, 2015 @01:04PM (#10259944)
Thanks. I wonder what Microsoft did for MS-DOS? I remember a few clones with 86's instead of 88's.
Re:A Story about Symbolics Lisp Machines by JoeAltmaier on Tuesday September 22, 2015 @03:11PM (#10260901)
DOS isn't an OS at all. The BIOS did the hardware-specific stuff, including the floppy driver(!) DOS was almost pure software services.

For cases where DOS had used an instruction that would change, Intel rolled over and crippled that instruction in 16-bit machines (at least for a while) so it would behave the same.