APL\B5500: The Language and Its Implementation – Gary A. Kildall (1970) [pdf] | Hacker News Reader

APL\B5500: The Language and Its Implementation – Gary A. Kildall (1970) [pdf]

cs.washington.edu

83 pointsvezzy-fnord27 comments

A few historical implementations of APL are very interesting. The Burroughs 220 version by Larry Breed was the defining factor of what became the APL environment. Jim Ryan's APL for the Burroughs 6500 was nothing short of amazing and pioneered many of the performance enhancement mechanisms now used in production optimizing compilers. John Walters (ex-IBM) built an APL compiler for HP which compiled very good code but retained the interactive character of APL. Much of the early history is documented in the ACM Quote-Quad.

The user manual for Jim Ryan's APL\700 implementation is here:

http://www.softwarepreservation.org/projects/apl/Manuals/APL...

If we can get a working Burroughs B6700 emulator (see elsewhere in this thread) we may have enough pieces to see APL\700 working again. We're seeking the APL\700 Installation Manual (Form 5000805) if anyone has a copy.

Note the use of the "\" in the name was intentional according to Jim Ryan, "APL "expands" the '700 series" (comp.lang.apl)

As drallison notes, APL\700 was a compiler to an intermediate language.

The likes of Hoare, Dijkstra and Knuth hacked on these Burroughs machines:

https://en.wikipedia.org/wiki/Burroughs_large_systems#Langua...

Pardon my slip; Dijkstra didn't "hack", of course.

If you have any interest in these machines, the web-based B5500 emulator is a fantastic toy:

http://www.phkimpel.us/B5500/

Source-code here: https://github.com/pkimpel/retro-b5500

including Kildall's APL implementation (still needs to be proof read), but also note that Niklaus Wirth's programming language predecessor to Pascal known as Euler is there too.

Another interesting piece of source-code you will find there too is a Burroughs B6500 emulator, B6500 MCP and B6500 ESPOL compiler written for the B5500. The B5500 was used by Burroughs as the development platform for the B6500 which was the follow-on from the B5000 family.

You can see the B6500 emulator running in this short film: https://www.youtube.com/watch?v=KOElQuCqj0Y#t=3m37s

I'm hoping we can re-create those few seconds of video once we have this B6500 software functional again.

We and others are working on emulators for the Burroughs B6700 too. If anyone can help find software for any of the Burroughs systems from that era, in particular the ESPOL compiler and the so-called INTRINSICS libraries it would be appreciated. Any software or documentation for Burroughs systems from that era would be helpful too, the related systems include Burroughs B6700, B7700, B6800, B7800, B6900, B5900 and B7900 (listed in order of release date). We may use later MCP revisions to help cross-compile to earlier versions or identify system related changes.

"Hacking" has at least three different sides. One variety is engineering by "hacking" up potentially kludgy solutions. Another is the process of learning about systems through hands on exploration. Yet another is exploitation of systems, typically involving the knowledge learned from the exploratory kind of hacking.

There's nothing shameful about the 2nd form of hacking, it's just a form of science.

Kinda off topic: Those who are interested in modern implementations of APL and similar vector languages are invited to join #kq on Freenode IRC.

We have John Earnest who implemented K5 in Javascript (oK.js) and Kevin Lawler who worked on both Kona (K3) and his new friendlier Kerf.

Some really interesting discussion about array languages has been happening there lately. Join us!

APL, J, and occasionally K conversation also happens on #jsoftware on freenode. Thanks for pointing to #kq, which I didn't know about.

Is this only for theoretical discussions about language design, etc.? Or can we ask dumb questions about how to do simple things in k and q? Not really a fan of Google Groups. It would be great to have an IRC channel for common questions.

We just happen to talk about other implementations due to the Q licensing thing from a few weeks back. Of course we love to help newbies learn the language too. There are some real k/q geniuses in there (I'm not one of them). Come say hi! I'm tlack.

Wow, I had no idea "Mr. CP/M" was into APL...

He was into quite a lot of things. One of the rare examples of a person who was simultaneously a skilled programmer, a computer scientist and a business-savvy entrepreneur (though one sadly plagued by a lot of obstacles and circumstance).

I would go as far as to say that, at least as far as individuals go, he was the single most important in initiating the PC revolution, doubtlessly more so than Bill Gates or Steve Jobs or even Woz from the perspective of fundamentals.

He is also virtually unknown to the public. Go figure.

He also invented dataflow analysis, which is the basis for a huge number of compiler optimization and program analysis techniques: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.102....

The contributions, from the Wikipedia entry about him:

"In March 1995, Kildall was posthumously honored by the Software Publishers Association (now the Software and Information Industry Association) for his contributions to the microcomputer industry:[3]

- Introduction of operating systems with preemptive multitasking and windowing capabilities and menu-driven user interfaces.

- Creation of the first diskette track buffering schemes, read-ahead algorithms, file directory caches, and RAM disk emulators.

- Introduction of a binary recompiler in the 1980s.

- The first programming language and first compiler specifically for microprocessors.

- The first microprocessor disk operating system, which eventually sold a quarter of a million copies.

- The first computer interface for video disks to allow automatic nonlinear playback, presaging today's interactive multimedia.

- The file system and data structures for the first consumer CD-ROM.

- The first successful open system architecture by segregating system-specific hardware interfaces in a set of BIOS routines.[22][23][24][25] "

He was a regular contributor to The Computer Chronicles and can be seen in many episodes.

https://archive.org/details/computerchronicles

I really enjoyed this retrospective on Gary Kildall. He was a really interesting guy, and I'll echo the sentiments of my sibling commenter-- a major influence on the development of the personal computer.

https://www.youtube.com/watch?v=Tdj8gh9GPc4

The list of people who implemented (Quasi-)APLs contains some further illustrious names:

- Ken Thompson (APL\11): https://groups.google.com/forum/#!topic/comp.lang.apl/u9sl6g...

- Dan Bricklin (Multics APL): http://lowendmac.com/2006/dan-bricklin-inventor-of-the-elect...

- Bill Gates (designed, never published): http://americanhistory.si.edu/comphist/gates.htm#tc30

- Robert Griesemer: https://en.wikibooks.org/wiki/Go_From_Scratch/The_Origin_of_...

- Rob Pike: https://github.com/robpike/ivy

Wonderful! But just think of the guy that had to manually type up these diagrams on on page 22 and 28 (at least I assume the text is done on a typewriter with the arrows and balloons being added later by hand?).

We had Burroughs mainframes at college (not sure why, although there was a Burroughs factory about 20 miles away so that might have been the reason). Algol 68. Good times..

Someone implemented ALGOL 68 for the Burroughs 5500 or descendant? I'm way envious.

If anyone finds a reference to its existence we would appreciate the details please. Bauer and DeRemer's Compiler Construction text speculates about ALGOL 68 on a B5500 but we've not found any evidence so far it was attempted. As you will see about the discussion of ALGOL-68R (https://en.wikipedia.org/wiki/ALGOL_68-R - the first implementation of ALGOL 68) it took considerable resources to support it, likely more than the B5500 could provision.

This was in the B6700 era, not B5500. Even so, that was a long time ago, and my memory is hazy. However, I do have an Algol68 book that dates from those times and I doubt I would have bought said book if I didn't have a machine to use it with. This old book also seems to imply that there was Algol68 on the B6800: http://www.softwarepreservation.org/projects/ALGOL/paper/ALG...

This is a great find! Not to hijack the thread - but does anyone know the status of "chasing men who stare at arrays"?

I like to think the director had a political epiphany concerning all the finance quant APL users.

Thank you to everyone who has kept my project alive in their hearts and minds. I am still on hiatus and while I do plan to start it up again I am not sure when.

In the mean time:

- Roger Moore has been working on his Wikipedia entry and notes: "It seemed easier to find citations from my brief time in academe than the remainder of my life. I will have to wait and see what Wikipedia editbot thinks of it." https://en.wikipedia.org/wiki/Roger_Moore_(computer_scientis...

-As for my father, there's not much new on the programming language front however his Alpaca business is holding its own: http://www.manitoulin.ca/2015/06/30/lobo-loco-alpaca-farm-th...

Mary and Sally. Hurray!

Reply on news.ycombinator.com