Videos from around the APL community.
The penultimate set of videos from Dyalog '16 has been released
These presentations are all related to APL's 50th anniversary next Sunday; for more on the golden anniversary, see more at DYALOG 16
L01: In the 1970s, Roy Sykes (Sykes Systems) worked for Scientific Time Sharing Corporation (STSC). He takes us back to the days of punch cards, when $250,000 bought a 300 MB disk complex and APL workspaces couldn't exceed 28 KB because the mainframe only had 32 KB and needed room for the operating system. On a shared mainframe, STSC and I.P. Sharp Associates (IPSA), competitors in the business world, partnered to offer time sharing services, supporting a whopping 30-40 simultaneous users (as long as there wasn't a thunderstorm). Looking back today, this may seem like a very difficult environment to work in, but at the time most users felt that the system made them almost omnipotent, and inhabited a friendly, sharing universe. APL was groovy, man!
L02: Datatypes are native to a language; they tend to be pervasive and can have a profound effect as all primitives have to be re-evaluated when a new datatype is introduced. Bob Smith (Sudley Place Software) looks at the evolution of APL through its supported datatypes; these include the versatile, compact and highly-performant Booleans that have been in almost every implementation of APL since APL\360, nested arrays, which are the only structural datatype as well as possibly the most significant and controversial addition to APL, complex (and hypercomplex) numbers, which extend the number of dimensions in which a number can be defined and are essential for real-time video games, and many others.
L03: In search of a more satisfying career, Bob Bernecky's (Snake Island Research) life was changed when he was hired by I.P. Sharp Associates (IPSA) in 1971. He started working on the COBOL compiler, investigating the data-type conversion algorithms and coding unit tests. At this point Roger Moore suggested he use APL to write the unit tests and pointed him towards Doug Forkes. Bob describes how he transitioned to life as a member of the APL Development Group, worked on improving the performance of various operations and developed high precision timers to eliminate billing oddities in the timesharing system. As to how the APL Development Group became known as The Zoo…
L04: Stephen Taylor (Equiniti Claybrook) takes on the role of chat show host when he interviews Dyalog's Geoff and JohnS about their careers in APL. Geoff was around for the formation of Dyadic Systems in the late '70s and the ensuing consultancy work and time-sharing systems; JohnS came on board soon afterwards to help write the first version of the Dyalog interpreter (written for a UNIX mini-computer). Before the internet and email, two "arrogant, fairly young, guys" implemented Bob Smith's blue book (they didn't know about APL2 at that point) and fortunately IBM choose to make language decisions that were very similar to those they had selected. Now in 2016 they're still working on the Dyalog interpreter (JohnS describes it as "almost finished"). Along the way there were nasty basements, arrests and too many ports…
L05: Twenty-five years ago, k and q were spawned from APL. Numerous small decisions made since then have had far-reaching and irreversible impacts, both within APL and within k and q. Simon Garland (Kx Systems Inc) looks at some of the choices made and their consequences today, choices that each had a ripple effect and that might have been made differently with the benefit of hindsight. Particular instances of divergence from APL include not having a workspace (no legacy support and a tiny 300 KB executable), not allowing users to change system variables, having timestamps with a nanosecond precision, permitting null values and introducing symbols (immutable strings).
L06: Niels Hallenberg (SimCorp) explains how business values drive the decision of what language to use. For SimCorp Dimension, SimCorp's integrated enterprise system for large investment managers, 250+ multi-discipline developers need to use the same language and that language has to be able to integrate with other technologies and be amenable to agile methodologies, as well as be future proof. APL was selected 20 years ago and is still the language of choice, although it is also supplemented by other 150 developers working with HTML, Ocaml, C++, C#, etc. Using APL allows the developers to focus on the business and is cost efficient. Niels then introduces the enhancements he would like to see within Dyalog to support SimCorp dimension for the next decades (increased parallelism, high precision numbers, cache shared by multiple Sessions, etc.).
L07: In the 1970s the Arab oil embargo meant that oil companies had to select the best crude oil from a range of options, each of which had a different price and composition (prior to this each refinery processed the most local crude oil). Composition-based models were devised and analytic techniques employed in the search for the most accurate representation, as this would result in identifying the crude oil that offered the largest profit. Stephen Jaffe (ExxonMobil – retired) explains his SOL (Structure Oriented Lumping) program that describes the reactions of hydrocarbon mixtures. Written in APL, individual hydrocarbon molecules are represented as vectors of incremental structural features – these provide a framework to construct rule-based reaction networks of arbitrary size and complexity.
|APL demonstration at Google Labs||Simple "Hello World" APL .EXE||APL Advanced data manipulation|
|APL program for the "Game of Life"||APL program for Robots using Raspberry PI||Fun and Games with APL to advance Learning|
|Morten Kromberg - PLDI 2016 ARRAY Workshop Keynote"|
© Copyright 2017 APLCloud.com, all rights reserved.