Email or username:

Password:

Forgot your password?
lj·rk's posts Post Back to profile
Top-level
lj·rk

Lynn developed "generalized dynamic instruction dispatch" for IBM in 1966. 2 years later she was kicked out, just after Robert Tomasulo published the "Tomasulo Algorithm" for out-of-order execution of floating point instructions, utilizing Lynn's work. Everyone knows Tomasulo (and he did great work, mind you!), but no-one knows Lynn.

Later, in technical compsci, you may stumble upon highly integrated circuits, everyone there knows #VLSI, but not the inventor, our dear Dr. Conway.

Her story, her struggle against IBM who took decades to apologize to her for her mistreatment. She transitioned in darker times and pioneered not "only" in compsci. She was what many would call "greater than life". She died a few days ago.

Today, let's remember Lynn 🏳️‍⚧️, tomorrow we'll fight on ✊

2/2

14 June at 11:03 | Open on todon.eu
15 comments
Wolfgang

@ljrk Sally Floyd, who did pioneering work in Internet congestion control.

en.m.wikipedia.org/wiki/Sally_

14 June at 15:10 | Open on darmstadt.social
Sobex

@ljrk Holy shit, Out of Order execution is based off her work ? What’s the share of contribution in the seminal paper between her and Tomasulo ?

14 June at 15:36 | Open on social.sciences.re
lj·rk

@Sobex Jup. Tomasulo himself did original work in his paper indeed and I wouldn't steal any credit from him, but nowadays multiple-issue OoO execution based on both Tomasulo's and Conway's work in roughly the same share is what's actually utilized. The name of Conway is usually dropped and both concepts (register renaming etc. and multiple-issue) subsumed under one, effectively erasing Conway's work. If it isn't then multiple-issue is often erroneously attributed to Yale Patt.

There's an interesting discussion going on since '12 on WP around that if you want to dive down that rabbit hole :'D

en.wikipedia.org/wiki/Talk%3AO

@Sobex Jup. Tomasulo himself did original work in his paper indeed and I wouldn't steal any credit from him, but nowadays multiple-issue OoO execution based on both Tomasulo's and Conway's work in roughly the same share is what's actually utilized. The name of Conway is usually dropped and both concepts (register renaming etc. and multiple-issue) subsumed under one, effectively erasing Conway's work. If it isn't then multiple-issue is often erroneously attributed to Yale Patt.

14 June at 18:20 | Open on todon.eu
David J. Atkinson #🟦

@ljrk I met Lynn Conway two times when she was at Univ of Michigan. Brilliant person!

14 June at 20:33 | Open on c.im
Eliot Lash

@ljrk Shout out to the original ENIAC programming team: Kay McNulty, Betty Jennings, Betty Snyder, Marlyn Wescoff, Fran Bilas and Ruth Lichterman.

I watched some interviews with them at the Computer History Museum. This doc also looks cool:
eniacprogrammers.org/

I have heard that some black women were also involved with this project but sadly this has not been well documented and information about them may have been lost in the historical record.

14 June at 23:12 | Open on social.coop
Eliot Lash

@ljrk Also shout out to Klára Dán von Neumann (John von Neumann's wife) also considered to be one of the first programmers:
en.wikipedia.org/wiki/Kl%C3%A1

She was Head of the Statistical Computing Group at Princeton, and worked at Los Alamos laboratory. She programmed the MANIAC I and ENIAC and coded the first monte carlo simulation.

The Lost Women of Science podcast devoted an entire season to her, I've been meaning to get around to finishing it: lostwomenofscience.org/season-

@ljrk Also shout out to Klára Dán von Neumann (John von Neumann's wife) also considered to be one of the first programmers:
en.wikipedia.org/wiki/Kl%C3%A1

She was Head of the Statistical Computing Group at Princeton, and worked at Los Alamos laboratory. She programmed the MANIAC I and ENIAC and coded the first monte carlo simulation.

15 June at 7:02 | Open on social.coop
Ulrich Junker

@ljrk I don’t know Tomasulo, but the book by Carver Mead and Lynn Conway was so important that it was used as the main reference for chip design in the CS courses that I took in the 1980ies,

15 June at 9:07 | Open on fediscience.org
lj·rk

@UlrichJunker Jup, they couldn't erase her from the title of the book. And yet, Mead was described as the Valley's founding father, Lynn not mentioned. Nowadays it's only Tomasulo (whose algorithm is an alternative to scoreboarding) who's remembered.

15 June at 9:42 | Open on todon.eu
maybeanerd

@ljrk it's funny you mention Conway, but not the "conway effect" she coined. This effect describes exactly what you're talking about: specific groups of people (often women) just "being forgotten".

Random first link I found on this: community.cadence.com/cadence_

15 June at 9:22 | Open on bumscode.com
lj·rk

@maybeanerd I... was not aware of that paper of hers. It's only briefly mentioned in the penultimate paragraph of "Legacy" on her Wikipedia page.

15 June at 9:32 | Open on todon.eu
maybeanerd

@ljrk it was also surprisingly difficult to find an example for my previous post on google as "conway effect" also returns mostly "conways law" results 😓

15 June at 10:30 | Open on bumscode.com
Thomas Fricke (he/him)

@maybeanerd @ljrk

Reading ieeexplore.ieee.org/stamp/stam

It is an eye opening article. We should have a Lynn Conway award for forgotten scientists.

15 June at 10:31 | Open on 23.social
John Seghers

@thomasfricke @maybeanerd @ljrk The above article by Lynn Conway is a remarkable story about her disappearance and about the beginnings of VLSI chip design. This was a story I did not know. Thank you for linking it.

15 June at 14:29 | Open on mastodon.social
Go Up