Allen Brady's Obituary

74 views
Skip to first unread message

Shawn Ligocki

unread,
May 19, 2024, 4:57:51 PMMay 19
to Busy Beaver Discuss
I am saddened to share the news that Allen H. Brady died last month on 21 Apr 2024. Brady discovered and proved the 4-state, 2-symbol Busy Beaver. His full obituary is available at https://www.rgj.com/obituaries/pnvs0805561 and mirrored at https://archive.ph/B0RRU .

On a personal note, Allen was a very welcoming and generous member of the Busy Beaver community. When my dad and I first announced our BB(2, 4) result in 2005, he shared his excitement and encouragement with us (even though we had just unseated his previous champion). Over the course of the following two years, we were in regular correspondence during which he shared his personal accounts of searches and techniques he had done as well as pointing us towards various resources and papers on TM simulation acceleration. These resources were invaluable to me in developing the earliest version of our own TM acceleration code. Thank you Allen, you will be missed.

nichol...@gmail.com

unread,
May 21, 2024, 2:59:27 PMMay 21
to Busy Beaver Discuss
In his dissertation, Brady mentions using an IBM 1620. He says it was "chosen primarily for its availability", which is pretty funny. Apparently the machine implemented addition by using lookup tables, i.e. stored precomputed values. This meant that it was possible to screw with a user's calculations by tampering with the lookup tables, I guess by setting 4 + 5 = 8 or something like that. And for the privilege of using such a janky machine, Brady had to drive 90 miles, because it was the only computer around.

Things were different in the old days!

More details about the hardware of early Busy Beaver researchers: https://nickdrozd.github.io/2021/12/08/busy-beaver-hardware.html

Terry Ligocki

unread,
May 29, 2024, 5:37:04 PMMay 29
to busy-beav...@googlegroups.com
I'm by this and agree that Allen was very welcoming to Shawn and I and our new (very non-canonical) results back in the day! I will miss him.

Allen predates me but I'm definitely solidly part of the OG. In primordial Fortran, everything was passed by address/pointer including constants AND, to save memory, a given constant, e.g. 2, was only stored once at one address, AND there was no notion of "const". So, if you passed 2 as an argument, say 'x', and set x to 3, then everyplace that used 2 would start using 3 instead?!?

Terry J. (Ligocki, tjli...@gmail.com)

--
You received this message because you are subscribed to the Google Groups "Busy Beaver Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to busy-beaver-dis...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/busy-beaver-discuss/4685d9e2-9c64-45c4-886e-569492abff0fn%40googlegroups.com.

Terry Ligocki

unread,
May 30, 2024, 1:31:27 AMMay 30
to busy-beav...@googlegroups.com
"I'm saddened by this..."
Reply all
Reply to author
Forward
0 new messages