Scored some great books from the department free table!
The Icon Programming Language: Icon is a direct influence of jq and I’ve been wanting to learn more
The C++ Programming Language by Bjarne Stroustrup
A Discipline of Programming by Edsger Dijkstra
I never saw Patrick Stewart at a convention and didn't come with him into direct contact in any way when working on #StarTrekPicard but there's one Patrick Stewart/Jean-Luc Picard related achievement that only I have ever made - find proof that Patrick Stewart exists in the #StarTrek universe. 1/10
On this day in 1969 ARPANET went live for the first time. It would later evolve into the Internet.
"On October 29, 1969, at 10:30pm Pacific Time, the first two letters were transmitted over ARPANET. And then it crashed. "
My heart is heavy. Walter Koenig, our Ensign Pavel Chekov, passed away early today.
Walter and I met on the bridge of the Enterprise in the 1960s, and we stayed friends for nearly six decades. When Brad and I married in 2008, Walter stood beside us as our best man. Just last month, on Sept. 14, he turned 90, on the same day Brad and I celebrated our 18th wedding anniversary. I will always treasure that he and I shared that date.
Walter was more than Chekov. He was a writer, a teacher, a thoughtful man, and a loyal friend. To the fans who loved him for so many years, thank you. He loved you too.
Rest well, my dear friend. Keep navigating a steady course.
– George
You've heard of Y2K and the 2038 problem, but what about the 1975 problem?
The UNIX form letter generator converts the current date to a string, but it only handles 1974. Why isn't it more general?
Well, the form utility was written around 1971 when UNIX used a different unit for time, jiffies (60ths of a second) since the current year instead of seconds since 1970. This 60× precision caused times to overflow a signed 32-bit integer every 13.6 months, so the epoch was bumped yearly. This yearly ritual of modifying the kernel meant it wasn't worth generalizing to support any year.
An earlier version of form uses jiffies since a 1972 epoch. The version from UNIX V4–V6 does the bare minimum to update it to the new system of seconds since 1970, while keeping the yearly overflow.
What are some games that never redraw the screen—only draw additively—and can be represented as text or ASCII art?
This is the class of games that can be played on a teletype, besides the obvious text adventures. With reverse line feed like the Teletype Model 37, a game board can be filled in, play by play.
* Minesweeper
* Picross
* Sudoku
* Voltorb Flip
* crossword
* regular expression crossword
* virus war / Paper Tactics
* tic-tac-toe
* hangman
This, and every reply is worth a boost.
As someone who worked in 20th century software preservation for twenty years, I am very tired of having to justify what I've done to institutional donors.
There were two institutions created back in the 1970's, The Computer Museum, who collected iron, and the Charles Babbage Institute, that collected paper.
NEITHER collected software.
"what have we forgotten from computer history that we can apply today?"
What the entertainment industry taught us. If you can't re-sell "intellectual property" in perpetuity, it isn't worth saving.
Picked up a neat book from the department free table today:
RE: https://machteburch.social/@netz39/116568278170901074
I should build a word clock! I saw a similar design in Zürich Altstadt and wanted one for some years, but I never got one, because they're so expensive. Now I could just lasercut and program one for cheap!
@mwichary Wow, this is quite the office, and the orange carpet is perfect! The CDC 6600's display is huge! Lovely photos.
@ronald I wonder whether the UNIX kernel from March 1972 supported EBCDIC translation? Let's take a look at its tty logic.
stty and gtty don't interpret their structures, just store them in rcsr (dcrsr), tcsr (dcpsr), and tty+4 (mode).
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u1.s#L476-L534
There were two kinds of ttys supported: the main console over KL-11 (tty) and up to 8 terminals over DC-11 (tty0–tty7). The inode read/write/open/close routines dispatch to rtty/wtty/otty/ctty if the device file is the main console or rcvt/xmtt/ocvt/ccvt for other terminals.
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u6.s
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u7.s#L192-L358
The main console is special. It's always connected and was apparently assumed to be a Teletype Model 33. Its input interrupt routine unconditionally normalizes to lowercase and handles the interrupt character (} or DEL).
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u4.s#L70-L109
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u6.s#L42-L181
The other terminals respected stty configuration and apparently the Teletype Models 33 and 37 and GE Terminet 300 were supported. All of the modes documented in V2, except for 001 (no delay for CRTs), are implemented. It is evidently between V1 and V2, as it supports 002 (tab expansion), which is not documented in V1. The high-order mode byte, which controls the IBM 2741 logic concerning chball in V2, is completely unused.
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u9.s
http://squoze.net/UNIX/v1man/man2/stty
http://squoze.net/UNIX/v2man/man2/stty
The canonicalization routine gets a line of characters and processes erase and kill. It is used on read for both kinds of terminals. After the C port in V4, this routine is where case normalization and Teletype Model 33 character translation was handled, but in 1972, the drivers weren't modular and this logic was specific to the devices.
https://github.com/dspinellis/unix-history-repo/blob/Research-V1/u7.s#L3-L33
Thus we can see that EBCDIC translation for the IBM 2741 was added to UNIX sometime between 17 March and 12 June 1972.
Earlier this year, I wrote a related piece about teletypes and their role in UNIX, which you may find interesting:
https://thalia.dev/blog/unix-braces/
Porting this kernel to C would be a fun project. It would also make it easier to see the architectural changes that were developed during the C port for V4.
Throughout this process, I fixed about a dozen bugs in the original assembly program and many other small issues. It seems it was less tested than the assembler, my last porting project, which had impressively few bugs in its 3,583 lines of assembly.
6/6
My process of reverse engineering this was quite methodical.
Since it is hand-written assembly, it employs numerous clever tricks, global program knowledge, and ad hoc calling conventions for every function.
Many functions have straightforward dataflow and a clear calling convention. Some require examining the demanded registers across all callers, to determine whether a register is written or clobbered.
But a nightmarish trio of functions are mutually recursive. To correctly analyze these, I annotated points with the register and stack state. Loops merge states on backedges, branches merge states on join. So far so good; this logic remains relatively local.
At recursive calls, I'd initially assume the call would update all state and I'd represent this with fresh values for all registers. This lets me do a dataflow analysis as usual. But some of these fresh values were observed, so I couldn't stop yet, as a full fixpoint was required. I repeatedly propagated constraints between the functions until the calling conventions were fully resolved.
Yes, I did this by hand. Yes, I'm crazy.
I'm working on a decompiler for PDP-11 assembly to period-appropriate C, but I haven't finished it yet, so I instead played as the static analyzer.
5/6
But why does form/fed implement an allocator? Now, we'd just load all the key/value pairs into memory, process, and dump, especially for a database so small. But memory was expensive and the PDP-11/20 had only 56KB.
Instead, it uses a file as the backing store and reads 512-byte blocks on demand, caching up to four of them, and flushing when evicted. It implements a buddy allocator with a full suite of functions, including coalescing and resizing. This was well before the standard library started to take form, so it was all bespoke to form/fed.
4/6
Now that I understand it fully, I'll build a forensic analysis tool in Rust to recover prior versions of the database by reversing allocator operations.
Much information is leaked through the slack data in freed blocks, the string copying patterns, and the ordering of the free lists. I hope to be able to construct a tree of possible operations that could have produced a form letter database.
With this, I may be able to refine my estimate of when UNIX was requested by the U.
3/6
Most notably, the oldest list of UNIX licensees was produced from a form letter database.
Ken Thompson used form/fed to simplify mailing to early users: rather than hand- or typewriting messages, this fancy tool would substitute each recipient and the date. In 1975, when UNIX News was starting up, he used this form letter database to dump the list of contacts that served as the initial subscriber list.
The circumstances of the UNIX V4 tape found at the University of Utah are still somewhat unclear, as it was apparently not used for research and only sparse traces of its acquisition survived. But some letters to Martin Newell are in that form letter database and I hope to recover more information through forensic analysis.
2/6
Whew! I just finished porting 2,852 lines of PDP-11 assembly to C by hand over the past week!
This is for form/fed, a form letter editor and generator with template substitutions, introduced in 1971. It's the largest UNIX utility that was never ported to C, and has since been forgotten.
1/6
I like finding and fixing bugs in 52-year-old software, especially when all later versions retain the bug.
RIP Margaret Hamilton
In memory and honor of Margaret Hamilton (1936–2026), American hero. https://news.mit.edu/2026/margaret-hamilton-computing-pioneer-dies-1007