2024-09-11 00:26:50 <unjust      > KipIngram: do you mean there just aren't enough of us demanding features that the average user doesn't need?
2024-09-11 00:34:05 <thrig       > cpu used to be aimed at humans, now cpu are aimed at compiler suites
2024-09-11 03:46:32 <KipIngram   > unjust: I just meant that the vast majority of customers wouldn't know what to do with a hardware monitor trigger if one stepped on their toe.
2024-09-11 04:03:08 <KipIngram   > When faced with the decision of whether to invest the development in including the feature, they're going to ask "how much will it increase our sales?"
2024-09-11 04:21:48 <unjust      > that's fair enough, they're businesses after all
2024-09-11 04:26:45 <unjust      > i know GPIO access is certainly not the norm in modern PCs, but at least some common SIO controllers must have a few unrouted (or just associated with unneeded peripherals) that could be picked off
2024-09-11 04:35:23 <xentrac     > back in the 90s you could use a parallel port
2024-09-11 04:36:13 <thrig       > there are things like https://en.wikipedia.org/wiki/Bus_Pirate
2024-09-11 04:36:30 <xentrac     > some researchers at Princeton were annoyed at how slow their Linux cluster was at doing barrier synchronization (where all the nodes finish computing up to a "barrier" before any of them continue beyond it) over Ethernet.  or maybe it was Myrinet
2024-09-11 04:36:35 <unjust      > that's true, even around 2010 some mainboards still had parallel port functionallity via the super i/o controller, but not routed
2024-09-11 04:37:11 <xentrac     > so they wired up a line from each machine's parallel port to a tree of NAND gates and wired the output of the tree back to all the machines
2024-09-11 04:37:26 <xentrac     > they called it PAPERS, which is unfortunately a terrible search term
2024-09-11 04:37:32 <xentrac     > well, AND gates, really
2024-09-11 04:37:53 <unjust      > a bunch of NANDs making ANDs?
2024-09-11 04:38:01 <xentrac     > yeah
2024-09-11 04:38:11 <xentrac     > that way they could get submicrosecond barrier synchronization, which was not something their networking stack was able to deliver
2024-09-11 04:40:57 <unjust      > thrig: those are definitely useful, though being reliant on USB for comms is a bit of a downer
2024-09-11 04:44:22 <unjust      > xentrac: this one? http://aggregate.org/TechPub/icpp96.pdf
2024-09-11 04:44:45 <xentrac     > yes!  sorry, I misremembered that it was Princeton rather than Purdue
2024-09-11 04:44:48 <xentrac     > thanks!
2024-09-11 04:46:35 <xentrac     > it's kind of amazing they were using TTL (74LS* and 74F*) rather than CMOS in 01996
2024-09-11 04:49:33 <unjust      > may have been easier to interface directly with the parallel port without level shifting at the time
2024-09-11 04:49:55 <xentrac     > dunno, do parallel ports use TTL levels?
2024-09-11 04:50:28 <unjust      > i think so
2024-09-11 04:50:45 <unjust      > i think the modern usb-to-printer adapters might be more CMOS-friendly though
2024-09-11 04:55:24 <thrig       > but the bus pirate folks are probably mostly selling to folks who have laptops with usb...
2024-09-11 04:57:46 <unjust      > this isn't relevant, but this was lurking in the results while searching for a copy of ieee 1284: http://ps-2.kev009.com/eprmhtml/index.htm
2024-09-11 04:58:29 <unjust      > thrig: that's true
2024-09-11 05:01:15 <unjust      > "IEEE Standard Signaling Methodhttps://kazus.ru/nuke/modules/Downloads/pub/148/0/IEEE%201284-2000.pdf
2024-09-11 05:01:18 <unjust      > for a Bidirectional Parallel Peripheral
2024-09-11 05:01:25 <unjust      > ^ https://kazus.ru/nuke/modules/Downloads/pub/148/0/IEEE%201284-2000.pdf
2024-09-11 05:02:24 <unjust      > Page 69: "Drivers shall operate at nominal 5 V TTL levels."
2024-09-11 12:22:03 <rendar      > why `1e0 fsin` gives me an error?
2024-09-11 12:23:46 <GeDaMo      > Which Forth? What's the error?
2024-09-11 12:24:49 <GeDaMo      > Did you try to print the result with . ? Use f. for floating point values
2024-09-11 12:27:02 <rendar      > gforth
2024-09-11 12:27:43 <rendar      > GeDaMo, oh right, that was the error, . instead of f.
2024-09-11 12:28:16 <GeDaMo      > :)
2024-09-11 14:12:08 <xentrac     > unjust: aha, thanks!
2024-09-11 14:17:34 <unjust      > the ibm nostalgia host also contained this: a reference to a real time UNIX from AT&T (Lucent?) and a phone switch that became eWaste -> http://kev009.com/wp/2024/07/Lucent-5ESS-Rescue/
2024-09-11 20:10:30 <KipIngram   > IBM nostalgia site?
2024-09-11 20:12:27 <KipIngram   > I started with IBM in 2012.  Turns out 2011 had been the companies "centennial year."  They made a video around that.  Honestly, it was a pretty damn impressive company up through the early 1980's.  Then their failure to keep hold of the PC market, and the explosion of said market, nearly did them in, and it took some fast footwork (transitioning from a hardware model to a "services" model) to keep
2024-09-11 20:12:29 <KipIngram   > them alive.
2024-09-11 20:12:56 <KipIngram   > They expected the IBM PC to sell MAYBE 200k units.
2024-09-11 20:13:13 <KipIngram   > I bet they would have handled it differently if they'd had a more accurate idea of what was coming.
2024-09-11 20:14:14 <KipIngram   > The TV show Halt and Catch Fire paints a pretty interesting picture of that whole business too.
2024-09-11 20:14:31 <KipIngram   > I really enjoyed the first couple of seasons of that, while it was dealing with stuff more up my alley.
2024-09-11 20:14:50 <KipIngram   > Then the show evolved, as the world did, into gaming and web apps and so on, and I kind of lost interest.
2024-09-11 20:54:03 <unjust      > KipIngram: it was in reference to this page with a late 90s design: http://ps-2.kev009.com/eprmhtml/index.htm (Electronic Pocket Reference Manual IBM - AUSTRIA PC-HW Support)
2024-09-11 21:54:06 <unjust      > veltas: working on anything interesting?
2024-09-11 22:00:21 <veltas      > I was thinking about a block-based protocol
2024-09-11 22:00:41 <veltas      > Over UDP
2024-09-11 22:00:48 <veltas      > blocknet
2024-09-11 22:01:05 <veltas      > Learning x86 encoding
2024-09-11 22:01:39 <veltas      > Also learned Go today at work, might write something for work in that, I'm impressed so far by it, it actually reminds me of NewB a lot
2024-09-11 22:01:48 <veltas      > Probably because Thompson was one of the co-designers
2024-09-11 22:02:12 <veltas      > I enjoyed the B throwback of Args being a global rather than arguments to main
2024-09-11 22:02:41 <veltas      > What's everyone working on?  What about you unjust?
2024-09-11 22:04:01 <KipIngram   > A guy I used to chat with who I felt was particularly "CS knowledgeable" also thought very highly of Go.
2024-09-11 22:04:15 <KipIngram   > Sounds like that got a lot of things right with it.
2024-09-11 22:06:29 <unjust      > veltas: been adding a few mnemonics to my amd64 machine code macros a couple of times a day recently, but less interestingly and more time consuming: working with something old and riddled with technical debt to sample contact closures over i2c
2024-09-11 22:08:00 <unjust      > i'm not familiar with Go, what are you planning to write in it?
2024-09-11 22:20:52 <veltas      > Proprietary ;)
2024-09-11 22:23:10 <unjust      > heh
2024-09-11 22:24:49 <veltas      > Less interesting but that's the bread and butter of Real Programming(TM)
2024-09-11 22:26:13 <unjust      > was the tutorial on go.dev good enough, or did you find something better?
2024-09-11 22:27:29 <veltas      > gobyexample.com and then just tried some stuff out / googling / blogs / reference manual
2024-09-11 22:27:43 <veltas      > I've only started but it's not a million miles away from other languages I've used
2024-09-11 22:28:24 <veltas      > I couldn't find a tutorial on go.dev, is there one?
2024-09-11 22:29:24 <veltas      > Huh it's linked at top of gobyexample.com, but I didn't see it on go.dev itself
2024-09-11 22:30:58 <veltas      > I don't understand new-style websites
2024-09-11 22:38:34 <dlowe       > https://go.dev/tour/
2024-09-11 22:51:04 <veltas      > That's not the tutorial ... is it(?)
2024-09-11 22:52:03 <veltas      > I already got directed to another introduction by go.dev so don't fancy readinga second, third, etc
2024-09-11 22:52:23 <unjust      > it's this one i was refering to: https://go.dev/doc/tutorial/getting-started
2024-09-11 22:53:16 <veltas      > I think really they should pick one
2024-09-11 23:08:04 <xentrac     > the way IBM tells it, nobody outside the company invented any computer stuff before the early 01980s
2024-09-11 23:08:38 <xentrac     > but that's not what actually happened of course
2024-09-11 23:09:29 <unjust      > i heard they were purveyors of fine abacuses since well into the stone age
2024-09-11 23:10:57 <xentrac     > they had become a major drag on progress by 01970 because of their internal politics
2024-09-11 23:12:29 <xentrac     > they were, for example, struggling to prevent the adoption of ASCII, which had been originally developed by an IBMer
2024-09-11 23:13:33 <xentrac     > also, for most of the 01970s, relational databases, though Codd stayed at IBM during that time instead of getting relegated to Siberia for inventing things that could decrease IBM's sales
2024-09-11 23:14:03 <veltas      > Dunno about stone age but I think an Ur IBM regional purchaser signed one of the Ea-nāṣir complaint tablets
2024-09-11 23:16:04 <unjust      > COBOL in its infancy: https://medium.com/tech-is-a-tool/the-dawn-of-computing-sumerian-abacus-83bdefb697ba