🇺🇦 haxadecimal 🚫👑

Proud to be Antifa, like my grandfather, who was killed fighting fascism in WWII, and was posthumously awarded the Air Medal and the Purple Heart.
#microcontroller #firmware #C #CPlusPlus
#FPGA #VHDL #Verilog #SDR (SW Defined Radio)
#retrocomputing
#nonpareil HP calculator simulation at #microcode level
Maker (with John Doran) of #NixieTube #RPN #calculator
CHM PDP-1 Restoration Team
cis male
atheist
Damned dirty ape
Not a tame programmer
Some assembly required
No user serviceable parts inside

pronouns: he/him
age: old enough to know better
religion: little endian, Emacs, Linux & BSD
amateur radio call sign: N2ES

Public Key

npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th

NIP-05 Address

brouhaha_at_mastodon.social@momostr.pink

Profile Code

nprofile1qqsgtr7wly5kyr6wyxme0xk3haf2rnprqkmpgylm2pqgpfxh0yu442cprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmnt5k67a4

Publishing to

relay.momostr.pink

Author Public Key

npub1sk8ua7ffvg85ugdhj7ddr06j58xzxpdkzsflk5zqszjdw7fet24slrh5th

Show more details

Published at

2026-04-22T03:12:36Z

Kind type

0 Metadata

Event JSON

{ "id": "6b29ba7e53aabf80f2fb0f0527ca121df29a8ecb473dfa4d3ba5ef4c549fb747", "pubkey": "858fcef929620f4e21b7979ad1bf52a1cc2305b61413fb504080a4d779395aab", "created_at": 1776827556, "kind": 0, "tags": [ [ "t", "c" ], [ "t", "retrocomputing" ], [ "t", "sdr" ], [ "t", "firmware" ], [ "t", "fpga" ], [ "t", "verilog" ], [ "t", "vhdl" ], [ "t", "cplusplus" ], [ "t", "nixietube" ], [ "t", "calculator" ], [ "t", "rpn" ], [ "t", "microcontroller" ], [ "t", "microcode" ], [ "t", "nonpareil" ], [ "proxy", "https://mastodon.social/users/brouhaha", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://mastodon.social/users/brouhaha", "pink.momostr" ], [ "-" ] ], "content": "{\"name\":\"🇺🇦 haxadecimal 🚫👑\",\"about\":\"Proud to be Antifa, like my grandfather, who was killed fighting fascism in WWII, and was posthumously awarded the Air Medal and the Purple Heart.\\n#microcontroller #firmware #C #CPlusPlus\\n#FPGA #VHDL #Verilog #SDR (SW Defined Radio)\\n#retrocomputing\\n#nonpareil HP calculator simulation at #microcode level\\nMaker (with John Doran) of #NixieTube #RPN #calculator\\nCHM PDP-1 Restoration Team\\ncis male\\natheist\\nDamned dirty ape\\nNot a tame programmer\\nSome assembly required\\nNo user serviceable parts inside\\n\\npronouns: he/him\\nage: old enough to know better\\nreligion: little endian, Emacs, Linux \u0026 BSD\\namateur radio call sign: N2ES\\n\",\"website\":\"https://mastodon.social/@brouhaha\",\"picture\":\"https://files.mastodon.social/accounts/avatars/000/027/729/original/95248bda6f9a51c7.jpg\",\"banner\":\"https://files.mastodon.social/accounts/headers/000/027/729/original/7fab8b351228fe8d.jpg\",\"nip05\":\"brouhaha_at_mastodon.social@momostr.pink\"}", "sig": "4ecda1ca291b41efbd378d7aebacc15d8c7b060f00caa5e216c34d5d50a9c88ea237ad4a769c82f752d8c29877f3a26cedf70255f41ca659ff0484fc215989a7" }

Last Notes

2026-04-07 04:51:45 UTC

Dear lazyweb,
Does anyone happen to know what company introduced the first TTL-compatible metal-can hybrid crystal oscillator in the DIP footprint (like a 14-pin DIP footprint, but with only the four corner pins), and when?
My web searches don't seem to yield relevant results. I'm not looking for the history of crystal oscillators in general. I think the type I've described appeared in the 1970s.
#hybrid #dip #crystal #oscillator

2026-04-03 20:11:26 UTC

If you want Cherry MX keyswitches for future use, you might want to stock up. Cherry announced last year that production was moving to China, but today Mouser Electronics sent me an obsolescence notice for Cherry MX Blue.
:-(
#MechanicalKeyboard #CherryMX #CherryMXBlue

2026-03-28 09:12:34 UTC

Dear Lazyweb,
Has anyone saved a copy of the M68060 software package (M68060SP), and its documentation? This includes exception handlers to provide compatibility for the 68K instructions and addressing modes that were omitted from the MC68060 processor.
#Motorola #MC68060 #M68060SP #MC68000

2026-03-23 19:35:24 UTC

- reply

I only have three days left to return it to Best Buy.

2026-03-23 19:32:45 UTC

- reply

I could return the device to Best Buy, but I'm not thrilled about doing that to them due to a f**kup on Google's part. But I guess they take that risk as a retailer of any product.

2026-03-23 19:22:47 UTC

- reply

They marked the case as "resolved" and closed the chat.
Maybe it's time to switch brands.

2026-03-23 19:17:32 UTC

- reply

They keep saying to be patient for this supposed "rollout", but won't commit to that happening within my 60 day from purchase window of eligibility to purchase the protection plan.

2026-03-23 19:16:00 UTC

- reply

They insist that they are a supervisor. I''ve now asked to be connected to someone with actual authority, though I don't expect that to happen.

2026-03-23 19:13:37 UTC

- reply

Yeah, some respones are word for word identical to the first "representative". Now I've asked to speak to their "supervisor", because I want more of the same.
#GluttonForPunishment

2026-03-23 18:56:48 UTC

- reply

Maybe a real person, but offered Ai-generated responses to send.

2026-03-23 18:54:40 UTC

- reply

Getting mixed vibes from "supervisor". Some responses seem human, some seem auto-generated.

2026-03-23 18:48:57 UTC

- reply

Google's web page for Pixel Care+ clearly states that devices purchased from Google's retail partners are eligible. It doesn't say "will be eligible when we gat around to it," or anything about a phased rollout.
When I put the serial number or IMEI into the form, it says not eligible.

2026-03-23 18:46:23 UTC

On chat with a Google "representative" (AI?) about why a device purchased new from Best Buy 11 days ago is not eligible for a Pixel Care+ protection plan. They/it keep telling me that Pixel Care+ is in a phased rollout and not yet available for my device, which is bullshit.
I've asked to speak to a supervisor. Will that be AI also?

2026-03-15 22:18:50 UTC

- reply

Yes, doing that requires manually rebuilding the repository, and effectively invalidates all client checkouts.
:-(

2026-03-11 19:24:45 UTC

- reply

People sometimes ask what my favorite programming language is. Probably Smalltalk, except that I prefer strong typing, and Smalltalk wants to be its own world, like FORTH, rather than being used to create "native" applications. I usually reply that I have no favorite, because all programming languages suck. Some suck less than others.

2026-03-11 19:21:06 UTC

- reply

I'm also feeling nostalgic for MPW Pascal, and the Macintosh Programmer's Workshop in general.
I was sad when Apple ditched Pascal for C.
Pascal was not perfect, by any means. I've programmed professionally in Ada, one of Pascal's successors, and liked it pretty well.
I wanted to use Modula 3, "an elegant weapon for a more civilized age," and dabbled in it a little, but there was no market for it.

2026-03-11 17:55:13 UTC

- reply

Improving LLMs will only make their responses slightly more answer-shaped.
People expect otherwise in part because they've seen so much technological advance in computers in general, and think that means more advanced LLMs will follow the usual progression. But mostly they expect otherwise because of the constant propaganda from Sam Altman and other supposed experts, telling them that AGI is right around the corner.

2026-03-11 17:50:09 UTC

- reply

Even if one ignores that, and somehow trains a new model on a pre-slop training set, LLMs still won't improve, in the sense that people expect. The main improvement everyone wants is for LLMs to have better understanding. That is impossible, as LLMs have NO understanding. A LLM doesn't answer a question or prompt. It generated a statistically plausible output of what an answer might look like. More CPU power, storage, or tweaks to algorithms can't fix that.

2026-03-11 17:46:26 UTC

- reply

The future of LLMs is that they suck worse than they do now, even if hardware gets faster, and regardless of whether they run locally or remotely.
The reason is that LLM generated slop is contaminating the only source of training data. That can only make future models worse.
There is literally no source of high-quality training data for new and improved models.

2026-03-11 05:31:20 UTC

For no particular reason, I am feeling nostalgic for Turbo Pascal.

2026-03-02 20:53:48 UTC

Erwin Schrödinger spent the rest of his life in a superposition of being and not being in jail for animal cruelty.

2026-02-26 06:51:21 UTC

CAUTION: EVENTS IN CALENDAR ARE SOONER THAN THEY APPEAR

2026-02-24 17:24:52 UTC

- reply

Model F is the general family and technology of a LOT of IBM capacitive keyboards made starting in 1980. The vast majority of model F keyboards are not the original IBM PC keyboard.
The original IBM PC keyboards is only identified by its part number, 1801449.

2026-02-24 17:07:34 UTC

I want an IBM PC (5150) keyboard for a project. Looking for genuine, not a clone. They are hard to come by, 45 years later.

2026-02-21 06:49:37 UTC

An optimist believes that we live in the best of all possible worlds.
A pessimist fears that that is true.

2026-02-20 16:54:05 UTC

Pizza. It's so much more than just a breakfast food.

2026-02-18 14:47:58 UTC

How did I go from "Any pillow will do" to "This pillow is too firm, that one is too soft, and the other one is 4mm too thick"?

2026-02-12 20:04:35 UTC

- reply

When I first read the Asimov story, I thought that it had to be set VERY far in the future. Now I wouldn't have any trouble believing that it was set in next week.
In the story, once he's proven his method, it generates a lot of interest, because computers are expensive. It is not surprising that in 1958, Asimov, and almost all other contemporary SF writers, did not anticipate integrated circuits and Moore's Law. It was common knowledge that computers would keep getting bigger.

2026-02-12 19:56:59 UTC

- reply

Reminds me of a different story, Asimov's "The Feeling of Power", in which a future mathematician discovers that it is possible to do arithmetic without a computer, for instance, using pencil and paper. He hypothesizes that at some point in the past, humans had to do arithmetic that way, because someone had to have programmed the computers to do it. At first, no one believes him, and they all consider him to be a crank. Everyone knows that you need a computer for arithmetic!

2026-02-12 10:39:54 UTC

- reply

I can hardly wait!
I just finished reading _Enshittification_, after going to your book signing at The Tattered Cover in Denver.
Previously, I had some inklings of why things were getting so awful, but you've done an excellent job of connecting the dots and explaining the many forces at play.
I hope that, as you've recommended, other countries repeal the DMCA equivalents the USTR forced upon them. Eventually, perhaps that could even lead to the US repealing it.

2026-02-03 06:49:21 UTC

- reply

I hadn't heard of dmsetup messaging before, and even after reading about it, don't feel like I understand the concept. What's the point of "sending a message" to a target? Is it something asynchronous? How does the target receive messages? Does it send back a reply? I'm just used to issuing commands directly.

2026-01-29 21:59:01 UTC

- reply

Why don't they just spend the $130 million on hookers and blow, instead? They'd at least have a good time, without fucking up the browser for users.

2026-01-28 01:45:17 UTC

- reply

To shoot, more likely. Preemptively prevent them from ever becoming illegal aliens.

2026-01-28 01:39:49 UTC

ICE is no longer satisfied with terrorizing people in the USA. Now they've decided to invade other countries.
https://www.timesnownews.com/world/us/us-news/watch-ice-agents-attempt-to-enter-ecuadorian-consulate-in-minneapolis-triggering-geopolitical-row-article-153517215

2026-01-23 10:12:12 UTC

My new plan for next week is to steal $100 million, and self-deport to South America. So long, suckers!

2026-01-22 05:46:42 UTC

- reply

Google says:
High-Purity Powder (99.5%+): $89–$138 for 10g, $294–$347 for 50g.
Industrial Grade: As low as $3/kg.
So will industrial grade provide that benefit, or does it have to be high purity?

2026-01-22 05:44:10 UTC

- reply

Wow! Hadn't heard of it before. Amazing stuff. But what does it cost?

2026-01-20 02:21:21 UTC

declaring marital law

2026-01-17 06:51:43 UTC

I accidentally mixed up the words jacuzzi and Yakuza, and now I'm in hot water with the Japanese Mafia.

2026-01-17 02:15:47 UTC

- reply

And don't get me started about how "software engineering" doesn't actually exist, and to the limited extent that it can be said to exist, isn't actually done in practice.
(There are, of course, a _very_ small number of exceptions.)

2026-01-17 02:14:54 UTC

- reply

I actually make fun of software developers far more than EEs. As a software developer, I feel entitled to do that.

2026-01-17 02:13:11 UTC

I got an EE at a place I worked a few years ago really angry with me, when I said my usual thing about "EE is trivial, you just wire together all the pins that have the same name." He must have known I was joking, but still reacted badly.

2026-01-15 06:17:43 UTC

- reply

Starship Troopers

2026-01-06 00:44:10 UTC

- reply

COBOL isn't particularly difficult. Just annoying.

2026-01-05 10:00:54 UTC

- reply

Tektronix includes in the 11301/11302 manual, under "physical characteristics", BOTH the weight and mass of the scope, in appropriate units: weight in Newtons and pounds, and mass in kilograms and slugs. I don't recall seeing this fone in any other equipment manual; most such manuals incorrect conflate weight with mass.
Surprisingly, however, this manual fails to state the conditions under which the specified weights are relevant.
https://files.mastodon.social/media_attachments/files/115/841/884/610/675/253/original/6156ea2fcbbaf0cd.jpg

2026-01-05 09:19:44 UTC

- reply

Most of the time, my Rigol 350 MHz MSO5354 (upgraded 5074) will still be my workhorse, and of course it's much smaller, lighter, and more portable, and I can get digital screen captures from it. There are rare times that I've needed more than 350 MHz bandwidth, so now I have that covered. I'll have to get some 500 MHz probes.

2026-01-05 09:15:18 UTC

Also tonight I picked up the Tektronix 11302 analog storage oscilloscope a friend offered me a while back. It's a 500 MHz mainframe with three slots for 11 series plugins. But it's not just any old analog storage scope. This has a microchannel plate (like the Tek 2467 portable 250 MHz), and has IEEE-488 and RS-232 ports for control. It also has an amazing self-calibration system, all the way to the display graticule.

2026-01-05 08:34:17 UTC

Friends and I just watched Star Trek VI: The Undiscovered Country tonight. While I still consider The Wrath of Khan to be the best of the series, IMHO VI is very good, and generally underappreciated.

2026-01-03 01:57:52 UTC

Someone asked me by email to make sure I package an item securely in "bubble warp". I didn't think warp technology was going to be invented until the 22nd century.

2025-12-28 08:29:16 UTC

- reply

So-called "AI" is just like the internet:
Garbage In, Garbage Out.
At least the name of the so-called "AI" is usually half-right, though there have been some notable exceptions even to that.
Sturgeon was an optimist.

2025-12-20 05:09:29 UTC

- reply

Many RISC processors don't have a hardware stack, and depend on software to maintain one if desired, which of course it almost always is.
On an interrupt or other exception, they capture just the content of only a few registers (including the PC) into registers reserved for that purpose, and then the handler software can push them onto a stack.
(I'm not advocating that you do this.)

2025-12-19 23:49:43 UTC

- reply

I'm not criticizing. I understand the fun of designing something minimal.
If you had interrupts, then you'd need data paths to store X and Y, possibly by transferring them to A.

2025-12-19 22:36:50 UTC

- reply

No interrupts, I assume?

2025-12-17 03:22:22 UTC

- reply

Google appears to have a lot of money. Maybe we could convince them to fund an open-source privacy-respecting browser.
@npub18ak…untn

2025-12-15 22:03:30 UTC

- reply

That was the point of the poll. I am writing an assembler for an architecture for which there is no (publicly) documented assembly language.

2025-12-12 04:21:10 UTC

preferred assembly language operand order

2025-12-11 02:33:29 UTC

[@mos_8502](https://studio8502.ca/@mos_8502 )
We've had tri-state buffers since the 1970s. Isn't it about time for quad-state buffers?

2025-12-10 01:06:27 UTC

In my ongoing research and development of computer peripherals, I have some improved types of printers, for which I am preparing patent applications:
* dot-vector printer (saves space!)
* daisy cutter printer
* chain saw printer
* commuter train printer
* pulse jet printer

2025-12-08 23:57:33 UTC

- reply

Money sent!
:-)
Many years ago, I really could have used a means to have like-minded people see what I was doing, and maybe offer a little support for my projects. That wasn't really possible then, and it's not like I have a ton of money now, but I like to occasionally support people working on interesting projects.

2025-12-05 00:08:16 UTC

- reply

The spikes are essentially a deciding race. I'll study it more. In the mean time, try inverting the clock into the FFs, but not the AND gates, and see if that's better.
The spikes also might be a result of simulated devices not having real propagation delay, but I wouldn't like to count on that.

2025-12-04 22:26:00 UTC

- reply

Ignore the top two FF in the '362 diagram; they're an extra divide-by-four for its crystal oscillator, unrelated to the divide-by-four used for the four-phase clock generator.

2025-12-04 22:22:43 UTC

- reply

In fact, the same thing can be done with the '74 FF. See the TI datsheet for the 74LS362(TIM9904), functional block diagram.
And obviously you can use a three input NAND (74x10) instead of AND, if you want active low phase output.

2025-12-04 22:16:44 UTC

- reply

Here's an untested idea for a three-chip TTL solution with no startup issues and needing no power-on reset, using a '109 dual J-/K edge triggered FF, and two '11 three-input AND gates. Being an off-the-cuff idea, it definitely needs to be simulated.
I've always liked the '109 because it is more flexible than the '74, but if you tie J and /K together, it behaves exactly like the '74.
https://files.mastodon.social/media_attachments/files/115/663/591/662/154/163/original/50cdc717b692ab9a.jpg

2025-12-04 21:51:02 UTC

- reply

On further study, it looks like the design is prone to startup problems in general, because flip-flops power up in random states in real life, but not in simulation. But even in simulation, inverting /Q should have the same result as using Q.
If there is apower on clear to all four of the flops in the ring, then you'd want feedback from /Q _directly_ to D of the first one, as otherwise you'd just cycle the same value continuously.

2025-12-04 21:47:21 UTC

- reply

The obvious drawback is that the PIC needs to be programmed, but iit's one chip instead of five.

2025-12-04 21:46:03 UTC

- reply

To a casual glance, that looks perfectly fine, though I wonder whether you really need the inverter from /Q of the right flop rather than just using Q.
The last time I needed a four-phase non-overlapping clock, to drive a Western Digital MCP1600 series chips (e.g., LSI-11, Pascal Microengine), I used a small PIC microcontroller, which is quite inexpensive and available in a 14-pin DIP.
https://github.com/brouhaha/4phaseclk
https://www.flickr.com/photos/_brouhaha_/albums/72157661646247410/
1/

2025-12-04 04:43:28 UTC

- reply

And all bets are off with virttual inheritance. Even P0149r3 disclaims that.

2025-12-04 04:42:50 UTC

I found and fixed my most recent C++ issue. It wasn't in the code I was working on at all; that just exposed an extremely stupid thing I'd done elsewhere.
Up-casting member pointers to a base type that doesn't have the member in question seems to work fine in GCC, although I think it might officially be UB, and that might be changed to not be UB in C++26 (P0149r3).
Obviously I can only actually dereference the member pointer with the correct object type.

2025-12-03 21:15:12 UTC

But I like MATE! I don't want to switch!
https://files.mastodon.social/media_attachments/files/115/657/700/580/793/829/original/cb1dc24fde3fed32.jpg

2025-11-29 09:46:31 UTC

- reply

Bit flip errors are the primary reason for use of ECC for on-die caches on high-end CPUs (i.e., x86, ARM 64-bit, but not small microcontrollers), and in DRAM memory systems in servers. Bit flips can occur in DRAM surprisingly frequently, because bits are stored by very few electrons.
I use ECC memory even in my desktop and laptop for that reason.

2025-11-29 09:42:13 UTC

- reply

However, radiation of various sorts can cause register and RAM bit flips, or signal errors, without causing CMOS latchup.

2025-11-29 09:40:16 UTC

- reply

As I understand it, CMOS latchup is caused by triggering parasitic SCR structures (PNPN), and that standard CMOS process design rules make it extremely unlikely to be internally triggered under normal operating conditions. An SEU can inject enough energy to trigger a parasitic SCR.
SOS and SOI CMOS fabrication eliminate the parasitic SCRs, but are typically only used for mil/aerospace due to expense.

2025-11-28 08:12:50 UTC

- reply

That a policy or plan can't possibly result in their stated objectives has never stopped the Republican Party, or their supporters, from anything, no matter how misguided.

2025-11-28 03:24:07 UTC

My oscilloscope seems to be getting closer to the edge of the bench every time I use it. I don't want it to fall, so I moved it back in a few inches. Maybe the bench isn't quite level? I've not before had firsthand experience of scope creep.

2025-11-26 23:31:39 UTC

"Soon AI agents will be writing better, cleaner code than any mere human can, just like compilers can write better assembly."
— Nick Hodges, in Infoworld, 2025-11-26
https://www.infoworld.com/article/4096265/writing-code-is-so-over.html
I look forward to a contined consulting revenue xtream, fixing the disasters that result from managers believing this bulles***.
He likens this to compileds writing better assembly code than humans. That works because the semantic level of high level languages like C, C++, etc. are not that far removed
1/

2025-11-23 00:25:51 UTC

- reply

SSDs are rated to retain data unpowered for six months for consumer grade devices, and for three months for enterprise grade. They need to stay powered up so that they can scrub (refresh) memory cells that lose their charge over time. This is because each data bit is actually stored by a surprisingly small number of electrons in modern NAND flash devices.
Spinning drives, despite their potential mechanical failure modes, are actually better for cold storage than SSDs.

2025-11-18 09:25:43 UTC

- reply

Some of us care that our browser is open source, in other words, not loaded up with vendor spyware/malware. It might be reasonable to run the open-source Chromium, but not actual closed-source Google Chrome.

2025-11-16 10:33:22 UTC

- reply

Sure, I don't mind the occasional tech question. Unfortunately, It's hard to know what's gone wrong. Since it's both HDMI and network, normally unrelated functions, that suggests that it might be a software failure, and that reinstalling the device drivers, or the whole OS, might fix it.
Or it could be corrupted BIOS settings.
On the other hand, depending on how the machine is built, there might be e.g. a bad connection to a board responsible for both.
Sorry I'm not more helpful.

2025-11-16 10:01:16 UTC

- reply

Feel better!

2025-11-16 10:00:50 UTC

- reply

Something I've had to wait three months for:
* EMG, hoping to determine cause of right shoulder and arm pain
Otherwise, more of the usual:
* work (FPGA stuff in VHDL, somewhat fun)
* programming my own projects, including some new cross-development tools for obscure obsolete microcontrollers
* reverse-engineering 1970s and 1980s calculators
* watching more episodes of Science Fiction Theatre (micd-1950s SF anthology show)

2025-11-09 02:06:30 UTC

- reply

Is there a memory expansion header? I think 512MiB is fine, if a daughterboard can add more.

2025-10-28 07:48:09 UTC

Germany is now advising people to stock up on cheese and sausages.
This is called the Wurst Käse Scenario.

2025-10-26 22:33:12 UTC

- reply

The PDP-11/20 could have been installed in racks with other equipment, but the PDP-11/20 processor and main memory (no peripherals) is a box 10.5" x 19" x 24" and weighing under 100 lb. The IBM 1401 processor and main memory (no peripherals) was 61" x 58" x 31" and weighed 2199 lb, so it is 23x the volume of the PDP-11/20 and around 20x the weight.
Not a very similar size at all.

2025-10-25 09:54:15 UTC

- reply

No, DHGR _almost_ fell out automatically from the basic design for 80 column text, just as double lores did. Apparently they realized very soon after the start of production, and before first customer shipment, that a very minor motherboard change would enable it. I am 99.9% certain that no ROM change was required, because the ROM doesn't even know whether it's in double mode.

2025-10-25 09:44:30 UTC

- reply

If it works on an enhanced IIe, it should work on unenhanced IIe. Nothing in the ROMs is required to make DHGR mode work. All IIe variants with 128K RAM can support DHGR, except for a very small number of unenhanced units that shipped with a very early revision motherboard.

2025-10-04 01:49:49 UTC

I'm thinking about putting everything on a bagel.

2025-09-21 07:09:07 UTC

- reply

Was common on "big computers", and very handy.
Of course, the 1401 is big only relative to microcomputers and small minicomputers, all of which postdate it.

2025-09-15 21:24:25 UTC

Humans (Homo Sapiens) are considered to be in the taxonomic order of primates, and family of hominids (great apes).
Lately I've come to believe that this is incorrect, and that the primates order should include another family, the not so great apes.

2025-09-15 09:39:16 UTC

- reply

Give a man a gun, and he'll rob a bank.
Give a man a bank, and he'll rob the world.

2025-09-13 20:46:10 UTC

- reply

That worked because it wasn't running any graphical interface. Just serial terminals. A current Linux kernel, even if you configured out absolutely everything not needed for that, probably wouldn't fit in 1 MiB, though I haven't tried. However, even stripped to a minimum configuration, a current Linux kernel provides a lot of features that BSD 4.2 did not.

2025-09-13 20:28:51 UTC

- reply

BSD 4.2 ran fairly well in 1 MiB on the VAX 11/780.

2025-09-13 20:25:21 UTC

I was suspended from Twitter early in the pandemic for replying to someone else's tweet that gave misinformation about hydroxychloroquine as a COVID treatment. I wrote something like: "Hydroxychloroquine is just as effective a COVID treatment as ivermectin. In other words, neither are effective at all." Even on appeal, Twitter refused to reinstate my account until I agreed to delete the tweet.

2025-09-13 03:52:19 UTC

- reply

That's why my hobby is #retrocomputing. Sure, the old computers from the 1960s and 1970s were thousands of times slower, and had a millionth as much memory, but they were, and still are, fun!

2025-09-11 04:52:23 UTC

- reply

And then deleted? I tried to reply to the edited post, but my reply failed,.

2025-09-02 20:23:04 UTC

I don't have an SR-71, but I do have an SR-52 and an SR-56, both fully operational.

2025-08-31 17:38:59 UTC

- reply

Alchemy shouldn't do absolutely nothing. It should consume a bunch of something valuable, and produce a little of something useless.

2025-08-22 19:00:03 UTC

- reply

Isn't the point of having some smaller ones like 0.1, 0.047, or 0.022 that the impedance vs frequency curve is better? Or do larger values like 4.7 in 0201 have impedance curves more like the 0.1 in 0603?

2025-08-22 18:48:45 UTC

- reply

My computer doesn't have a halo, and probably should have horns and a forked tail.

2025-08-22 00:03:27 UTC

I haven't lost my mind. I have it backed up on floppies somewhere.

2025-08-18 20:10:32 UTC

Breaking:
Leapfrogging the competition's six-blade and seven-blade razor cartridges, Gillette announces the new FusionNano. Using the latest advances in both fusion energy and nanotechnology, the FusionNano incorporates over one billion cutting edges in each razor cartridge. Developed in cooperation with the Lawrence Livermore National Laboratory (LLNL), the nanotechnology blades of the FusionNano cartridge are forged by 500 terawatt (TW) laser pulses at the National Ignition Facility.