Stack Exchange
log in

dirkt

 Discussion on question by cjs: Were C…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Nov 20, 2024 09:01
"which have ROM in the first 16K" means that $4000 is the first RAM address, so the memory maps you found and the assumptions for that particular CP/M seem at odds. Often ROM routines use the low RAM, so I guess that would be the reason to start at $4200, $4300 or whatever works with the existing ROM.
 

 Discussion on question by rwallace: D…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Jul 5, 2024 08:12
"the color of something is determined more often by what it is than where it is" these things are called "sprites". Making the whole screen consist of sprites is impractical. But you also have "mixed modes" where 2 bits in the character determine the color (trading resolution for colors), and on top, there's an attribute cell which modifies (some of) the colors.
 

 Discussion on question by scruss: Pro…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Jun 19, 2023 16:55
@supercat this is as bad as using 360 degrees or 400 gons for one full rotation, mathematics doesn't give you any way to calculate trig with these units I know of. And yes, you make a case distinction on the range before performing the calculation, which you need to do anyhow. See e.g. here for an actually used algorithm.
dirkt
Jun 19, 2023 16:55
@supercat I meant radians as in "2*pi radians = 360 degrees = 1 complete rotation". I guess that's what you mean by "2*pi radians". I only know algorithms for trig that use this representation. I don't know which algorithms you know.
dirkt
Jun 19, 2023 16:55
Note that Logo also had radsin etc. with the angle in radians. The reason most languages prefer trig functions with radians is that the internal calculation uses radians, and if you use degrees that's an extra multiplication before the internal calculation, so they leave it to the user to decide if they want to do this extra multiplication.
 

 Discussion on question by Noob_Guy: W…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Jun 17, 2023 08:28
You are describing what lead to the development of the Apple homecomputer: Steve Wozniak wanted to make a cheap video terminal, and ended up with a complete computer (except with the 6502, not the 8080). So no, it doesn't have to be connected to a host computer, but the itch that the VT100 wanted to scratch is "I want a video terminal instead of a teletype".
 

 Discussion on question by javanerd: W…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
May 23, 2023 14:49
We had Javascript first (which was designed on a paper napkin and was intended to make some things blink, and then sort of outgrew it's design spec...), and now we have WebAssembly, WebGL, WebGPU, you name it. Just please not another high-level language, Javascript is quite enough. So "it was never added" is false. Also, not retrocomputing.
 

 Discussion on question by Hennes: "Ma…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Mar 18, 2023 18:36
Rewrite of the OS would even take longer... and it's much more attractive to use a OS that's made to fit the Z80. Like CP/M. I think there are a few rewrite attempts for that. Or small variants of Unix.
dirkt
Mar 18, 2023 18:36
Having to rewrite all the /360 code into Z80 is a non-trivial task and will probably take a few man-years at least, so I'd be surprised if the answer is anything but "no".
 

 Discussion on answer by Scrooge McDuc…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Dec 6, 2022 07:32
I don't have a playstation to test it... that's why it would be interesting to know what exactly has gone wrong, instead of "just do those particular steps with those tools". And now looking at it again, it might still be just an issue with the tools itself. I know 90% of Q&A on the internet today is just "Don't try to understand, just do those steps, turn off your brain", but it's not a particular good way to work with computers...
dirkt
Dec 6, 2022 07:32
After looking at the link, it looks to me like it is a big-endian vs. little-endian issue, and the Playstation only seems to support little-endian. So that's what the answer should start with. The rest just looks like a somewhat complicated way to achieve this with particular tools, I can think of at least two different ways out of my head to do differently with different tools (e.g. without any cue files).
dirkt
Dec 6, 2022 07:32
It would be nice if you could explain why your earlier attempts failed, and what your recipe does to avoid those problems. That would allow to replicate this with different tools.
 

 Discussion on question by Kyndel: Why…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Nov 29, 2022 17:31
@Justme You are again back to your definition "it's non-interlaced because all fields are identical". But each field contains the marker that says "start on the even line" by virtue of timing and sync pulses. You could equally well make identical fields that all start on the odd line. And since you can do that, and since the signal still contains only half the possible lines, it's an interlaced signal (according to my definition, but not yours). It does not have to contain different fields to be an interlaced signal. I am sorry, I can express it any simpler... No really EOT.
dirkt
Nov 29, 2022 17:31
@Justme have a look at e.g. this. The signal contains a different number of pre- and post-equalizing pulses on vertical sync depending on the field. The C64/NES/Apple send this format, and they send the number of pulses for one field all the time. So e.g. instead of sending a second "odd" field, they send the even field again. But it still contains the marker for even/odd. Your "non-standard non-interlaced" signal is an interlaced signal., with all the markers. That's where your definition differs from mine. EOT?
dirkt
Nov 29, 2022 17:31
@justme you are just using a different definition. For me, a signal is interlaced if it allows to send different fields for half-images, and if there is an indicator that determines which half image it is. Which is the case for the analog TV signal, and which is the only format the TV understands, so you got to use it. For you, a signal is interlaced if it actually contains both half images. And according to your definition, the point the OP is asking about actually never happened - very few computers actually sent both half images, so the "local" signal was always "progressive".
dirkt
Nov 29, 2022 17:31
@Justme it's a matter of definition: the TV accepts interlaced, with each half-image being progressive. So if you are only sending one half image all the time, you are only sending every second line all the time, and you are loosing the vertical resolution of the TV mode, which is interlaced. The signal itself is still interlaced: It uses the vertical blanking bumps to decide between even/odd half-image. No way around that. And yes, it works fine, but again: the whole "point" is that the TV doesn't accept anything else.
dirkt
Nov 29, 2022 17:31
The point is that for analog TVs, there was no alternative. You wanted to connect any kind of computer to an analog TV in the 80s, you had to send an interlaced signal, because no other format was accepted. Though most home computer cheated and only sent the first half-image, halving vertical resolution.
 

 Discussion on question by Richard T: …

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Aug 11, 2022 08:52
In unix, both shell return status and OS call return status have multiple "non-ok" codes and just one "ok" code. And while you could make 42 the "ok" code and the "non-ok" codes start with 0, somehow having the unique single "everything is ok" code at 0 make more sense, doesn't it? In particular as checking zero/non-zero is a very efficient operation. So while this is also speculation, it least it seems pretty obvious to me.
 

 Discussion on answer by manassehkatz-…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Jan 11, 2022 11:38
@Piovezan Wikipedia has a generation list for ICs with transistor count and also RAM chip capacity. By 1969, available RAM chips had ca. 256 bits (so 32 8-bit-words, though words didn't have 8 bits then). That's enough for a register bank, but not enough for main memory.
 

 Discussion on question by Piovezan: W…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Jan 11, 2022 11:38
(1) Core memory was hand threaded, and there was a natural limitation to density. (2) RAM chips size scaled with the process (wikipedia has a year/char. length table on the right). You just couldn't put enough transistors on early ICs to make RAM of meaningful size.
 

 Discussion on question by Leo B.: Wha…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Apr 8, 2021 17:31
Also note that, historically, the incentive to package up files differently started with telecommunications. First you needed to make sure that binary files were safe to transfer (expanding them instead of compressing them), and after LZW was published and patented, compression took off. This is very different from keeping uninitialized sections out of executable formats, which is what SpaceMaker does, and which other executable formats don't need because it's baked in. So be careful not to compare apples and oranges when looking for "earliest".
 

 Discussion on question by OmarL: Why …

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Mar 29, 2021 17:56
@supercat "people would think it desirable to cut the speed of additions in half" they don't. That's why everyone switched from one's complement to two's complement. "where one wants to add readouts to a lot of registers." I think that was never the main use case. Mechanical calculators used "nine's complement", because it's reasonably cheap in terms of gears (but you need to perform additional cranks). When binary computers came up, copying this was one obvious choice (sign bits were the other). Two's complement wasn't well understood. Once it was understood, everyone switched.
dirkt
Mar 29, 2021 17:56
@supercat "On a serial system, one could perform end-around-carry with minimal hardware if every value gets shifted through the adder twice, but that would take twice as long as using a two's-complement addition.". Yes, exactly. And that's what you wanted in old systems - you trade expensive hardware components for speed. And it doesn't make sense to do that with a parallel adder. So you just get rid of the one's complement. And one's complement was used for traditional reasons, not because it was faster or more efficient than two's complement - it never was.
dirkt
Mar 29, 2021 17:56
@supercat "I don't see how one could efficiently process ones'-complement serially". Have a look at end-around carry.
dirkt
Mar 29, 2021 17:56
An important aspect could also be that one's complement addition and subtraction with end-around carry and borrow makes much more sense for bit-serial ALUs than for bit-parallel ALUs with carry-look-ahead, where this just can't be parallelized. So I'd imagine that contributed to the overall move from one's complement to two's complement. Though I am not sure if this already applied to the PDP-4.
 

 Discussion on question by Schezuk: We…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Mar 27, 2021 18:04
Implementing a generic key/value mechanism in a low-level languages like C, which most OS are written in, is much more effort than just implementing static attributes. So basically all OS started out with static attributes, for simplicity, and to save space on disk.
 

 Discussion on question by Paul Ghobri…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Feb 21, 2021 11:50
Does "magazine" mean "something provided for money by some kind of publishing house"? If not: Electronic distribution of the kind of information you'd pay for in a magazine, but between peers, happened long before that.
dirkt
Feb 21, 2021 11:50
@user3840170 "The answer may not be easy to research or verify, but at least in principle, an objective answer should at least exist" - which is exactly why I don't like those "which is the first X" questions, at all. A better question would be "what are some of the earliest X". On top of that, sometimes "being X" isn't particular well-defined, because contrary to what some people think, invention proceeds by improvements. So finding an exact point in time when something "became X" often doesn't make sense. But I guess all of this is a meta-discussion, and doesn't belong here.
 

 Discussion on question by Schezuk: Ch…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Feb 19, 2021 09:13
@Schezuk the PDP-8 can access 4 kWords of 12 bits = 48 kbits in total, without bank switching. The memory extension controller increases this by a factor of 8, for a total of 368 kbits. So where ever the 1Mbits where stored (guess: DEC tape, and loaded on demand), it's not in some kind of directly addressable ROM-like memory.
dirkt
Feb 19, 2021 09:13
Also keep in mind that core memory survives loss of power. So early computers basically used core memory as "ROM" - when you turned the computer back on, the OS etc. was still there. If it got corrupted, you'd bootstrap, either by manually entering a short program into the core, or by having core rope/a coldstart panel/some module with it somewhere. But that was a handful of words, often less than a kbit. So "price per kbit" for that wasn't important...
 

 Discussion on question by Schezuk: Wh…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Sep 13, 2020 17:24
It was a gradual process, and results vary much depending on the particular code, so I don't think you can pinpoint a certain compiler, or a certain year... And the "4 times slower" is probably the result of some particular benchmarks, not a universal truth.
 

 Discussion on question by Nelson Blak…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Jun 10, 2020 08:31
I mean, MS-DOS can shutdown quicker...
 

 Discussion on question by apschultz: …

Imported from a comment discussion on unix.stackexchange.com/q...
dirkt
Apr 13, 2020 06:42
What I'd do in this situation is put the app with the raw socket in a network namespace, connect it up with a veth pair to the main namespace, and then re-arrange the VLANs on the veth pair end and/or the wifi interface to make everything work. Then you've cleanly separated the raw socket and the other master/slave constructions.
dirkt
Apr 13, 2020 06:42
@apschultz "sending a tagged packet using a raw socket on the base interface doesn't write the packets to the vlan interface." I'd be surprised if this even works for normal interfaces. Raw sockets don't integrate too well with the rest of the add-ons.
dirkt
Apr 13, 2020 06:42
Dummy devices are really only intended as a place to assign IP addresses. I am not sure if they work with VLANs attached, at all. I guess the standard setup would be to create the VLAN on the wireless interface, and then bridge as needed.
 

 Discussion on question by bitmask: Ur…

Imported from a comment discussion on german.stackexchange.com...
dirkt
Aug 27, 2019 15:49
@bitmask: Wenn man "Judensau" als die Herkunft von "Sau" als Schimpfwort sieht (und nicht umgekehrt als einen Zusammensetzung, wann auch immer sie stattgefunden hat), muss man aber schon erklären, warum es soviele Schimpfworte mit Ferkel/Schwein/Sau gibt, die mit Juden nichts zu tun haben...
dirkt
Aug 27, 2019 15:49
Da es Ausdrücke wie "Du Ferkel!", "den inneren Schweinehund überwinden", oder "das interessiert kein Schwein" gibt, die alle keinerlei Verbindungen mit Juden oder einer sonstigen Gruppe vom Menschen haben, würde ich einen Zusammenhang mit Antisemitismus als eher fragwürdig betrachten. "Judensau" wäre dann einfach eine Zusammensetzung, genau wie andere Zusammensetzungen mit "-sau" ("Saupreiß"!).
 

 Discussion on question by Leo B.: Wha…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Feb 24, 2019 12:14
There are quite a few variations of the Lunar Lander game. "In Pascal" is an unusual one, as is "tracking attempts" (when do the attempts reset? When the program is restarted?). But with more information one might at least narrow down where this version split off: Can you rotate the craft? Can you adjust parameters, e.g. the moon gravity, or fuel reserve? Did it keep the crater depth calculation for a bad landing? Etc.
 

 Discussion on question by Minix: Why …

Imported from a comment discussion on unix.stackexchange.com/q...
dirkt
Jan 10, 2019 16:50
What are your kernel command line options when you boot either kernel? (You must look in the configuration of your boot loader to see this, not at the filesystem).
dirkt
Jan 10, 2019 16:50
If the kernel itself is identical, the difference must be somewhere else: For example in the initial ramdisk (initrd), or in the kernel command line options provided by the bootloader (grub etc.). Only you can find out what else you have, and if it is different or not.
 

 Discussion on question by rjhwinner03…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Sep 9, 2018 12:02
@AlexHajnal: Your comments are enough for an answer, even though it's just partial at this point, but it will already be valuable for anyone else in a similar situation. You can edit your answer and add to it as the OP gives more info.
 

 Discussion on question by nbloqs: Sma…

Imported from a comment discussion on retrocomputing.stackexch...
dirkt
Aug 25, 2018 10:42
Note that the first computers (no matter if they had relays, tubes, or other elements) generally used some other ways of storage (CRT tubes, mercury delay lines, drums, ...); so replacing this by SRAM instead of trying to replicate the storage method is probably fair game. Of course, you can implement storage with relays, and add this storage to all of the designs you mentioned, and the "smallest machine" would be the one with the least amount of storage, but I'm not sure if this helps...
 

 Discussion on answer by Raffzahn: Que…

Imported from a comment discussion on retrocomputing.meta.stac...
dirkt
Jan 29, 2018 09:33
Maybe to illustrate: Here's a series of videos of the restoration of a Xerox Alto from Digibarn, with work on the PSU and the CRT. All perfectly fine retro-computing stuff IMHO, and all "just electronics".
dirkt
Jan 29, 2018 09:33
@Raffzahn: And I'd say the same line of reasoning applies to power-supplies, mechanical or electronic parts of the floppy disk drive, mechanical or electronical parts of tape drives, etc.
dirkt
Jan 29, 2018 09:33
@Raffzahn: I'm not saying the Electrical Engineering is off-topic for such a question, or not some place where you should ask. I'm just saying that questions about CRT repair related to retro-computing are on-topic here. IMHO, YMMV.
dirkt
Jan 29, 2018 09:33
@Raffzahn: The criterion IMHO is simply "is it computer related?". If it's a question about a TV set, with no connection to any computer, it's off-topic. If it's a question about a CRT used in a retro computer, it's on-topic. If you have a retro-computer that somehow needs a car carburator to function (can't imagine how that should work, but there's an analog computer that uses liquid and pipes, so who knows), it's on-topic. If no retro-computer is involved with the car carburator, it's off topic. Transport clearly doesn't count. Simple, no?
dirkt
Jan 29, 2018 09:33
I agree with @wizzwizz4. Questions about repair and maintenance of CRTs, e.g. as used inside terminals, are perfectly on-topic, even if it's "just" electronics. And "modern" electronics doesn't really deal with this kind of stuff, you are more likely to find an audience who can answer these questions on retrocomputing. A question, say, about how to make changes to an Apple motherboard to achieve something particular would also be on-topic. Even if the answer involves "just" electronics.
 

 Discussion on question by Brad: Silen…

Imported from a comment discussion on superuser.com/questions/...
dirkt
Mar 28, 2017 19:19
Then another wild guess: If it happens on ARM (RaspPi), but not on x86 (PC), it might be some bug in the driver that only shows up when compiled on some architectures. For example, ARM and PC have different memory coherence protocols.
dirkt
Mar 28, 2017 19:19
Wild guess: Can there be anything on the Behringer side that you need to configure properly? If somehow playback and recording is confused, or they influence each other, it may be a bug in the driver, or the card may need a quirk. File a bugreport with the developers of snd-usb-audio.
dirkt
Mar 28, 2017 19:19
Can't find a quirk for this, so I'm running out of ideas. Things to test: (1) Reduce to 8/8 or even 2/2 channels in the Behringer UI, reconnect, and see if it makes a difference. (2) Try to play some WAV files in 8/8 or 2/2 and see at least this direction works.