Recent content by Svenska

That seems like you are doing something you really should not do, then trying something else to fix it up afterwards. I see this approach in professional software too, and it always causes issues down the line - because the fix will never be fully free of side effects.

We all have to thank you (and everyone else involved) for the Lisa work. Without insider access, these things are completely impossible, and since "no" is always the safe option, everything involving large cooperations is an uphill battle. So thank you!

Sure, they were able to produce working binaries (which is very impressive, don't get me wrong), but I was wondering whether they are able to reproduce the binaries exactly. In particular, and the video also mentions that, the code segment order (linking order) is sometimes significant. I don't...

Cool stuff. It would have been interesting to know how closely they have been able to reproduce the original binaries.

Neither CP/M nor DOS shipped with such a tool. However, both shipped with primitive development tools (ASM, DEBUG) and it is easy enough to write if needed.

I've played with the KC85/2-4 series. While they can handle huge amounts of memory, their basic architecture is not a good match for CP/M: The bottom 32 KB are fixed RAM, followed by a switchable 16 KB bank, followed by 16 KB fixed ROM. Remaining system memory, video memory, segmented memory...

That's a fair point and a great example of missing the point - we were discussing C compilers in the context of CP/M and, by extension, comparatively small systems. Turbo Pascal does not qualify because it's not a C compiler (does it even support object files?); shared libraries do not qualify...

If a linker is incapable of throwing out unused objects, I simply consider it badly written. Whether you choose (in your local environment) to disable that optimization to gain a little more speed is your own decision. If you disagree with that, then I will simply have disagree with you on that...

Yet none bothered to access both sides simultaneously, which is enough I need to know about the idea. :)

Your argument holds about the same amount of water as a hardware company intentionally producing badly written software to promote sales of better hardware. Yes, such a case can be made, yes, it has been done and will be done again, and yes, I will dismiss it as invalid anyway.

Only in the form of a ramdisk or through system-specific interfaces (such as RomWBW), as we have discussed before - not in the general case. And please don't start another huge discussion about this.

You don't need to do case studies, you can read the blog posts by David himself. He describes the language and the design choices made specifically to avoid the limitations of 8080 / Z80 / 6502 processors. Giving up reentrancy and stack-based parameter passing, then adding function nesting...

To most people, CP/M means CP/M-80. CP/M-80 runs on 8-bit computers and DOS does not. That's the major difference. If you look at x86, DOS is a superset of CP/M-86 - anything CP/M-86 can do, DOS can do as well. I own a strange machine which can run either MS-DOS 3.31 or DOS Plus (an extended...

Given the current state of most media, I consider the verification done as part of formatting worth doing anyway. Sure, but if you're going to replace everything (interface, controller and drive electronics) to get a dual-track floppy drive, you should have designed a tape drive instead. Gets...

Sure, but we were talking about PC-style and USB/UFI hardware. :) An Amiga or 1541 could do substantially better... You only have one set of wires for read data, write data, write enable. The selection logic in both the controller and the drive is built around that, and they both only handles a...