Comments by "Lawrence D’Oliveiro" (@lawrencedoliveiro9104) on "ForrestKnight" channel.

  1. Turing also published a groundbreaking paper in an entirely different field, namely biology. He was considering the issue of how the amorphous mass of cells making up a fertilized embryo can suddenly decide that this is its “front” and this is its “back” and that is “the left side” and that is “the right side”.
    51
  2. 4:34 Perhaps more importantly, MINIX was only available under a non-Free licence: you had to buy a copy of Tanenbaum’s textbook in order to get the licence. Linus wanted something he could use and redistribute freely, and that others could also use and redistribute freely.
    34
  3. Prophets are not necessarily people you can get along with personally.
    29
  4. 1:10 Multics had some features that would still be considered advanced today. For example, something still lacking in present-day Linux is the ability to share ownership controls over a set of files/directories between more than one user. Linux has ACLs, but there is no ACL entry conveying ownership privileges, like there was on Multics.
    23
  5. 7:51 Church and his student Kleene did some interesting work with the λ-calculus. For example, they showed how mathematical paradoxes (like Russell’s paradox) could be represented by expressions that could be manipulated logically without the whole world collapsing about your ears. Mathematicians go to great lengths to try to ensure that their theories are free of paradoxes. But I think λ-calculus shows how you can tame the paradox and not be afraid of it.
    16
  6. I doubt the toenails part. Remember, he was living in a country that offered free Socialist-style healthcare to all its citizens.
    9
  7. 6:18 Worth making something very clear here: the program (call it D, the “deciding program”) that decides whether a given program (call it P, the “problem program”) will terminate for input data I is taking both P and I as input data. In other words, a program is input data to another program! This is a key point about the nature of the Turing machine, and also of all our electronic digital computers: programs and data are both represented using the same set of symbols that can be stored in the memory of the machine. The only difference between the two is, a stream of symbols becomes a “program” only because you point the CPU at the start of that stream and say “run this as a program”.
    7
  8.  @Digger-Nick  It’s the other way round. WSL2 is the first step towards getting rid of the Windows kernel. Unless you’re talking about WSL1, in which case ... I pity you.
    7
  9. 0:14 Also worth watching, if you like the dramatization approach, is a BBC TV movie about Turing from 1996, called Breaking The Code . This was based on a stage play from 1986.
    6
  10.  @shallex5744  The GPL doesn’t prevent commercialization by large companies, either. And the FSF is quite happy to include non-copyleft licences under its definition of “Free software”. Apart from that, you are just regurgitating the same old Stallman cant. So what exactly is the problem with “Open Source”, again?
    6
  11. 1:17 There were other groundbreaking OSes from around this time--you have already mentioned Multics, let me add also TENEX and ITS. Why is it most people only remember Unix nowadays, and none of the others? It is because Unix was the only one to make the leap to becoming a portable OS. The others were developed on particular hardware architectures; when those died, they died with them. While the original PDP-11 architecture that Unix started on is a museum piece nowadays, the OS architecture itself was sufficiently adaptable to make the jump to 32-bit, and then 64-bit, hardware. And Linux has carried that portability tradition to the next level, being ported to an average of one new major architecture for just about every year of its existence.
    5
  12. 5:17 Actually, Stallman was not keen on Linux at all. He was working on his own “Hurd” kernel, which was going to be have a “microkernel” architecture, which would supposedly make it much easier to develop, test and maintain. Better than an old-fashioned “monolithic” kernel, like Linux. Well, here we are, about 30 years later, and we’re still waiting for Hurd to reach production quality. Meanwhile, Linux started becoming popular just a couple of years after that initial announcement, and spread to a whole bunch of non-x86 architectures at the same time.
    5
  13. I don’t understand what exactly Stallman disagrees with. If you look at the Open Source definition, what is there that goes against his philosophy?
    5
  14.  @gauravshah89  He didn’t want to name it after himself. He wanted to call it “Freax”. He is certainly no “narcissist”, as the quality of his contributions can attest.
    4
  15. Stallman was responsible for the “POSIX” name, too.
    4
  16. I remember when the Raspberry π first came out, there was no good OS for it. You could run a version of Debian for ARM, but that was not optimized for the particular combination of an older version of the ARM instruction set, combined with hardware floating-point, that was implemented on the Broadcom processor used in the π. So a couple of guys set out to rebuild the whole of Debian just for that hardware. As I recall, it took them about six weeks. And remember, all this building had to be done on the Raspberry π itself--the original single-processor model. The result was called “Raspbian”, but these days I believe the Raspberry π Foundation have taken it over and officially dubbed it the “Raspberry π OS”. So that’s an example of the kind of work it takes to create your own distro.
    3
  17.  @liquidmobius  That’s FreeBSD-based. Why not mention MINIX? That is a “true” microkernel design.
    3
  18.  @liquidmobius  Worth mentioning, then Intel’s use of MINIX as the OS on its “Integrated Management Engine” included in many of its processors. Remember you claimed it was “far superior” in terms of “security”? Well, it turned out the IME had some rather glaring security holes. And given that the subsystem has total control over the CPU, and can totally bypass any privilege protections of the OS running on there, the consequences have been rather ... unfortunate.
    3
  19. BSD development has a whole different culture around it -- more centralized, less adaptable. I doubt it could have filled the gap.
    3
  20. 4:06 Gödel’s name sounds more like “guh-del” then “goo-del”.
    3
  21. Is anybody still using Windows for serious work? Other than games?
    3
  22.  @johnwellbelove148  None of which is relevant to the issue of IDEs and compilers that you initially raised. Whatever you are developing for, Linux is a more versatile development platform than Windows. And this scares Microsoft, which is why it is putting so much effort into WSL.
    3
  23.  @elliotts5574  It’s MINIX. MINIX is a microkernel -- there is no “confusion” about that.
    2
  24. Yeah, but Turing’s tape was erasable.
    2
  25. IBM’s OSes were never “open source”. Source code may have been available to customers, but under restrictions.
    2
  26.  @danielrhouck  Initially IBM made its software available to customers free of charge. But this was seen as unfair to independent developers who wanted to create a software industry. So as a result of the 1956 “Consent Decree”, I think it was, IBM were forced to “unbundle” -- charge extra for the OS etc.
    2
  27. He is a prophet. He may not be nice to know personally, but he has been right about the dangers of proprietary computer systems more than once.
    2
  28.  @bazoo513  Companies like Microsoft were always going to badmouth anything that they saw as a threat to their business model, even if someone like Stallman did not exist. There were those who even claimed that Free software was “communist”!
    2
  29. Seems like games are the only thing left. Isn’t anybody doing serious work on Windows any more? . . . (Crickets)
    2
  30.  @gauravshah89  You “may as well be a lie”. You like BSD, fine. Don’t try demonizing Linux, or Linus.
    2
  31.  @gauravshah89  You are seriously out of touch.
    2
  32. 4:50 Not just “short-term” in the server space. Linux is the reason Windows Server, at least the on-prem version, is no longer getting the love it once was. It’s the reason why the HPC version and Windows Home Server both died a quick death.
    2
  33. 5:20 That “which implies” part has not actually been proven. It’s called the “Church-Turing thesis”, and the mathematical term for it is a “conjecture”. It seems to be true, as far as we can tell, in all the examples so far, but, as for the general case, we can’t be sure either way.
    2
  34. 18:20 The “Get The Facts” campaign also put out an ad in the form of a fictitious newspaper titled “The Highly Reliable Times”, touting the site-wide installation of Windows NT at the London Stock Exchange. They were not quite so self-congratulatory a little while later when the system suffered a complete meltdown which Microsoft tried to blame on “network issues”. A little while after that, the LSE ripped out their Windows NT servers and replaced them with Linux.
    2
  35. The common factor in all of these variants is ... Linux. Hence the name.
    2
  36. Games seem to be the only real reason to continue using Windows. Which leads to the question: is anybody still doing serious work on Microsoft’s platform?
    2
  37.  @roastxbeefx7744  Is Dotnet still part of UWP? Or Project Reunion? Or whatever new grand scheme Microsoft is pushing this week?
    2
  38. This is why Microsoft is trying so hard, with WSL, the new Windows Terminal, PowerShell and all the rest of it, to turn Windows into something more like Linux. But still, for some reason, they believe that 26 drive letters ought to be enough for anybody ...
    1
  39. Microkernels are pretty much a relic of history at this stage. ∗cough∗ GNU Hurd ∗cough∗
    1
  40.  @liquidmobius  Nobody has got one working properly to production quality. Nobody. FreeBSD I think has the vestiges of a microkernel at its heart, but that had to be bastardized somewhat in the name of performance.
    1
  41.  @liquidmobius  I wonder why you say that such characteristics of the software have nothing to do with the software itself. That’s a very dubious assertion to make. The objective facts are that the IME runs MINIX, and it has vulnerabilities.
    1
  42.  @liquidmobius  Sounds like you are trying to argue in circles: first claiming that microkernels are more “secure”, then, when faced with evidence to the contrary, trying to claim that those vulnerabilities don’t count. If serious vulnerabilities don’t count, then which ones do?
    1
  43.  @trollerjakthetrollinggod-e7761  If a microkernel cannot safely run in a “privileged state”, then how is it supposed to be used?
    1
  44. Worth mentioning that “operating system” seems to be essentially synonymous with “kernel” -- at least according to Andrew Tanenbaum, the author of the MINIX Book. So the GNU project was (and still remains) a userland in search of an OS.
    1
  45. Oh heck no. I suspect Nokia is better known--and much missed by many former users, victims of Microsoft’s hamfisted destruction of the company. But yeah, Linux is certainly longer-lived.
    1
  46. EWD could be a bit of a troll sometimes, as in his comments on programming languages. For example, on BASIC, he said “It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration.”
    1
  47.  @D0J0P  Sure it was, but remember open-source OSes were not heavily used back then. So it’s not that people “abandoned” BSD, it simply didn’t grow as fast as Linux. I don’t think AT&T were ever fans of BSD. That’s why you had the AT&T-versus-BSD divide among flavours of proprietary Unix.
    1
  48.  @leogama3422  Not a chance. Open-source is about enhancing competition, not killing it.
    1
  49. Linux runs on them all.
    1
  50. MacOS is “official” Unix, but Linux works more like you would expect a Unix system to work.
    1