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

  1. Windows is not that popular with developers. Why do you think Microsoft is trying to add a Linux layer to it?
    2
  2. youtube-dl is still available.
    2
  3. If they were using Microsoft Team Foundation Server, you can understand why anything would be seen as an improvement on that ...
    2
  4. I use Linux. I have used a range of languages with Linux. Before I used Linux, I used a range of languages on a bunch of other platforms. I not only understand what a “curried function” is, I have used them in real-world situations. Would you like a “so what?” lecture? I could give you one.
    2
  5. Fun fact: Functional programming is older than OO.
    2
  6. @Michael Hansen I think the earlier ILLIAC IV research machine was also SIMD, with four processors executing the same instruction set on four data streams. Seymour Cray looked at the difficulty of programming that before creating his architecture for the Cray-1.
    2
  7. 8:24 The trouble with the hardware blitters in the Amiga and Atari ST is that they were never as powerful or flexible as software. The Macintosh had its QuickDraw graphics routines, which implemented sophisticated drawing modes and nonrectangular clipping. When you tried to build similar functionality on top of a more limited hardware blitter, the performance would often end up worse than if it was all done in software.
    2
  8. Modula-2.
    2
  9. I recall a colleague, a former IBMer, mentioning how some of those pages went beyond the usual “THIS PAGE INTENTIONALLY LEFT BLANK” business. Instead, they said “DESTROY THIS PAGE”.
    2
  10. If you look at the three different kinds of address numbering needed for memory, they are * byte addresses within memory (call this number B) * bit numbers within a byte for masking (call this number b) * bit numbers within a byte for integer digit values (call this number a) In all little-endian architectures, we have a = b and B = int(b / 8) which is very simple and straightforward. But in big-endian architectures, the situation is more complicated. Also, another peculiarity of big-endian architectures is that CPU registers are still effectively little-endian! For example, consider an instruction sequence like move_word a → b move_byte b → c Does c end up with the high or low byte of a? In little-endian architectures, it is always the low byte. In big-endian architectures, it is the high byte if b is in memory, but the low byte if b is a register! Thus, the only truly consistent byte/bit layout is little-endian.
    2
  11. 0:37 Blur ... focused ... heh-heh.
    2
  12. Phosphoric acid dissolves your teeth!
    2
  13. 4:31 Marketing travesty generator! Is there a copy of that text anywhere? ;)
    2
  14.  @RobertMilesAI  Fun fact: Python allows semicolons, so you can put multiple (simple) statements on one line! Its rule for omitting semicolons is a bit saner than the one in JavaScript.
    2
  15. Isn’t there such a thing as an echo-cancellation microphone? The front one picks up your voice, the back one picks up the room reverb, and by subtracting the latter from the former, you get a much clearer, less boomy recording.
    2
  16. But will they help him find the Golden Fleece?
    2
  17. No, it’s really the opposite of antialiasing.
    2
  18. 1:24 Someone who has no recollection of Life Before Linux ...
    2
  19. Luckily, there are Open Source alternatives that are much more modern, like R, Octave, and of course Jupyter notebooks using Python with Matplotlib.
    2
  20.  @TrueThanny  Yes, address books are precisely my point. Why do we have to do the lookup of the number, and then have to dial that number? Why not have the computer do the lookup for us? That’s how the DNS works -- it’s like an address book that automatically keeps itself up-to-date!
    2
  21.  @Richardincancale  I’m sure there were smart researchers at IBM, but somehow their shipping products rarely reflected that -- they were clunky, resource-hungry and overcomplicated compared to, say, DEC and Unix systems.
    2
  22.  @Harry-lu7hj  C# does it all for you ... on only one platform. “26 drive letters ought to be enough for anybody!”
    2
  23. Java doesn’t “ban” pointers, it only obscures them. As a consequence, it suffers from the aliasing problem, e.g. a = b; a.field = «value»; System.out.println(b.field); /* what does this print? */
    2
  24. One drawback with DNS A records, in particular, is that they only give you an IP address, not a port number. So they are not sufficient to identify a service, only a machine which might provide that service. This was remedied later with the introduction of SRV records, but they are not heavily used.
    2
  25. Interlacing should die already.
    2
  26.  @muhamednourezzeddin2784  The whole point with comments is that they’re not part of the “data structure” (or indeed any semantic content).
    2
  27.  @Carewolf  No, that’s not error diffusion. Error diffusion involves keeping track of the error between the true pixel value and what you have actually displayed, and accumulating (“diffusing”) that into subsequent pixels. Hence the name.
    2
  28.  @Carewolf  Floyd-Steinberg is one of many algorithms under the category of “error-diffusion”. It’s not the only one.
    2
  29. 0:39 Two places where it’s not used: SSH and VPN connections. They have their own encryption handshake protocols. Make that three: wi-fi also has its own security system(s). Bluetooth? Four. Among the places where it is not used ... nobody expects the encryption inquisition!
    2
  30. That’s called “key management”. It’s an entirely separate issue. Perhaps you do it by turning up in person at the registration centre with suitable ID.
    2
  31. There is a more fundamental aspect of the “laziness” principle not mentioned here, and that is that a computer is a _universal machine_, capable of being programmed to perform any kind of tedious and repetitive task you can describe to it. Consider the number of people who use GUI tools to perform repetitive tasks, for example, when a command-line tool could be scripted to perform the same thing automatically, leaving the human free for more interesting tasks. GUIs have their uses, but they cannot be automated.
    2
  32. 4:09 Ah, the joys of Baudot code ... Baudot gave his name to the unit “baud” ... which is another term you don’t hear used much any more ...
    2
  33. 2:02 The I/O problem mostly went away of its own accord. Once punch cards, paper tapes, magtapes, line printers and all the rest of it went away, everybody pretty much settled on something Unix/POSIX compatible (stdio). Writing to that gives you a nice, portable common denominator.
    2
  34. The Debian emacs-common package is just about 95MB.
    2
  35. I can’t even see anything in standard SQL that allows columns to be referenced by numbers, only names. Not in the SQL 2003 spec, anyway.
    2
  36. The private key never leaves your custody.
    2
  37. Sonic screwdriver!
    2
  38. One end still has to own both halves of each public/private-key pair.
    2
  39.  K.D.P. Ross  Because pure-functional languages tend to be awkward to use. Procedural languages which allow functional constructs (like Python does) tend to work better.
    2
  40. 1:34 Well, by “completely changed” you mean that somewhere around 50% of the bits in the hash flip to different values. You don’t want them all to flip (or flip according to any discernible pattern), otherwise that’s no longer quite so random.
    2
  41. No, C++ and Python do not let you define new operators, they only let you overload existing ones. ALGOL 68 did allow you to come up with new operators. But you couldn’t use non-alphabetic/numeric characters for them.
    2
  42. 6:09 Idea: why not let the robot loose with that idea, and see what solution it comes up with? This is how we give birth to the evil robot overlords, isn’t it?
    2
  43. Many implementations did “lazy buffering” to get around this. For example, DEC’s VMS compiler (which I liked because of its system-programming extensions).
    2
  44. 11:54 Yup, all the Free Software desktop environments can agree on common standards as prescribed by freedesktop.org, I think is the main place where these things are discussed. They’re not out to wrest control of the marketplace from each other. So your choice of DE doesn’t even have to dictate your choice of what apps you can run.
    2
  45. Macs are the ultimate example of how something can officially license the “Unix” name, yet work nothing like people expect a traditional “Unix” system to work.
    2
  46. Since you ask an off-topic question, let me give an off-topic answer: Have you tried turning it off and on again?
    2
  47. An 11/70 was the main campus machine when I started my Comp Sci degree (1979). I learnt a lot from messing around on it--stuff that wasn’t even taught in lectures. Several of the lecturers seemed to have expertise elsewhere. For example, one handed out notes on how to do system service calls on a UNIVAC machine, because that’s all he knew. Which was completely useless, given we had no access to such a machine. Nobody gave a lecture on how to do service calls on RSTS/E — what a wasted opportunity, when there was an actual living, breathing machine running that very OS just outside the lecture hall, that the students could have run actual exercises on.
    2
  48. Yes, they do talk to ordinary nobody clients like you or me. In fact, a lot of their traffic is completely unnecessary, coming from misconfigured clients that should be sending their queries elsewhere. But they can’t block it.
    2
  49. Ah, is that why the limit.
    2
  50. 7:30 Uncanny: I had a graduate CS lecturer who reported the exact same experience.
    2