3 ms·

An anecdote about backward compatibility

82 points by speckx 10mo ago

mikelitoris 10mo ago
For those who don’t get it: It’s referring to the ink soaked ribbon that would print characters on a piece of paper, similar to a typewriter. This is a preceding technology to digital consoles. Also why most programming languages refer to outputting a string to stdout as “print”.
- bitwize 10mo ago
  It was a typewriter ribbon, and the type of terminal it was designed to be used with was a typewriter with communications circuitry, called a "teletypewriter". This is why the controlling terminal of Unix CLI/TUI processes is called a tty or pty (pseudo-tty).
- jibal 10mo ago
  Similar? It is in fact a typewriter ribbon: https://www.amazon.com/Olympia-Typewriter-Ribbon-Black-Red/dp/B08Y33BKHX?th=1 https://www.amazon.com/Olympia-Typewriter-Ribbon-Black-Red/d...
- reddalo 10mo ago
  It's almost the same reason Windows still uses CR LF characters for new lines. Not one character, but two: Carriage Return and Line Feed. Literally the action of moving the printer back to the beginning of the line and then the action of making the sheet of paper go "up" by one line.
  - randallsquared 10mo ago
    That's why those characters exist, but not why Windows uses both: Unix already used LF only, and the Apple II (and Mac, for a while) used only CR. The choice to use both was, as far as I know, Gary Kildall's, in CP/M, and various DOSes including MS-DOS inherited that decision without much examination.
- TZubiri 10mo ago
  Man I'd love to have a computer that just prints stdout to a typewriter. Even if modern linux.
  - bitwize 10mo ago
    Score a GE TermiNet 30 or similar teletypewriting terminal off eBay, hook it to your PC via USB-to-RS232, and Bob's your uncle. You can even do a getty to it, log in, and run shell commands: https://www.youtube.com/watch?v=-Ul-f3hPJQM https://www.youtube.com/watch?v=-Ul-f3hPJQM
bob1029 10mo ago
IBM is the undisputed king of backward compatibility. There is code running on mainframes right now that is going on 50 years old. Microsoft is a close #2 with windows. I'd probably consider using IBM if it wasn't so goddamn weird and expensive. I suppose all that backward compatibility does have its downsides. Windows feels a bit weird in some places too, but at the same time it didn't start out life as a typewriter.
- reddalo 10mo ago
  >Windows feels a bit weird in some places too Windows 11 still has some dialogs that haven't been touched (and they can't ever be, in order to prevent backward compatibility breakage) since Windows 3.1: https://www.windowsonwindows.com/forum/viewtopic.php?t=44 https://www.windowsonwindows.com/forum/viewtopic.php?t=44
  - nottorp 10mo ago
    The ones that work every time, right?
    - duskdozer 10mo ago
      Hey, what's wrong with a little breakage every now and again and again and again, as long as it's new and fresh?
      - wiseowise 10mo ago
        It’s neumorphism, mom, and it’s important!
nye2k 10mo ago
I love this absolute example of old systems interfering with new systems, rewriting old systems. My old man started his tech work on hot rods, then mechanical typewriters, calculators, eventually continuing into mainframe electronics and nearly followed all the transitions up to today’s AI. The number of times I’ve scratched my head at a problem and he had a clear understanding of where the logic broke… based on a historical decision that could not physically be undone.
- fph 10mo ago
  If you try profiling almost any program that does linear algebra (something that uses Numpy, for instance), you will see a lot of calls and CPU time in functions with names like DGETRF or SGESVX. These obscure names stand for stuff like Single-precision GEneral matrix Solve Vector eXtended; i.e., solve a linear system of equations with a full, dense matrix. Why are they so difficult to parse? Couldn't they come up with a friendlier name? They come from Lapack, the standard linear algebra foundation library, which is written in Fortran 77. That library was first written in 1992, when the Fortran 90 standard was still new and not supported everywhere, so they stuck with the earlier version. Lapack has become the standard library for dense non-parallel linear algebra; it is still maintained and updated, but the basic math algorithms haven't changed much, so there was no need to replace it entirely. Today there are also processor-specific libraries like MKL or Apple Accelerate, but they still all follow the same Lapack API. When Fortran-77 was standardized, they decided to keep function names at most 6 letter long, to "ensure portability". I.e., they wanted to support certain compilers and architectures that were already considered old in 1977. TL;DR: if you can't read easily those flame graphs today, it's because of backward compatibility with certain mainframes that probably date back to the late 1960s.
  - AnimalMuppet 10mo ago
    Wasn't there a very similar library earlier than 1992? I seem to recall Linpack back in the early 1980s that sounded very similar.
    - buescher 10mo ago
smiley1437 10mo ago
Strangely, this made me think about the recurrent laryngeal nerve in giraffes. The nerve takes a 15-foot detour down the long neck and loops under the aorta near the heart before it travels back up because evolution needed to stay backwards compatible with previous iterations of protogiraffes as environmental selection pressure lengthened the neck.
- fredley 10mo ago
  I love this fact. If you're a fish with no neck, the route it takes is the most direct and obvious. But as evolution gradually lengthened necks the route remained the same!

82 points by speckx 10mo ago

For those who don’t get it: It’s referring to the ink soaked ribbon that would print characters on a piece of paper, similar to a typewriter. This is a preceding technology to digital consoles. Also why most programming languages refer to outputting a string to stdout as “print”.

bitwize 10mo ago
It was a typewriter ribbon, and the type of terminal it was designed to be used with was a typewriter with communications circuitry, called a "teletypewriter". This is why the controlling terminal of Unix CLI/TUI processes is called a tty or pty (pseudo-tty).
jibal 10mo ago
Similar? It is in fact a typewriter ribbon: https://www.amazon.com/Olympia-Typewriter-Ribbon-Black-Red/dp/B08Y33BKHX?th=1 https://www.amazon.com/Olympia-Typewriter-Ribbon-Black-Red/d...
reddalo 10mo ago
It's almost the same reason Windows still uses CR LF characters for new lines. Not one character, but two: Carriage Return and Line Feed. Literally the action of moving the printer back to the beginning of the line and then the action of making the sheet of paper go "up" by one line.
- randallsquared 10mo ago
  That's why those characters exist, but not why Windows uses both: Unix already used LF only, and the Apple II (and Mac, for a while) used only CR. The choice to use both was, as far as I know, Gary Kildall's, in CP/M, and various DOSes including MS-DOS inherited that decision without much examination.
TZubiri 10mo ago
Man I'd love to have a computer that just prints stdout to a typewriter. Even if modern linux.
- bitwize 10mo ago
  Score a GE TermiNet 30 or similar teletypewriting terminal off eBay, hook it to your PC via USB-to-RS232, and Bob's your uncle. You can even do a getty to it, log in, and run shell commands: https://www.youtube.com/watch?v=-Ul-f3hPJQM https://www.youtube.com/watch?v=-Ul-f3hPJQM

bob1029 10mo ago

IBM is the undisputed king of backward compatibility. There is code running on mainframes right now that is going on 50 years old. Microsoft is a close #2 with windows. I'd probably consider using IBM if it wasn't so goddamn weird and expensive. I suppose all that backward compatibility does have its downsides. Windows feels a bit weird in some places too, but at the same time it didn't start out life as a typewriter.

reddalo 10mo ago
>Windows feels a bit weird in some places too Windows 11 still has some dialogs that haven't been touched (and they can't ever be, in order to prevent backward compatibility breakage) since Windows 3.1: https://www.windowsonwindows.com/forum/viewtopic.php?t=44 https://www.windowsonwindows.com/forum/viewtopic.php?t=44
- nottorp 10mo ago
  The ones that work every time, right?
  - duskdozer 10mo ago
    Hey, what's wrong with a little breakage every now and again and again and again, as long as it's new and fresh?
    - wiseowise 10mo ago
      It’s neumorphism, mom, and it’s important!

nye2k 10mo ago

I love this absolute example of old systems interfering with new systems, rewriting old systems. My old man started his tech work on hot rods, then mechanical typewriters, calculators, eventually continuing into mainframe electronics and nearly followed all the transitions up to today’s AI. The number of times I’ve scratched my head at a problem and he had a clear understanding of where the logic broke… based on a historical decision that could not physically be undone.

fph 10mo ago
If you try profiling almost any program that does linear algebra (something that uses Numpy, for instance), you will see a lot of calls and CPU time in functions with names like DGETRF or SGESVX. These obscure names stand for stuff like Single-precision GEneral matrix Solve Vector eXtended; i.e., solve a linear system of equations with a full, dense matrix. Why are they so difficult to parse? Couldn't they come up with a friendlier name? They come from Lapack, the standard linear algebra foundation library, which is written in Fortran 77. That library was first written in 1992, when the Fortran 90 standard was still new and not supported everywhere, so they stuck with the earlier version. Lapack has become the standard library for dense non-parallel linear algebra; it is still maintained and updated, but the basic math algorithms haven't changed much, so there was no need to replace it entirely. Today there are also processor-specific libraries like MKL or Apple Accelerate, but they still all follow the same Lapack API. When Fortran-77 was standardized, they decided to keep function names at most 6 letter long, to "ensure portability". I.e., they wanted to support certain compilers and architectures that were already considered old in 1977. TL;DR: if you can't read easily those flame graphs today, it's because of backward compatibility with certain mainframes that probably date back to the late 1960s.
- AnimalMuppet 10mo ago
  Wasn't there a very similar library earlier than 1992? I seem to recall Linpack back in the early 1980s that sounded very similar.
  - buescher 10mo ago

smiley1437 10mo ago

Strangely, this made me think about the recurrent laryngeal nerve in giraffes. The nerve takes a 15-foot detour down the long neck and loops under the aorta near the heart before it travels back up because evolution needed to stay backwards compatible with previous iterations of protogiraffes as environmental selection pressure lengthened the neck.

fredley 10mo ago
I love this fact. If you're a fish with no neck, the route it takes is the most direct and obvious. But as evolution gradually lengthened necks the route remained the same!