4 ms·

The leap between IBM punchcards and VT100 makes no sense. Teletypewriters used 72 characters per line. IBM mainframe line printers used 132 characters per line.

by gok 7y ago

The leap between IBM punchcards and VT100 makes no sense. Teletypewriters used 72 characters per line. IBM mainframe line printers used 132 characters per line. DEC and IBM probably just happened to pick 80 character columns for unrelated reasons (1970s CRT resolution and 1920s printing technology, respectively)

jchw 7y ago
But wait, I thought CRTs don’t really have resolution per se. Especially for horizontal, I’d imagine things are quite flexible. The same might not be true of an NTSC signal, where 80 columns probably would not work terribly well, but why would it matter for 72 vs 80 vs etc.?
- tyingq 7y ago
  VT220s had a 132 column mode. It was readable, albeit not as readable as 80 columns.
  - NeedMoreTea 7y ago
    So did vt100/vt102 - though it may have been an option on vt100, and a reverse video mode. It was only vt52 stuck at 80. It was readable, but not something for your default setting - 10 mins was more than enough!
  - masswerk 7y ago
    However, the character matrix was the same for both modes, just the letter spacing was different. – Did this cause certain characters to bleed one into another at times, or was this just a general readability issue?
    - tyingq 7y ago
      Hard on the eyes: https://imgur.com/IYULtxA https://imgur.com/IYULtxA
      - masswerk 7y ago
        Ah, I see. Apparently, more of a general contrast / density / grey level issue. Also, probably some issues with bleed and character separation, but this may be also due to the photographic exposure. Thanks for linking the image!
- devnulloverflow 7y ago
  don’t really have resolution per se A CRT doesn't have to have resolution in the sense of pixels per inch (although many of them do). However "resolution" is a more general limit on how far away two distinct points can be. A discrete grid of pixels imposes one kind of limit. But another kind of limit is that two slightly fuzzy dots will become indistinguishable if they blur into one another sufficiently.
  - jchw 7y ago
    Yeah, this was my understanding, but I kept it vague since I don’t claim to be any kind of expert. Otoh, I suspect for this reason that the choice of 80 columns particularly was not motivated by CRT limitations. Could be wrong but it’s not like CRT tech was new in the 70s, and we’re talking about monochrome here.
    - jacobush 7y ago
      Most monitors were pretty small. So too many characters per horizontal real estate would have been hard to read.
      - gok 7y ago
        The core claim of this article is that this is not why 80 was picked though
Taniwha 7y ago
It makes lots of sense - as we moved from batch systems to working from terminals (mid to late 70s) some people were editing cards on card punches, while others were editing card images on terminals
- gok 7y ago
  So DEC thought of this as a way to edit IBM cards but IBM didn't? The implicit claim is that somehow the machines for manufacturing punch cards were reused for computer output and that is what terminals were trying to emulate. But surely there were more displays and non-card-sized printers?
  - Taniwha 7y ago
    Actually I was using Burroughs systems (in pre-VT52 days) There certainly were non card size printers - but they scaled linearly with technology But no we didn't repurpose card punches to be terminals, but we did need terminals to edit card images (fortran was 72 characters plus 8 digits of sequence numbers)
- partdavid 7y ago
  Having used such a system: yes.
masswerk 7y ago
The VT100 actually supported both 132 characters and 80 characters per line, in a 9 x 10 and a 10 x 10 character matrix, respectively. (The character matrix in ROM is just 8 x 10, but pixels are stretched horizontally before being displayed. The limiting factor here isn't resolution, but the signal response of the screen phosphor and the circuits involved in its activation. Therefor the horizontal pixel stretch in order to allow the phosphor to reach its maximum activation level. In fact, a monochrome CRT has just an even phosphor coating and there isn't such a thing as hardware pixels or a native resolution. However, if activation phases are too short to allow the phosphor to gain a suitable activation, the display will be dim in those regions and overall unbalanced in brightness.)

by gok 7y ago

jchw 7y ago

But wait, I thought CRTs don’t really have resolution per se. Especially for horizontal, I’d imagine things are quite flexible. The same might not be true of an NTSC signal, where 80 columns probably would not work terribly well, but why would it matter for 72 vs 80 vs etc.?

tyingq 7y ago
VT220s had a 132 column mode. It was readable, albeit not as readable as 80 columns.
- NeedMoreTea 7y ago
  So did vt100/vt102 - though it may have been an option on vt100, and a reverse video mode. It was only vt52 stuck at 80. It was readable, but not something for your default setting - 10 mins was more than enough!
- masswerk 7y ago
  However, the character matrix was the same for both modes, just the letter spacing was different. – Did this cause certain characters to bleed one into another at times, or was this just a general readability issue?
  - tyingq 7y ago
    Hard on the eyes: https://imgur.com/IYULtxA https://imgur.com/IYULtxA
    - masswerk 7y ago
      Ah, I see. Apparently, more of a general contrast / density / grey level issue. Also, probably some issues with bleed and character separation, but this may be also due to the photographic exposure. Thanks for linking the image!
devnulloverflow 7y ago
don’t really have resolution per se A CRT doesn't have to have resolution in the sense of pixels per inch (although many of them do). However "resolution" is a more general limit on how far away two distinct points can be. A discrete grid of pixels imposes one kind of limit. But another kind of limit is that two slightly fuzzy dots will become indistinguishable if they blur into one another sufficiently.
- jchw 7y ago
  Yeah, this was my understanding, but I kept it vague since I don’t claim to be any kind of expert. Otoh, I suspect for this reason that the choice of 80 columns particularly was not motivated by CRT limitations. Could be wrong but it’s not like CRT tech was new in the 70s, and we’re talking about monochrome here.
  - jacobush 7y ago
    Most monitors were pretty small. So too many characters per horizontal real estate would have been hard to read.
    - gok 7y ago
      The core claim of this article is that this is not why 80 was picked though

Taniwha 7y ago

It makes lots of sense - as we moved from batch systems to working from terminals (mid to late 70s) some people were editing cards on card punches, while others were editing card images on terminals

gok 7y ago
So DEC thought of this as a way to edit IBM cards but IBM didn't? The implicit claim is that somehow the machines for manufacturing punch cards were reused for computer output and that is what terminals were trying to emulate. But surely there were more displays and non-card-sized printers?
- Taniwha 7y ago
  Actually I was using Burroughs systems (in pre-VT52 days) There certainly were non card size printers - but they scaled linearly with technology But no we didn't repurpose card punches to be terminals, but we did need terminals to edit card images (fortran was 72 characters plus 8 digits of sequence numbers)
partdavid 7y ago
Having used such a system: yes.

masswerk 7y ago

The VT100 actually supported both 132 characters and 80 characters per line, in a 9 x 10 and a 10 x 10 character matrix, respectively. (The character matrix in ROM is just 8 x 10, but pixels are stretched horizontally before being displayed. The limiting factor here isn't resolution, but the signal response of the screen phosphor and the circuits involved in its activation. Therefor the horizontal pixel stretch in order to allow the phosphor to reach its maximum activation level. In fact, a monochrome CRT has just an even phosphor coating and there isn't such a thing as hardware pixels or a native resolution. However, if activation phases are too short to allow the phosphor to gain a suitable activation, the display will be dim in those regions and overall unbalanced in brightness.)