Hacker News Reader

Comment by smnrchrds | Hacker News Reader

I have a tangential question: why were × and ÷ not included in keyboards and ASCII standard?

It would make sense for them to be. For one, they already have + - = < > % etc. It would make sense to add signs as fundamental and as frequently-used as × and ÷.

Also, some of the signs that were included do not seem to be remotely as important. Do we really need 3 different quotation signs (" ' `), 3 bracket styles ({} [] ()), 2 different newline characters (CR and LF), and various somewhat obscure symbols and many more non-printable characters? It doesn't seem like the address space was too small and they had to make hard choices about which symbols to include. It looks like they had all the capacity they could have wanted, yet decided not to include × and ÷.

Does anyone know why?

View on HN

ASCII was designed for typewriters, not video terminals.

"×" is the same as "x". "÷" can be made by overstriking ":" and "-". The quotation marks needed to also function as accents, by overstriking: "ë é è".

CR and LF are separate operations on a typewriter. You might want to simulate bold type by sending CR and retyping part (or all) of a line.

[ and ] were added for the benefit of Algol, and other programming languages with an Algol-derived syntax. Though it was recognized that these characters were less essential, and these code points were left undefined in ISO 646, the international version of ASCII. This is also why the brackets were put in the alphabet part of the ASCII table, in other countries these code points would've been used for letters that weren't part of the English alphabet.

It was anticipated that some manufacturers would use a 6-bit subset of ASCII. (Even 5-bit and 4-bit subsets were considered.) This would mean that the lower case block would be collapsed on the upper case block. So they needed to find characters that could be useful but would still make sense when rendered as [ and ]. Hence { and }.

Hope this helps. A lot of thought went into ASCII, but many of the considerations became obsolete during the 70s. And you could say that any terminal that does not support overstriking, does not properly support ASCII :)

> It was anticipated that some manufacturers would use a 6-bit subset of ASCII.

Not only anticipated but quite common on 36-bit machines where you could put six characters in a word. In those days “byte” meant a contiguous string of bits so a byte pointer could point to one, two, ... 63 bits in some architectures so SIXBIT characters were not at all inconvenient

> You might want to simulate bold type by sending CR and retyping part (or all) of a line.

Some video displays (“glass teletypes” — these took character data and were not pixel addressable) supported this feature. Printers pretty much all did — tolerances were not great so overprints were always darker. Backspace (^H) was used for this too. This is one reason why backspace and delete have been different characters to me.

Note that the 0xFF value for delete goes back to paper tape: it punched out all the holes in a column so could overwrite any other character. Yes, you deleted by doing backspace delete.

The character × is not universally used for multiplication - in various places in Europe we use . for multiplication, and not the · character, literally . as in 12.34 = 408. The only time × is used is for vector cross-product. And similarly we don't use ÷, we use :. I always thought * and / were a good compromise - the former for looking just like a fat dot, and the latter because it looks like a fraction written on a single line.

I'm guessing it's carry over from the typewriter days. Lower case x would double as the multiply symbol, and over-striking dash and colon would approximate the divide symbol.

Well, I guess we should be happy that 1 and l are separate characters then.

Indeed many typewriters lacked 1 and 0,; you just used l and O

I'd bet that people were already uses to * and / instead of × and ÷ due to their existence in 64 bit codes.

https://en.wikipedia.org/wiki/Six-bit_character_code

Reply on news.ycombinator.com