Hacker News Reader

Comment by robertlagrant | Hacker News Reader

Hot take: @ might've been a good alternative.

I'd argue @ is more suitable for a dereference operation than an "address of" one. It functions that way in an email address too.

Also, then "int @p" would read "integer at p" which is exactly what "int *p" means.

That's precisely what Clojure did, where @foo is syntactic sugar for (deref foo)

So address-of could be an "@" symbol, but upside down.

Or a spanish question mark ¿.

Ah yes - that's true.

Yup. But as I understand it @ wasn't available on the terminals used to create C. Likewise #. They were used for later features, once they could actually be typed. Happy outcome, that there were characters available for fresh features, since C used absolutely everything available on the original (ADM?) terminals.

Another curious thing: Unix came shortly after and was entirely typed with lower case. Because those same terminals had only one font (font ROMs were tiny back then, and expensive). It looked like UPPERCASE but in fact the keyboard produced lowercase. Once better terminals were used to edit the code it was noticed for the first time(?) that it's all be edited as lowercase.

Or so the story went, back in those days, when it was all new.

That corroborates what I've read.

The DEC PDP-11 assembler used the @ symbol for dereferences of registers ("@Rn" or "(Rn)" dereferenced a register, for example).

However, in Unix, the terminal convention was to use the @ symbol to delete the current line, and they didn't have a DEC assembler yet, so they used the asterisk (*) instead in B, as well as in the Unix PDP-11 assembler, which was written in B.

I'm paraphrasing this Quora post (which lists primary and secondary sources): https://www.quora.com/Why-did-the-developers-of-C-decide-to-...

As far as I know @ on modern keyboards above the two from the IBM Selectric[1], released in 1961; the Model D and the VT-52 both copied the keyboard layout. The 1963 ASCII draft had @, despite not having lowercase yet.

Prior to IBM, it was on at least the Underwood typewriters[2], so it was never quite absent.

1: https://en.wikipedia.org/wiki/IBM_Selectric#/media/File:IBM_... Note that the number-row is only different from modern keyboards with the cent-sign being replaced with a caret (the cent-sign was never included in ASCII)

2: https://upload.wikimedia.org/wikipedia/commons/a/a4/Ernest_H...

1963 ASCII was upper-case only, and many early terminals (notably for Unix, including the Teletype 33) were upper case only. `stty lcase`, which maps upper-case input to lower case, existed early — https://www.tuhs.org/cgi-bin/utree.pl?file=V1/man/man2/stty....

Instance-variable heavy Ruby code does look rather like someone has sneezed on the monitor, so I think it would have been a bad move, aesthetically ...

Nethack or code? You decide

At the very least, &/* are in the same place on many keyboards, I've been bitten by muscle memory many times going back and forth between us/uk/fr keyboard layouts.

For reference, it's above the ' and between the ;: and #~ keys on my current board, rather than on shift-2 (where " lives).

Maybe this would have stayed consistent worldwide, if it was used in c.

Pascal uses this syntax.

Also "pointer to integer" is written as "^integer" which is better than "int".

Syntax also doesn't allow confusing stuff like like "int a, b;".

> "^integer" which is better than "int"

I agree that it reads a little better. But as a small-handed person, it is unfortunately much more uncomfortable to type.

Honestly that fully depends on your locale and keyboard layout. So it really shouldn't influence programming language syntax design.

My computer has buttons, and characters printed on it. So I can't really change the distance between characters.

Some dialects of BASIC used @ as a shortcut for the `VarPtr` operation.

My immediate thought as well.

Where’s that value in memory? It’s @ the address.

I agree, big missed opportunity in C’s development.

Which is what Pascal dialects use.

I didn't think @ took on the "at" meaning until email came around.

The oldest trace of @ being used to mean "at" can be traced to typeriters for commerce around 1880, where it was used as "5 apples @ 10 p" meaning "5 apples at 10 pence each."

At least that's what German Wikipedia Claims while the corresponding paragraph in english Wikipedia is short.

https://de.wikipedia.org/wiki/At-Zeichen

https://en.wikipedia.org/wiki/At_sign

The formal name of the Unicode character U+0040 '@' is, in fact, COMMERCIAL AT.

https://en.wikipedia.org/wiki/At_sign#Unicode

It's also the "a commercial" in Quebec French [1]. Despite lacking the accent on the A (à), the intention of the A is to stand in for the single letter French word "à" (at) [2].

[1] https://vitrinelinguistique.oqlf.gouv.qc.ca/fiche-gdt/fiche/... [2] https://www.noslangues-ourlanguages.gc.ca/fr/cles-de-la-reda...

French wikipedia traces the usage of @ to signify "à" (at) back to Renaissance !

More recently (but still 1968, before email): "In ALGOL 68, the @ symbol is brief form of the at keyword; it is used to change the lower bound of an array. For example: arrayx[@88] refers to an array starting at index 88."

RFC 20 shows "@" defined as "commercial at" in 1969. https://www.rfc-editor.org/rfc/rfc20.html . That was two years before Tomlinson used it for email addressing.

EDIT: "Creation of Computer Input in an Expanded Character Set" at https://ejournals.bc.edu/index.php/ital/article/download/292... from 1968, p112, describes it simply as "at".

EDIT #2: "TRANSLATION FROM MONOTYPE TAPE TO GRADE 2 BRAILLE" at p83 of https://archive.org/details/researchbulletin05lesl/page/82/m... from 1964 describes that symbol as the '"at" sign'.

I think that is early enough to say that email did not influence the "at" meaning but rather the other way around.

There are claims it comes from the latin "ad" similar to how & comes from the latin "et" but the etymology is far less certain, with the french à being another possible source.

What I can't find is when it was first used in the US for indicating home-team for sports. E.g. [1] where there is vs. or @ depending on whether it is a home or away game. I suspect it's relatively modern, but not sure how far back it goes.

1: https://www.mlb.com/redsox/schedule/2023-09

What do you think it meant before that?

Arroba:

https://en.wikipedia.org/wiki/Arroba

As for me, I was taught it meant "alias", used when introducing nickname together with real name.

Reply on news.ycombinator.com