Proposal3: Make $_ a valid identifier character

Richard A. O'Keefe ok at atlas.otago.ac.nz
Thu Jun 1 01:12:46 UTC 2000


	This is historically wrong. (See eg. "The early history of Smalltalk".)
	Smalltalk did not adopt ascii until ST-80. Before this they used their own
	set, including "eyes" and smiley faces in ST-72.

What Smalltalk*72* used or was able to use is one thing.
The fact remains that the Altos did have an arrow key,
did not have a smiley key, and did not have an underscore key.

	Hence, they could obviously use whatever they wanted.

Not *and* share files with Mesa or Lisp people.
Not *and* share files with Bravo.

	A look at some ST-76 code shows that the
	intercappedStyle was used already then,

Only relevant to the discussion *if* underscore was available.

	I think it is a definite possibilty that underscores came to be used because
	there were no lowercase characters avaiable, only uppercase. (Lowercase?
	Luxury.)
	
While *some* equipment didn't have lower case, Flexowriters did.
Perhaps more interestingly, the American military codeset FIELDATA
had lower case before 1963.

WeDontWriteEnglishThisWaySoWhyShouldWeWriteProgramsThisWay?

Note in particular that the underscore was placed in character sets
to support typewriter-style underlining (ASCII-63 and ASCII-67 are
both paper-type/typewriter codes rather than punched card/line printer codes).
The hypothesis that it has some connection with upper case is falsified by
APL, which uses an up triangle to separate words (LIKE/\THIS), even though
it has an underscore (the up triangle in APL is a single character, the same
width as the letters).





More information about the Squeak-dev mailing list