portability and `echo'

[Chris Torek]

In article <720 at ttrdc.UUCP> levy at ttrdc.UUCP writes:
>In article <3032 at umcp-cs.UUCP>, chris at umcp-cs.UUCP (Chris Torek) writes:
[followed by my sample portable program, namely a simplified
reimplementation of `echo'.]

> What about a system where it is impossible to give command line
> arguments to a program, like the early-stone-age card reading
> IBM systems?

On such a system an `echo' program is useless, since it always
prints a blank line.  However, note that my sample program would
indeed successfully print a blank line on such a system.

> Is C only allowed to run on machines that DO allow command line
> arguments?

No; but such a system requires severe design constraints which
rule out a program like `echo'.  Yet that does not make echo
unportable.

> Would this program just get run with an argc of 0 in the case
> of a stone age system?

As to that I cannot say, but again, my code will still print a
blank line.

> As a matter of fact, must a system support both upper and lower
> case characters to support C?

No.  But it would no doubt break quite a bit of code.  That code
can be considered unportable in that sense; but here I think one must
begin to consider just what extremes of portability are worthwhile.
Would you say that screen editors should not exist because they are
unportable to some machines?

I think the real point is that one should write code as portably
as possible given the constraints under which that code is to
operate.  Window systems cannot run on teletypes; bitmapped graphics
cannot run on character-oriented CRTs.  Yet I can write a portable
routine that does bitmapped graphics using the MIT X window system
as a base.  I claim there is no contradiction here: that portability
does not mean that code runs on any machine anywhere; rather, that
it is not *unnecessarily* limited.
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 1415)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris at umcp-cs		ARPA:	chris at mimsy.umd.edu