Using identifiers with more than 7 chars. #$%@
Barry Shein
bzs at bu-cs.UUCP
Mon Mar 17 08:05:27 AEST 1986
[sorry, this got long, skip the numbered parts if you like]
> I agree with this except for the word `FREE'. TANSTAAFL,
> y'know. The net community as a whole pays for distribution of
> software. This shouldn't discourage any generous person from
> submitting the results of his efforts but he does bear some
> responsibility for the cost he incurs by so doing. I think
> it is reasonable to say that software posted to the net
> should be applicable to as many net users (supporters) as
> possible. Thus, software contributed should be as portable
> as possible. If you have a specialised piece of software to
> give away, and it's big, then you can announce it and e-mail
> it to anyone who requests it.
>--
> Peter Kendell <pete at stc.UUCP>
This is one of those reasonable sounding arguments that can end
civilization as we know it. What about Mac and Atari and Amiga
postings? I know, some sites don't subscribe, but that's THEIR choice
(or at least one would hope it is.)
How about a little tolerance, there are any number of things that
come across this network that may not be suited to you or your needs
be they opinions, job offerings, for sale ads or programs. It is
critical that we try our best to just kind of let the unwanted stuff
pass as much as possible and try not to get too excited. Alternatively,
if you are concerned about traffic helping you get useful things surely
there are places to start cutting down on before net.sources!
I think this discussion should be ended, everyone has now heard:
1. There are some sites who cannot compile some of the
sources that come across the net due to long identifiers.
2. Similarly due to other variants in their software.
3. A poster should probably be careful to put exactly
what systems s/he knows the software runs on near the
beginning of the post as a matter of courtesy. On the
other hand, there is some responsibility on the person
who is extracting the software to verify this (eg. a
piece of software could easily depend upon a system bug-fix
which has not been applied at the other site, such is life.)
4. There exist several pieces of free or very low-cost software
to help you port, among them shortc for long->short identifiers
(which seems to work, I used it on hack a while ago successfully)
and of course Doug Gwyn's SYS/V emulation for 4.2bsd. I am sure
there are other useful packages floating around and hints, this
is part of what the software biz is about (like a bunch of #defines
to help with small curses incompatibilities, readdir() et al
routines for earlier SYS/V ports.)
5. Remember, it is often the very existence of a piece of software's
wide popularity on one system that leads to someone else porting
it to another. A great example is GNU/Emacs which first came out
for 4.2bsd only (oh how they moaned) and owing to it's popularity
was subsequently ported to SYS/V (not perfect yet, but getting there,
and the commitment is clear with 17.49.) I mean, first opinions were
that it would *never* run on anything but 4.2bsd, but there you have
it!
There's an old maxim that system programmers use to cover this type
of situation which I think is applicable:
"Stand on each other's shoulders, not each other's toes"
Also loosely translated to:
"If you're not part of the solution, you're part of the problem"
I hope this will end this so far unproductive discussion and you people
would start sending in all your wonderful ports (I run both 4.2 and SYS/V
here at BU, I can use your help and I think I have endeavored to make it
clear that you are always welcome to mine.)
-Barry Shein, Boston University
More information about the Comp.sources.bugs
mailing list