Re: [TUHS] Any Good dmr Anecdotes?

["ches@Cheswick.com" <ches@cheswick.com>]

[-- Attachment #1: Type: text/plain, Size: 2882 bytes --]

I am a fan of these routines, and use the regularly, but I didn’t write them.

Message by ches. Tappos by iPad.


> On Jul 10, 2018, at 9:50 PM, Noel Hunt <noel.hunt@gmail.com> wrote:
> 
> I'm surprised why anyone would bother with these routines
> anymore, given the startling simplicity of Plan9's arg(3).
> One stands in awe of such simplicity. I believe it was
> William Cheswick who designed it, but I may be wrong.
> 
> 
>> On Tue, Jul 10, 2018 at 5:25 PM <arnold@skeeve.com> wrote:
>> RFS vs. NFS and sockets vs. STREAMS were much more serious; they were
>> about the directions Unix would take going forward, where interoperability
>> (RFS/NFS) and code portability (sockets/STREAMS) were big either/or issues.
>> 
>> Had AT&T been smarter about its licensing, both RFS and STREAMS might
>> have "won", but they weren't, and those technologies have all but
>> disappeared.
>> 
>> GNU getopt can be used in a source-compatible way with POSIX getopt;
>> having long options is up to the programmer.  I agree, there were
>> aesthetic arguments, altough long options have mostly "won".  I'm about
>> as long-time a Unix aficianado as anyone else here, and for many things
>> I find long options easier to remember than short ones.
>> 
>> (To their credit, at least initially, the GNU project asked its developers
>> to use the same long options in all programs for operations that were
>> the same.)
>> 
>> Arnold
>> 
>> 
>> George Michaelson <ggm@algebras.org> wrote:
>> 
>> > ... and then somebody GNUified it. I seem to recall three huge
>> > flamewars in UUCP days: RFS vs NFS, STREAMS (the original) vs sockets,
>> > and getopt
>> >
>> > --no -noo --nooo=please --dont-make-me=do-that
>> >
>> > On Tue, Jul 10, 2018 at 3:54 PM,  <arnold@skeeve.com> wrote:
>> > > Clem Cole <clemc@ccc.com> wrote:
>> > >
>> > >> BY the time dmr adds stdio, it was
>> > >> still early enough in the life to displace the randomness for something as
>> > >> important as I/O, whereas lack of use of something.like getopt would not
>> > >> become clearly deficient until after widespread success.
>> > >
>> > > I think "widespread access" is more like it for getopt.  Getopt dates
>> > > to 1980; it was in System III (I just checked). That's only about two years
>> > > after V7 which was circa 1978.
>> > >
>> > > Here are the dates:
>> > >
>> > > -rw-rw-r-- 1 arnold arnold 1073 Apr 11  1980 usr/src/lib/libc/pdp11/gen/getopt.c
>> > > -rw-rw-r-- 1 arnold arnold 2273 May 16  1980 usr/src/man/man3/getopt.3c
>> > >
>> > > But the world outside the Bell System didn't have System III. Getopt
>> > > didn't become "popular" until System V or so, and became much easier to
>> > > adopt once Henry Spencer published his public domain rewrite of the code
>> > > and man page.
>> > >
>> > > Just a nit, (:-)
>> > >
>> > > Arnold

[-- Attachment #2: Type: text/html, Size: 4183 bytes --]