[comp.lang.perl] Standard Distribution

srs@ticnj.UUCP (Scott Strool) (02/05/91)

I have been hacking with perl for a few months and find it very useful
and extremely powerful. I try to use it wherever I can, but have a problem
convincing others its purpose. One arguement is that if it is so good
then why isn't it part of a standard UNIX package. Has anyone been interested
in making part of the group of standard UNIX development tools, i.e lex awk
sed etc. One reason management is reluctant to write stuff with perl is
because they have the burden of insuring it is on the target system.
Fo in house stuff its great but for systems that are to be delivered it 
could be trouble.

Comments?

-- 
Scott Strool      	(908) 747-4700		srs@iex.com	
IEX Corporation
125 Half Mile Rd.	
Red Bank, NJ 07701

tchrist@convex.COM (Tom Christiansen) (02/05/91)

From the keyboard of srs@ticnj.UUCP (Scott Strool):
:I have been hacking with perl for a few months and find it very useful
:and extremely powerful. I try to use it wherever I can, but have a problem
:convincing others its purpose. One arguement is that if it is so good
:then why isn't it part of a standard UNIX package. Has anyone been interested
:in making part of the group of standard UNIX development tools, i.e lex awk
:sed etc. One reason management is reluctant to write stuff with perl is
:because they have the burden of insuring it is on the target system.
:Fo in house stuff its great but for systems that are to be delivered it 
:could be trouble.

I understand the dilemma.  My solution was to make it part of the standard
distribution, bundled up along with /bin/cat.  Now all Convex customers
are guaranteed to have it on their Convexen, and anywhere else, since we
ship it with source.  Rick Adams has an amusing comment that the real
reason they did this was because it was easier than arguing with me any
more.  He may have something there. :-)

In any event, we at Convex can (and do) now freely use perl to write
utilities, sysadmin support tools, install scripts, diagnostic scripts,
test suites, plus plenty more, and we can be assured that perl will be
there both for internal use and for customer use.  Since most of these
were written in sh or (ug) csh before, it's truly a breath of fresh air.

I'm absolutely convinced that perl is destined to join that list of tools
you mentioned, like lex, yacc, sed, awk, and sh.  The only question is
when.  If history is any example, it'll show up first on the BSD tape
(probably the fabled 4.4 release).  Shortly thereafter it will come with
Sun distributions.  Or maybe Sun will do so even shortly before BSD if
Kolstad is half the evangelist I am :-).  Didn't I just read somewhere how
some Sun salesman was giving out perl to his customers?

I fear, however, that only after many years will we see it from AT&T, at
least for customer shipments.  Actually, I wouldn't be surprised if some
of the folks sequestered at Murray Hill were already using perl, but the
migration path from Research Version X (and other sources) to System VrY
seems very long:  think how long it took for them to get virtual memory.
This is a problem, because too much of the non-university world thinks
(to their loss, IMHO) that only AT&T makes UNIX.  I don't think we'll
be seeing 1003.2 or 1003.7 taking up perl any time real soon now, so 
POSIX standards won't help.  (Nor would I want to see what POSIX would
do to perl.)

One thing I've been trying to learn is how many other vendors, either
hardware or software (and what size customer base) ship perl with their
product.  Anyone know of any besides us?  

So how to get perl more widely distributed as standard software?  If you
work for a vendor, you could lobby internally for it.  If you're a
customer, you can send in enhancement requests (we call then "wishes"
here) to have perl included.  Maybe if enough people ask, their wish may
be granted.

--tom
--
"Still waiting to read alt.fan.dan-bernstein using DBWM, Dan's own AI
window manager, which argues with you 10 weeks before resizing your window." 
### And now for the question of the month:  How do you spell relief?   Answer:
U=brnstnd@kramden.acf.nyu.edu; echo "/From: $U/h:j" >>~/News/KILL; expire -f $U

bob@MorningStar.Com (Bob Sutterfield) (02/06/91)

This is the age-old problem of technology adoption, and it hits in
every area, not just sysadm tools.  For example, we ship documentation
to our customers as *roff source, which means I must contort my poor
aching brain into writing in *roff, even though LaTeX is so much more
expressive and would yield many more lines of better-quality
documentation per man-hour of writing time.  Why ship *roff source?
Because our customers must be able to print it or browse it on-line
themselves, and not all have *TeX (to make paper) and Emacs (to use
the embedded Info browser).  To pick another example, our support guy
(an Emacs user from way back) forced himself to learn ed, ex, and vi,
because he often dials into customer sites to fix their broken
installations of our products, and not all of them can be counted upon
to have Emacs installed.

Perl faces the same slow migration problem.  To really mangle some
metaphors, there are chickens and eggs involved as well as critical
masses to worry about.  Our company isn't as fortunate in this regard
as Convex, where they define a single platform and must support only
that one target, because our products run on dozens of platforms of
widely varying pedigree.  Hopefully, we can expect modern tools (e.g.
perl, Emacs, *TeX, expect) to begin appearing on enough of our target
systems to make using them worthwhile.  Until then, we're forced to
use what's available on the least of them.

tneff@bfmny0.BFM.COM (Tom Neff) (02/07/91)

Wonderful as Perl is, I'm sure glad we're not stuck with version 2.0
being distributed with everyone's system tapes and disks today, as would
be the case if major vendors had "adopted" it!  Two years is chicken
feed for these guys.  Perl's still changing.  When Larry's tired of
changing it, vendors should and will embrace it.

-- 
"I'm not sure I've even got the brains to   #:#   Tom Neff
 be President." -- Barry Goldwater, 1964    #:#   tneff@bfmny0.BFM.COM

allbery@NCoast.ORG (Brandon S. Allbery KB8JRR) (02/09/91)

As quoted from <1991Feb05.111212.12131@convex.com> by tchrist@convex.COM (Tom Christiansen):
+---------------
| This is a problem, because too much of the non-university world thinks
| (to their loss, IMHO) that only AT&T makes UNIX.  I don't think we'll
+---------------

And too many university types think (to their loss, IMHO) that only BSD makes
*useful* UNIX.  I've seen a few BSD attempts to invent AT&T's sdiff; until
someone on the BSD end finally discovered sdiff and reimplemented it with
non-AT&T code, the only usable effort was Emerge (a package for GNU Emacs).
I've also encountered a number of other useful utilities which are *standard*
in System V and lacking in BSD.  Why else would "dual universe" and "merged"
Unixes be so popular?

Neither AT&T UNIX nor BSD UNIX is complete in and of itself.

+---------------
| POSIX standards won't help.  (Nor would I want to see what POSIX would
| do to perl.)
+---------------

Ugh.  I can just imagine the monstrosity they'd leave us with.

+---------------
| One thing I've been trying to learn is how many other vendors, either
| hardware or software (and what size customer base) ship perl with their
| product.  Anyone know of any besides us?  
+---------------

Perl is one step from being shipped on all the systems Telotech, Inc. sells.
(That one step depends on events in the next few months.)  I already use it
for development scripts, and have a few utility scripts in an accounting/
order entry/inventory control system I've been working on for the past 1 1/2
years.  (Then again, another small but significant part of that system is
written in SBprolog.  Convention be d*mned, I go with what works and does so
in a reasonable amount of time.)

+---------------
| U=brnstnd@kramden.acf.nyu.edu; echo "/From: $U/h:j" >>~/News/KILL; expire -f $U
+---------------

Some time back, I asked Dan "Whose dolly have you busted?" (context: the
replacement moderator for comp.sources.misc).  I guess I know now.  Why?

++Brandon
-- 
Me: Brandon S. Allbery			    VHF/UHF: KB8JRR on 220, 2m, 440
Internet: allbery@NCoast.ORG		    Packet: KB8JRR @ WA8BXN
America OnLine: KB8JRR			    AMPR: KB8JRR.AmPR.ORG [44.70.4.88]
uunet!usenet.ins.cwru.edu!ncoast!allbery    Delphi: ALLBERY

allbery@NCoast.ORG (Brandon S. Allbery KB8JRR) (02/10/91)

As quoted from <BOB.91Feb5111721@volitans.MorningStar.Com> by bob@MorningStar.Com (Bob Sutterfield):
+---------------
| widely varying pedigree.  Hopefully, we can expect modern tools (e.g.
| perl, Emacs, *TeX, expect) to begin appearing on enough of our target
| systems to make using them worthwhile.  Until then, we're forced to
| use what's available on the least of them.
+---------------

One nice thing about PD and even copylefted tools:  if we expect 'em and they
aren't there, we can just take a tape over to the site and install 'em.  Emacs
is a popular candidate for this; Perl is just one step away from it.  And at
least one site is about to get its buggy, broken DWB replaced with TeX.

Of course, it helps that we don't take customers we can't drive to in under an
hour....

++Brandon
-- 
Me: Brandon S. Allbery			    VHF/UHF: KB8JRR on 220, 2m, 440
Internet: allbery@NCoast.ORG		    Packet: KB8JRR @ WA8BXN
America OnLine: KB8JRR			    AMPR: KB8JRR.AmPR.ORG [44.70.4.88]
uunet!usenet.ins.cwru.edu!ncoast!allbery    Delphi: ALLBERY