From list Thu Dec  2 04:26:23 1993
Received: by feenix.metronet.com id AA21985
  (5.65c/feenix for wjm); Thu, 2 Dec 1993 04:22:44 -0600
Received: by feenix.metronet.com id AA21977
  (5.65c/feenix for perl-packrats); Thu, 2 Dec 1993 04:22:42 -0600
Return-Path: <wjm>
Message-Id: <199312021022.AA21977@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Thu, 2 Dec 1993 04:22:41 -0600
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: perl-packrats
Subject: informal setup of perl archivists list 
Status: OR



  Greetings.  I've setup a simple procmail recipe to handle this
list to start.  I will happily add the functionality of majordomo,
or some other package, if needed.

  Mail to perl-packrats@metronet.com will deliver to the following
initial subscribers:


Tom Christiansen <tchrist@wraeththu.cs.colorado.edu>
Henk Penning <henkp@cs.ruu.nl>
jhi@snakemail.hut.fi (Jarkko Hietaniemi)
spp@dragonfly.cis.ufl.edu (Stephen P. Potter)
Lee McLoughlin <lmjm@doc.ic.ac.uk>
Mark <mark@blackplague.gmu.edu>     <- formerly of coombs
Bill Middleton <wjm@metronet.com>


  Does anyone care to offer opening commentary, suggestions, or guidelines?



Bill







From list Thu Dec  2 05:08:45 1993
Received: by feenix.metronet.com id AA22632
  (5.65c/feenix for wjm); Thu, 2 Dec 1993 05:08:00 -0600
Received: from blackplague.gmu.edu by feenix.metronet.com with SMTP id AA22628
  (5.65c/feenix for <perl-packrats@metronet.com>); Thu, 2 Dec 1993 05:07:57 -0600
Return-Path: <mark@blackplague.gmu.edu>
Received: by blackplague.gmu.edu (NX5.67d/NX3.0M)
	id AA11929; Thu, 2 Dec 93 06:10:41 -0500
From: Mark <mark@blackplague.gmu.edu>
Message-Id: <9312021110.AA11929@blackplague.gmu.edu>
Subject: informal status of the coombs archive...
To: perl-packrats@metronet.com
Date: Thu, 2 Dec 1993 06:10:41 -0500 (EST)
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1767      
Status: OR

Hi guys,

I guess you're wanting to know the real story about the coombs archive..me
too :) As I said earlier it was accidentally deleted coz I was in the 'guest'
group when all guest accounts were nuked.. fair enuff... anyway Ive been
calling, emailing and whatnot since then (a few months ago now) trying to 
get the sys admin to A) contact him, and B) get the account back/get the
files... I know he will eventually do it but for some reason he's dev/nulling
my calls/email.. *shrug* I guess he's gotten moody about something Im rumoured
to have done, he tends to ignore those he's upset with. (No idea whats up, just
guessing)

Anyway my game plan is to physically travel interstate and talk to him and at
the very least get a tar file from it all... the actual archive and the 
comparably sized work dir with yet-to-be-added scripts. Im wondering if email
from people concerned about the archive to him might speed him up but then it
might annoy him too.. he's a nice guy and all, just rather quiet at the moment
in respect to answering me.

Anyway, this is purely for your info, dont spread this letter around as it
might cause people to get the wrong idea about the anu people, they are just
over worked and this isnt a huge priority.

When I get the files back I'll be looking at USA sites to have the archive
on I think, more central and less load on the aussie link. It'd probably
mean having an account on the machine with the archive so I could maintain it
so keep that in mind if you're offering space as some of you have. Thats 
a bit aways in the future anyway.

Hope this clears things up a bit. Tom, can you edit out the coombs references
in the FAQ until this is settled please? I still get about 4-5 emails a week
asking about it. :/

Thanks,
Mark

From spp@dragonfly.cis.ufl.edu Thu Dec  2 03:06:19 1993
Received: from dragonfly.cis.ufl.edu by feenix.metronet.com with SMTP id AA27138
  (5.65c/feenix for <wjm@feenix.metronet.com>); Thu, 2 Dec 1993 07:06:07 -0600
Return-Path: <spp@dragonfly.cis.ufl.edu>
Received:  by dragonfly.cis.ufl.edu (5.61ufl/4.12)
	id AA24545; Thu, 2 Dec 93 08:06:19 -0500
From: "Stephen P. Potter" <spp@dragonfly.cis.ufl.edu>
Message-Id: <9312021306.AA24545@dragonfly.cis.ufl.edu>
To: wjm@feenix.metronet.com (Bill Middleton)
Subject: Re: informal setup of perl archivists list 
In-Reply-To: Some random ramblings on Thu, 02 Dec 93 04:31:20 -0600.
Organization: Department of Impossible Probably Facts
Reply-To: Stephen P Potter <spp@cis.ufl.edu>
Date: Thu, 02 Dec 93 08:06:19 EST
Status: O

Strange sunspot activity caused wjm@feenix.metronet.com (Bill Middleton) to wri
te:
| 
| On Dec 2,  4:22am, Bill Middleton wrote:
| } Subject: informal setup of perl archivists list
| } 
| } spp@dragonfly.cis.ufl.edu (Stephen P. Potter)
| 
| This address didn't work for you Stephen.  I've changed it to cis.ufl.edu.
| 

What error message did you get?  Although cis.ufl.edu is preferred, mail to
any machine on the cis network should work.  Unless that machine was down
when you sent it.

Thanks for adding me,
Steve

From lerami!altai!permian!permian!kevin@utacfd.uta.edu Wed Dec  1 17:10:39 1993
Received: from utacfd.uta.edu by feenix.metronet.com with SMTP id AA01167
  (5.65c/feenix for <root@feenix.metronet.com>); Thu, 2 Dec 1993 08:04:44 -0600
Return-Path: <lerami!altai!permian!permian!kevin@utacfd.uta.edu>
Received: by utacfd.uta.edu (/\==/\ Smail3.1.25.1 #25.1)
	id <m0p5EXH-000JSGC@utacfd.uta.edu>; Thu, 2 Dec 93 07:57 CST
Received: by lerami.lerctr.org (/\==/\ Smail3.1.28.1 #28.3)
	id <m0p5DRC-000M9HC@lerami.lerctr.org>; Thu, 2 Dec 93 06:47 CST
Message-Id: <m0p5DRC-000M9HC@lerami.lerctr.org>
Received: by altai.com (1.65/waf)
	via UUCP; Wed, 01 Dec 93 23:20:44 CST
	for root@feenix.metronet.com
Received: by permian.lerctr.org (1.65/waf)
	via UUCP; Wed, 01 Dec 93 22:10:44 CST
	for root@feenix.metronet.com
Date: Wed, 1 Dec 93 22:10:39 EST
From: kevin@permian.lerctr.org (Kevin Bomar)
Subject: Re: Test email
Reply-To: kevin@permian.lerctr.org
To: root@feenix.metronet.com
Lines: 16
X-Mailer: Alamito Mail and News Manager (V1.1)
Status: O

>Just a test
>srl
>
Hello again.

This may be a duplicate message.  If not, thanks for all the help
getting me set up.  I'd like to get some news flowing so I can quit
fooling with this for a while.  If you can add alt.politics.*  I'll
let that go for a couple of days, and the add some more.

thanks,
-kevin
--
----------------------------------------------------------------
Kevin Bomar                 | e-mail: kevin@alamito.metronet.com

From list Thu Dec  2 08:10:37 1993
Received: by feenix.metronet.com id AA01373
  (5.65c/feenix for wjm); Thu, 2 Dec 1993 08:09:45 -0600
Received: from swan.doc.ic.ac.uk by feenix.metronet.com with SMTP id AA01273
  (5.65c/feenix for <perl-packrats@metronet.com>); Thu, 2 Dec 1993 08:07:35 -0600
Return-Path: <lmjm@doc.ic.ac.uk>
Received: from doc.ic.ac.uk by swan.doc.ic.ac.uk id <17044-0@swan.doc.ic.ac.uk>;
          Thu, 2 Dec 1993 14:07:09 +0000
From: Lee McLoughlin <lmjm@doc.ic.ac.uk>
Date: Thu, 2 Dec 1993 14:07:05 +0000
In-Reply-To: Mark <mark@blackplague.gmu.edu> "informal status of the coombs archive..." (Dec 2, 6:10am)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: Mark <mark@blackplague.gmu.edu>, perl-packrats@metronet.com
Subject: Re: informal status of the coombs archive...
Message-Id: <"swan.doc.i.052:02.11.93.14.07.12"@doc.ic.ac.uk>
Status: O

My mirror of the coombs scripts is still intact.  As mirror couldn't get
into the directory /pub/perl it hasn't done any deletions.  Have a look
in:
        src.doc.ic.ac.uk:computing/programming/languages/perl/coombs-scripts
	(aka packages/perl/coombs-scripts)

-- 
--
Lee McLoughlin.                          Phone: +44 71 589 5111 X 5085
Dept of Computing, Imperial College,     Fax: +44 71 581 8024
180 Queens Gate, London, SW7 2BZ, UK.    Email: L.McLoughlin@doc.ic.ac.uk

From list Thu Dec  2 08:16:52 1993
Received: by feenix.metronet.com id AA01914
  (5.65c/feenix for wjm); Thu, 2 Dec 1993 08:16:01 -0600
Received: from crane.cis.ufl.edu by feenix.metronet.com with SMTP id AA01908
  (5.65c/feenix for <perl-packrats@metronet.com>); Thu, 2 Dec 1993 08:15:53 -0600
Return-Path: <spp@crane.cis.ufl.edu>
Received:  by crane.cis.ufl.edu (5.61ufl/4.12)
	id AA05767; Thu, 2 Dec 93 09:15:42 -0500
Message-Id: <9312021415.AA05767@crane.cis.ufl.edu>
To: "Perl Packrats,the Archivers" <perl-packrats@metronet.com>
Subject: Re: informal status of the coombs archive... 
In-Reply-To: Some random ramblings on Thu, 02 Dec 1993 14:07:05 +0000.
Organization: Department of Impossible Probably Facts
Reply-To: Stephen P Potter <spp@cis.ufl.edu>
Date: Thu, 02 Dec 1993 09:15:41 EST
From: Stephen P Potter <spp@crane.cis.ufl.edu>
Status: O

Strange sunspot activity caused Lee McLoughlin <lmjm@doc.ic.ac.uk> to write:
| My mirror of the coombs scripts is still intact.  As mirror couldn't get
| into the directory /pub/perl it hasn't done any deletions.  Have a look
| in:
|         src.doc.ic.ac.uk:computing/programming/languages/perl/coombs-scripts
| 	(aka packages/perl/coombs-scripts)
| 

Also, assuming that script is a good mirror, I've got it at
ftp.cis.ufl.edu:/pub/perl/coombs.  I've also got the mirror from the Hebrew
University in Jerusalem.  Once I get everything sorted out (get rid of all
the duplicates, I'll probably just drop the site names and make one big
archive.  The administrator of the HUJI archive tells me that his is
stable, and that he would rather I didn't mirror it.

Steve

From list Fri Dec  3 02:35:06 1993
Received: by feenix.metronet.com id AA00117
  (5.65c/feenix for wjm); Fri, 3 Dec 1993 02:30:30 -0600
Received: from blackplague.gmu.edu by feenix.metronet.com with SMTP id AA00101
  (5.65c/feenix for <perl-packrats@metronet.com>); Fri, 3 Dec 1993 02:30:25 -0600
Return-Path: <mark@blackplague.gmu.edu>
Received: by blackplague.gmu.edu (NX5.67d/NX3.0M)
	id AA15167; Fri, 3 Dec 93 03:33:06 -0500
From: Mark <mark@blackplague.gmu.edu>
Message-Id: <9312030833.AA15167@blackplague.gmu.edu>
Subject: Weird...
To: perl-packrats@metronet.com
Date: Fri, 3 Dec 1993 03:33:06 -0500 (EST)
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 680       

re guys,

Never rains but it pours huh... anyway I talked to one of the sysadmins there
and he put me onto another guy and also the original sysadmin called me from
his sick bed today and everyone is telling me at once that the files will be
back rsn. Ive emailed the guy responsible for the backups and Im expecting 
him to do a restore when he has time. Looks like it's all coming back soon.

I'll run off and back up the uk mirror now incase it decides to do something
weird before I can get to the coombs files to fix any problems.

Now as for that USA site... any suggestions? Might be worth figuring out some
details in that area now. Close to or in CA would be nice.

Mark

From perl-packrats Mon Dec  6 10:17:25 1993
Received: by feenix.metronet.com id AA16068
  (5.67a/feenix for wjm); Mon, 6 Dec 1993 04:17:27 -0600
Return-Path: <perl-packrats>
Received: by feenix.metronet.com id AA16064
  (5.67a/feenix for perl-packrats); Mon, 6 Dec 1993 04:17:26 -0600
Message-Id: <199312061017.AA16064@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Mon, 6 Dec 1993 04:17:25 -0600
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: perl-packrats
Subject: plots, plans, dreams, and schemes



  I just finished cross-referencing my archive with what's on the
current coombs list.  I'll be mirroring what i don't have within
a few days.  Then i'll find a home for the stuff in my "waiting area",
which i try to do every now and then.  It's rather different, running
an archive as a gopher heirarchy, simultaneously with ftp access.
This is the first time i've gone thru and created a comprehensive
master list of the entire tree.

  Which leads in to some other potential discussion areas, which
i'd really like to know y'all's thoughts about.  Maybe we don't
need any of this stuff, or maybe not.  Of course, it'd mean some
serious coordination, but i think we could pull it off.  And
automation is certainly thinkable too, once the processes/definitions
are given.


- Archiving (in general)
    a. mirroring, updating mirror sites
    b. versions and version control (pointers to actual sites)
    c. usability, portability of items
    d. testing 
    e. reinventing the wheel (other scripts do the same thing)

- Indexing
    a. general index record format - naming, author?, version, master archive
    b. descriptions - current one-liners or (possible) abstract format 
    c. more?

- Classification
   a. scripts 
   b. libraries
   c. info/announcement files
   d. instructional postings (hi Tom!)
   e. discussion and explanations 
   f. others

- gopher, WWW, and mailserver access 
    a. heirarchical structure for an archive (gopher)
    b. http WWW documents for what purposes?
    c. mailserver request format (common?)
    d. waisindexing, other search engines?
    e. other retrieval methods?

  
  Mark, i'll be happy to create you an account here for the purpose of
coordinating on this, and once you get coombs back, you could mirror
whatever we decide to do here and elsewhere.  Could someone get 
Tim Bunce and other's interested possibly?  Tom, maybe a posting
from you would encourage others to join us.


Bill

From perl-packrats Mon Dec  6 13:04:34 1993
Received: by feenix.metronet.com id AA21265
  (5.67a/feenix for wjm); Mon, 6 Dec 1993 07:04:16 -0600
Return-Path: <perl-packrats>
Received: from dragonfly.cis.ufl.edu by feenix.metronet.com with SMTP id AA21261
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 6 Dec 1993 07:04:13 -0600
Received:  by dragonfly.cis.ufl.edu (5.61ufl/4.12)
	id AA06702; Mon, 6 Dec 93 08:04:35 -0500
From: "Stephen P. Potter" <spp@dragonfly.cis.ufl.edu>
Message-Id: <9312061304.AA06702@dragonfly.cis.ufl.edu>
To: "Perl Packrats,the Archivers"	<perl-packrats@metronet.com>
Cc: Jared_Rhine@hmc.edu
Subject: Canonical Archive
Date: Mon, 06 Dec 93 08:04:34 EST

If he isn't already on packrats, someone might want to send Jared the
information....   I am forwarding a message from him about canonicalizing
the archives.  This is definite discussion material, and something I would
definitely be interested in pursuing further.  I'm a bit busy the next
couple of days, so I'll have to get back to this later in the week.

Steve

------- Forwarded Message

Date: Mon, 6 Dec 1993 02:03:16 -0800
Message-Id: <199312061003.CAA12657@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
Subject: Archive

I congratulate on your effort to create a new Perl archive site.  I too,
have begun to create a comprehensive Perl archive.  It is this about which I
wish to speak.  As you no doubt have noticed, there are far too many
different structures of Perl archives out there and none are canonical.  I
propose that we cooperate to create a unified structure, much like the CTAN
project which has managed to create a collection of canonical sites for TeX
(which is equally as expansive and complex as a Perl archive -- the total
archives are ~400 Mb, I believe; our site is also in the process of becoming
a CTAN site, but that's a side note).

The advantages to the Perl community would be numerous, I imagine.  There
would be no searching for the correct random directory to find something,
and users would not have to go to multiple sites looking for the particular
bit they are looking for.  With proper configuration of Mirror (I'm using
2.1), it will take almost no effort to keep our sites in sync.  Cooperating
on the archive like this will also half the work for each of us.  Fame,
fortune, all this will be ours! 8^) I'm sure we could each come up with
excellent, independent archives on our own, but why torture both ourselves
and the Perl community this way?

Would you like to do this?  We will need to work some things out (like your
.message, which claims the title of the North American Archive 8^), but
nothing insurmountable, I imagine.  I'm reasonable, and willing to talk.

- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             | To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- --


------- End of Forwarded Message

From mark@blackplague.gmu.edu Tue Dec  7 05:11:02 1993
Received: from blackplague.gmu.edu by feenix.metronet.com with SMTP id AA05991
  (5.67a/feenix); Mon, 6 Dec 1993 23:08:10 -0600
Return-Path: <mark@blackplague.gmu.edu>
Received: by blackplague.gmu.edu (NX5.67d/NX3.0M)
	id AA27325; Tue, 7 Dec 93 00:11:03 -0500
From: Mark <mark@blackplague.gmu.edu>
Message-Id: <9312070511.AA27325@blackplague.gmu.edu>
Subject: Re: plots, plans, dreams, and schemes
To: wjm@feenix.metronet.com (Bill Middleton)
Date: Tue, 7 Dec 1993 00:11:02 -0500 (EST)
Cc: perl-packrats@feenix.metronet.com
In-Reply-To: <199312061017.AA16064@feenix.metronet.com> from "Bill Middleton" at Dec 6, 93 04:17:25 am
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1122      

>  Mark, i'll be happy to create you an account here for the purpose of
>coordinating on this, and once you get coombs back, you could mirror
>whatever we decide to do here and elsewhere.  Could someone get 
>Tim Bunce and other's interested possibly?  Tom, maybe a posting
>from you would encourage others to join us.

Well I talked to the person responsible for the archives at coombs and he 
tells me that my home dir is fine but the scratch dir which housed 10 megs
of yet-to-be-added scripts, plus the ftp directory are gone. I guess Ive
lost about 12 hours of work or more as at the time the dir's disappeared I 
was literally working full time updating the archive. *ob grumble*

Anyway Ive decided to cut my losses and take the archive to somewhere in
the USA, and seeing how metronet.com already has a good archive plus seems
to be offering me shell access it looks like that might become the defacto
script site... I'll talk to Bill.

Coombs is no more a perl archive. Please remove references from faq's and 
README's and maybe dir structures... time to start another one more central
to the net.

Thanks,
Mark

From perl-packrats Tue Dec  7 05:11:02 1993
Received: by feenix.metronet.com id AA05996
  (5.67a/feenix for wjm); Mon, 6 Dec 1993 23:08:13 -0600
Return-Path: <perl-packrats>
Received: from blackplague.gmu.edu by feenix.metronet.com with SMTP id AA05991
  (5.67a/feenix); Mon, 6 Dec 1993 23:08:10 -0600
Received: by blackplague.gmu.edu (NX5.67d/NX3.0M)
	id AA27325; Tue, 7 Dec 93 00:11:03 -0500
From: Mark <mark@blackplague.gmu.edu>
Message-Id: <9312070511.AA27325@blackplague.gmu.edu>
Subject: Re: plots, plans, dreams, and schemes
To: wjm@feenix.metronet.com (Bill Middleton)
Date: Tue, 7 Dec 1993 00:11:02 -0500 (EST)
Cc: perl-packrats@feenix.metronet.com
In-Reply-To: <199312061017.AA16064@feenix.metronet.com> from "Bill Middleton" at Dec 6, 93 04:17:25 am
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1122      

>  Mark, i'll be happy to create you an account here for the purpose of
>coordinating on this, and once you get coombs back, you could mirror
>whatever we decide to do here and elsewhere.  Could someone get 
>Tim Bunce and other's interested possibly?  Tom, maybe a posting
>from you would encourage others to join us.

Well I talked to the person responsible for the archives at coombs and he 
tells me that my home dir is fine but the scratch dir which housed 10 megs
of yet-to-be-added scripts, plus the ftp directory are gone. I guess Ive
lost about 12 hours of work or more as at the time the dir's disappeared I 
was literally working full time updating the archive. *ob grumble*

Anyway Ive decided to cut my losses and take the archive to somewhere in
the USA, and seeing how metronet.com already has a good archive plus seems
to be offering me shell access it looks like that might become the defacto
script site... I'll talk to Bill.

Coombs is no more a perl archive. Please remove references from faq's and 
README's and maybe dir structures... time to start another one more central
to the net.

Thanks,
Mark

From perl-packrats Tue Dec  7 19:39:50 1993
Received: by feenix.metronet.com id AA06037
  (5.67a/feenix for wjm); Tue, 7 Dec 1993 13:39:59 -0600
Return-Path: <perl-packrats>
Received: by feenix.metronet.com id AA06033
  (5.67a/feenix for perl-packrats@metronet.com); Tue, 7 Dec 1993 13:39:51 -0600
Message-Id: <199312071939.AA06033@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Tue, 7 Dec 1993 13:39:50 -0600
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: CTAN@shsu.edu
Subject: Can we get some advice?
Cc: perl-packrats@metronet.com


George and other respectable members of CTAN,
  
  Us perl archivists are trying to coordinate and plan how to implement
a decent archiving scheme for perl things.  Would y'all care to send
some advice?  We'd like to get a decent mirroring scheme going, and
generally keep up with versions.  My archive here is more schematically 
laid out for gopher retrievals, but the general format for archive is yet not 
decided firmly.  

  Have you folks done any sort of guide?  


Thanks very much, and please send replies to perl-packrats@metronet.com.


Bill Middleton





From perl-packrats Tue Dec  7 23:18:03 1993
Received: by feenix.metronet.com id AA26039
  (5.67a/feenix for wjm); Tue, 7 Dec 1993 17:17:54 -0600
Return-Path: <perl-packrats>
Received: from sun2.nsfnet-relay.ac.uk by feenix.metronet.com with SMTP id AA26022
  (5.67a/feenix for <perl-packrats@metronet.com>); Tue, 7 Dec 1993 17:17:48 -0600
Via: uk.ac.aston; Tue, 7 Dec 1993 23:13:36 +0000
Received: from ftp.tex.ac.uk by email.aston.ac.uk with SMTP (PP) 
          id <03444-0@email.aston.ac.uk>; Tue, 7 Dec 1993 23:12:27 +0000
Received: by ftp.tex.ac.uk (4.1/SMI-4.1) id AA06495; Tue, 7 Dec 93 23:18:03 GMT
Date: Tue, 7 Dec 93 23:18:03 GMT
From: spqr@ftp.tex.ac.uk (Sebastian Rahtz)
Message-Id: <9312072318.AA06495@ftp.tex.ac.uk>
To: CTAN@SHSU.edu
Cc: perl-packrats@metronet.com
Subject: Re: Can we get some advice?
In-Reply-To: <199312071939.AA06033@feenix.metronet.com>
References: <199312071939.AA06033@feenix.metronet.com>
Sender: spqr@ftp.tex.ac.uk

a guide to archiving? sadly not. we run on luck and spit, mostly. our
tools consist of:
 - an agreed structure to put new things in. this is crucial. even the
   three of us misinterpret it at times. you must have an unamiguous
   place to stash a new goody
 - moving towards ISO9960 directory structure. forget you are Unix :-}
 - Joachim Schrod's version of the Lee M's `mirror' script. it works.
 - a script to allow each of us to install things on all 3 CTANs
 - enough tolerance to keep us sane
 
does this help?

sebastian

From perl-packrats Fri Dec 10 05:30:41 1993
Received: by feenix.metronet.com id AA26385
  (5.67a/feenix for wjm); Thu, 9 Dec 1993 23:30:54 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA26366
  (5.67a/feenix for <perl-packrats@metronet.com>); Thu, 9 Dec 1993 23:30:50 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id VAA14097; Thu, 9 Dec 1993 21:30:41 -0800
Date: Thu, 9 Dec 1993 21:30:41 -0800
Message-Id: <199312100530.VAA14097@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@metronet.com, spp@dragonfly.cis.ufl.edu
Subject: CPAN archive

Recently, I proposed to Stephen Potter that some effort be made in
integrating some of the perl archive sites and creating the Perl equivalent
of the Comprehensive TeX Archive Network (CTAN), a collection of sites which
maintain the same structure and same information in a well-organized,
coordinated, documented way.  CTAN has been a tremendous boon to the TeX
community and I think the Perl equivalent would be similarly useful to the
Perl community.

I have received no correspondence regarding that proposal, so I thought I
might further expand on the topic with a few comments.  Please reply, if
only to tell me it is a stupid idea.

The benefits of a canonical structure are numerous, I think.

   * Archive users need only learn a single structure, making access more
     enjoyable and more likely to yield sucess.

   * It distributes the responsibilty of maintenance among multiple
     maintainers (very important, in my mind; I'm a busy guy and as much as
     I would like to devote all of my time to maintaining an archive,
     reality is not likely to let that happen).

   * It collects all the relevant information on a topic together in one
     place, meaning that more comprehensive coverage is given to each topic.

   * Since all the information is in one place, it is much easier to index
     and document.

   * It allows for more effort to be put into the documentation of the
     archive and its scripts.

   * Multiple canonical sites reduce the load on each individual server.  

   * Additional maintainer effort can be invested in creating access points
     via other information systems access points, such as Web and gopher.

Florida is to be commended for their efforts and has done an excellent job
in beginning the process.  I would like to make an initial proposal for a
directory structure, based on that structure.

   archives : For collected bodies of perl references, such as
              comp.lang.perl postings.  This directory
   announce : announcements of new perl packages and administrative perl
              messages.
   documentation : Text designed to serve as reference material or
                   tutorials for perl-related topics
      reference : For material geared towards being reference material
      tutorials : For material geared towards beginning and teaching people
                  how to use perl
      ora : Documentation stemming from the O'Reilly series
   incoming : for submissions of all types
   misc : Miscellaneous, by definition.  This should be kept as small as
          possible.
   scripts : For source code written in perl (note, the structure below is
             likely not the ultimately desired one; I just threw something
             together as a basis for discussion.  I think it is useful to
             adopt subdirectories for the scripts directory, since it is
             quite large.  Most often, archive users are interested in a
             particular kind of script.  Also, in my humble opinion, I don't
             think directories specifically for gods like Muir and Tchrist
             are appropriate.  I _rarely_ want to find a tchrist script,
             rather I would want to find a man replacement and if tchrist
             happens to have written one, all the better.  This may be a
             topic of some debate, however).
      sysadmin : For general system-adminstration oriented scripts
      networking : For scripts that deal primarily with the network
      gateways : For programs that function as gateways, such as ftp
      text-proc : For text processing programs, including convertors
      mail-news : Programs which process or deal with mail
      other : catchall.  Should be kept as small as possible
      file-utils : File utilities
      development : software development tools (perl or otherwise)
   implementations : For various incarnations of the perl language
      perl4
      perl5
      ntperl
      sqlperl

Please note this is intended to be only a starting point for discussion.  A
good structure is the keystone of a cooperative archive effort.  I would
like to discuss how to integrate having a cooperative effort archive network
with still retaining some local flexibility as to the content and structure.
I am certainly not suggesting the all perl archives either convert or
perish; I don't think that would be healthy.  How can these issues be
balanced?

I would also propose that 'gzip --best' be adopted as the quasi-standard
archive format.  Member archive sites should be, of course, free to store
the information as they wish, but in general, I think it is a good option,
since it reduces transfer time.  Most sites probably run wu-ftpd in any case
so that archive clients will be able to access files in their preferred
format, regardless of the format in which it is stored on the archive.

I think adoption of Lee McLoughlin's mirror.pl would be an excellent way to
coordinate the activities of the archive network.  It is flexible enough
that it could be configured to keep sites up to date with minimal maintainer
effort.  Updates made at one site could propagate to the other sites in a
relatively short time.  It is only proper that we use a perl program to do
the mirroring 8^)

One final comment.  Although I don't wish to be negative, I feel I must
register protest against ftp.cis.ufl.edu's widely advertised name of ~the~
North American Perl Archive.  I don't feel that the various archives should
be in some sort of competition to be the best.  I feel that this name shows
disrespect to maintainers of other Perl archives in North America who, I
imagine, have put a considerable among of effort into developing their
archives for the benefit of the Perl community, and not so they could be
called the best.

From perl-packrats Fri Dec 10 06:14:01 1993
Received: by feenix.metronet.com id AA29619
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 00:14:08 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA29614
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 10 Dec 1993 00:14:05 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id WAA16218; Thu, 9 Dec 1993 22:14:01 -0800
Date: Thu, 9 Dec 1993 22:14:01 -0800
Message-Id: <199312100614.WAA16218@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@metronet.com
Subject: CTAN and CPAN
References: <199312100544.AA28327@feenix.metronet.com>

Bill> Here's the 2 mailings you missed.  (Not very active so far)

Ok, thanks.  I'm pleased to see some thought had already gone into this area.
I _thought_ that it was pretty strange that nobody had suggested
consolidating the archives...

When Bill asked CTAN about how they managed to do it, Sebastian suggested
the following things were important:

Sebastian>  - an agreed structure to put new things in. this is
Sebastian>  crucial. even the three of us misinterpret it at times. you must
Sebastian>  have an unamiguous place to stash a new goody

I agree.  It has been my experience that a well-structured directory
structure is extremely useful.  I've standardized the directory structure
across multiple platforms at our site and it has been a tremendous help.
After a while of working with a good structure, locations begin to be
intuitive.  

Sebastian> - moving towards ISO9960 directory structure.

Yikes!  I better tell the rest of our CTAN people...

Sebastian> - Joachim Schrod's version of the Lee M's `mirror' script.

Hmmm, can anyone on this list tell me more about this?  Perhaps a pointer?
I've recently installed Lee's Mirror 2.1 and am gaining more experience by
the night.  What are the differences?

Sebastian> - a script to allow each of us to install things on all 3 CTANs

Hmmm, I hadn't thought much about this.  I kind of hoped to let Mirror take
care of all that.  I can see some problems, of course.  I imagine CTAN has
much more experience with the matter and I'm inclined to trust their
judgement if they feel this kind of thing is necessary.

Sebastian> - enough tolerance to keep us sane

Again, I agree.  The most important goal here, in my mind, is that we create
a service that will be as useful to the Perl community as possible.

How many archivists on this list would consider adopting a standardized
structure for their archive, were such a thing to come to pass?

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Fri Dec 10 07:18:50 1993
Received: by feenix.metronet.com id AA02838
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 01:18:55 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA02834
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 10 Dec 1993 01:18:53 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id XAA19461; Thu, 9 Dec 1993 23:18:50 -0800
Date: Thu, 9 Dec 1993 23:18:50 -0800
Message-Id: <199312100718.XAA19461@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@metronet.com
Subject: Re: plots, plans, dreams, and schemes
References: <199312100635.AA01335@feenix.metronet.com>

Bill> On Dec 6,  4:17am, Bill Middleton wrote:

Bill> } - Indexing
Bill> }     a. general index record format - naming, author?, version,
Bill> }        master archive 
Bill> }     b. descriptions - current one-liners or (possible) abstract format 
Bill> }     c. more?

I'd like to propose using the experimental Bunyip extension to Archie.  It
allows for fairly good index files to be placed anywhere on the server to be
automatically indexed by archie.  Some pointers to information about the
project may be found at http://www.hmc.edu/www/internet/ftp/index.html.

Bill> } - Classification
Bill> }    a. scripts 
Bill> }    b. libraries
Bill> }    c. info/announcement files
Bill> }    d. instructional postings (hi Tom!)
Bill> }    e. discussion and explanations 
Bill> }    f. others

I think you've done a fairly good job with your server structure.  Could you
consolidate it onto a page for examination?  Hmmm, I just poked around
a little more on your server and offer my congratulations; extremely well
done.  I'm a little confused now because it seems that most of the work of
consolidating the servers has been done by you -- at this point, I think
your server is an excellent model for the way the CPAN (I hate that name)
should look.  What do you view as the major problems?  What needs to be
fixed or improved?

Bill> } - gopher, WWW, and mailserver access 
Bill> }     a. heirarchical structure for an archive (gopher)
Bill> }     b. http WWW documents for what purposes?
Bill> }     c. mailserver request format (common?)
Bill> }     d. waisindexing, other search engines?
Bill> }     e. other retrieval methods?

This is my speciality -- integrating information systems via gateways.  I've
pretty much given up on gopher, since it is, for all intents and purposes, a
dead-end information service, but I still acknowledge its widespread use on
the Internet and will support it.  I am focusing primarily on WWW at this
point, and on coordinating the information services via gateways.

Could you answer a question about your WAIS indexing: how much space does it
take?  As a percentage of the text indexed?

I'm going to try an experiment tonight.  I am going to duplicate your server
structure and information and trying configuring mirror to place other
site's information within yours.  Hmmm, I immediately forsee some problems;
it is going to be difficult to integrate any sites which use a different
subdirectory structure.  It would be easy to mirror a bunch of directories
into a single directory, but single->multiple is impossible and
multiple->multiple is very difficult.

Perhaps mirror could be exteneded to view the directory structure from
above, so to speak.  It would check a remote directory, and not mirror it if
a local copy existed anywhere in the local structure, not just the directory
that it was mirroring into.  It would place all changed files into a
"Changed" directory, which the maintainer would have to sort manually.  Of
course, if two site had the same directory structure, it would not be a
problem.

Perhaps you can use this kind of system to mirror other sites.  How are you
currently mirroring?  If we can get this kind of system setup, your
maintainers could keep an eye on all the other archives, quickly turning
your site into a canonical site, and yet everything will be sorted into your
local structure.  Comments?

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Fri Dec 10 07:24:49 1993
Received: by feenix.metronet.com id AA03594
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 01:24:57 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA03578
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 10 Dec 1993 01:24:53 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id XAA19631; Thu, 9 Dec 1993 23:24:49 -0800
Date: Thu, 9 Dec 1993 23:24:49 -0800
Message-Id: <199312100724.XAA19631@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@metronet.com
Subject: Re: listing
References: <199312100650.AA01756@feenix.metronet.com>

[ Bill, it would nice if all such discussions about these kind of archive
  issue would take place on the list, instead of directed to me personally.
  There are certainly other people besides myself who could make valuable
  contributions to the discussion, I'm sure. ]

Bill> I just put out a better listing, that i've been working on a bit.
Bill> But, as you can see this archive is already pretty well set... perhaps
Bill> symlinks would cure any changes that i _had_ to make to conform tho.

Bill> The problem being that gopher bookmarks by the thousands have been
Bill> made to the current paths here, along with links.  It wouldnt be nice
Bill> to all the folks that have been using this archive via gopher for over
Bill> a year.

This is a problem with gopher in general; it is of an extremely static
nature.  A WWW server, on the other hand, could perform automatic
redirection for links that have been outdated, allowing for a dynamic
structure to evolve if it needed to.  This takes place automatically by the
client; the user never has to worry about such problems.

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Fri Dec 10 07:38:50 1993
Received: by feenix.metronet.com id AA04739
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 01:38:52 -0600
Return-Path: <perl-packrats>
Received: by feenix.metronet.com id AA04735
  (5.67a/feenix for perl-packrats); Fri, 10 Dec 1993 01:38:50 -0600
Message-Id: <199312100738.AA04735@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Fri, 10 Dec 1993 01:38:50 -0600
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: perl-packrats
Subject: greetings to 2 new folks on our list


  We now have Jared Rhine <Jared_Rhine@hmc.edu>, and
Jonathan Magid <jem@sunsite.unc.edu>.  Welcome!

  Jared has some very good commentary this evening.  I do in fact
have a (semi) structured archive already.  My fear has been that
this harmonic convergence we're having here would result in a lot
of folk's gopher bookmarks and links being useless, due to our
necessary conformity forcing me to change names.

  As i mentioned to Jared tho, i think symlinks can overcome most
of my troubles.   But, can we please use some (most?) of my current 
directory format?  huh?  pretty please?

  More on this tomorrow, i welcome y'all to retrieve my listing 
in /pub/perl, to kinda get an idea of how it works.  It's very crude,
still, and i hope Jared, Mark, and the rest of you can  offer more
commentary and guidance.  Mark, i've setup your account here, per
your specs.  

  I reckon you .edu folks are gonna be pretty busy for the next few
weeks, perhaps we should allow the discussion to broaden before
devising methods/plans.

Bill




  

From perl-packrats Fri Dec 10 08:21:24 1993
Received: by feenix.metronet.com id AA05980
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 02:22:48 -0600
Return-Path: <perl-packrats>
Received: by feenix.metronet.com id AA05895
  (5.67a/feenix for perl-packrats); Fri, 10 Dec 1993 02:21:25 -0600
Message-Id: <199312100821.AA05895@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Fri, 10 Dec 1993 02:21:24 -0600
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: lwall@scalpel.netlabs.com, merlyn@ora.com
Subject: Care to joiin the archivists list?
Cc: perl-packrats


Larry and Randal, 

 As you've probably noticed in the group, we've managed 
to get a list going for perl archivists.  There's about a dozen
of us now, and the discussion is interesting. 

  I'd be honored to put your name(s) on it, and send you the stuff
that's already been sent to it, if you'd like.  Or, if you'd
just like to comment, send mail to perl-packrats@metronet.com.

Thanks, 

Bill
will archive perl 4 food





From perl-packrats Sat Dec 11 01:05:11 1993
Received: by feenix.metronet.com id AA01096
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 19:05:21 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA01089
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 10 Dec 1993 19:05:17 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id RAA02250; Fri, 10 Dec 1993 17:05:11 -0800
Date: Fri, 10 Dec 1993 17:05:11 -0800
Message-Id: <199312110105.RAA02250@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@metronet.com
Subject: Mirroring technology

I recently said:

Jared> Perhaps mirror could be exteneded to view the directory structure
Jared> from above so to speak.  It would check a remote directory, and not
Jared> mirror it if a local copy existed anywhere in the local structure,
Jared> not just the directory that it was mirroring into.  It would place
Jared> all changed files into a "Changed" directory, which the maintainer
Jared> would have to sort manually.  Of course, if two site had the same
Jared> directory structure, it would not be a problem.

An easy way to handle this, I think, would be to periodically run a cron job
which would build a directory full of links to all the files on the system.
This would be the destination directory for mirror into on the local site.
This would allow implementation of the "overview" view discussed above.  You
would have mirror compare the remote directory with this "symlink-hell"
directory, and it would be able to compare timestamps with the real files
since that's where they would resolve to.  This would require a flat
namespace, but that's probably a good idea in general for the perl archives.

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Sat Dec 11 01:35:07 1993
Received: by feenix.metronet.com id AA04085
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 19:33:06 -0600
Return-Path: <perl-packrats>
Received: from blackplague.gmu.edu by feenix.metronet.com with SMTP id AA04081
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 10 Dec 1993 19:33:01 -0600
Received: by blackplague.gmu.edu (NX5.67d/NX3.0M)
	id AA01009; Fri, 10 Dec 93 20:35:08 -0500
From: Mark <mark@blackplague.gmu.edu>
Message-Id: <9312110135.AA01009@blackplague.gmu.edu>
Subject: Re: CPAN archive
To: Jared_Rhine@hmc.edu
Date: Fri, 10 Dec 1993 20:35:07 -0500 (EST)
Cc: perl-packrats@metronet.com, spp@dragonfly.cis.ufl.edu
In-Reply-To: <199312100530.VAA14097@osiris.ac.hmc.edu> from "Jared_Rhine@hmc.edu" at Dec 9, 93 09:30:41 pm
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 574       

>community and I think the Perl equivalent would be similarly useful to the
>Perl community.

I like the idea, some standardization would be useful I think.

>   scripts : For source code written in perl (note, the structure below is

It's sometimes hard to sub divide this.. I wimped out and just put
everything into the one directory and had a comment line for each
script in the INDEX which people could grep for keywords. I found
it easier for maintenance and managing filenames.

>I would also propose that 'gzip --best' be adopted as the quasi-standard

Agreed.

Mark

From perl-packrats Sat Dec 11 03:57:01 1993
Received: by feenix.metronet.com id AA10984
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 21:57:09 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA10980
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 10 Dec 1993 21:57:06 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id TAA09470; Fri, 10 Dec 1993 19:57:01 -0800
Date: Fri, 10 Dec 1993 19:57:01 -0800
Message-Id: <199312110357.TAA09470@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@metronet.com
Subject: Division of scripts
References: <199312100530.VAA14097@osiris.ac.hmc.edu>
   <9312110135.AA01009@blackplague.gmu.edu>

Jared> scripts

Mark> It's sometimes hard to sub divide this.. I wimped out and just put
Mark> everything into the one directory and had a comment line for each
Mark> script in the INDEX which people could grep for keywords. I found it
Mark> easier for maintenance and managing filenames.

Just because something isn't easy doesn't mean it is not worth doing.  I
think the advantages are worth it.  It allows for much easier browsing, when
you're not going in looking for a specific package.  It also allows cool
things like Bill's gopher menu files.

I believe Bill has implemented a good scripts division structure for
metronet.  I propose that we adopt Bill's structure (here pulled off of
gopher):

    1. ADMINISTRATION scripts /
    2. ASSORTED scripts (many authors and purposes)/
    3. DATABASE and archived file access with perl/
    4. DATE and TIME conversion/manipulation/
    5. FILE handling, devices, and links with perl/
    6. FIND type scripts and find2perl info/
    7. FTP scripts and library/
    8. GOPHER servers, clients, and tools/
    9. MAIL implementations and various scripts for mail/
   10. MATH/STAT routines in perl/
   11. MENU and curses applications /
   12. NETWORK programming with perl /
   13. NEWS scripts and utilities/
   14. TEXT handling and text processing/
   15. perl MAN replacement /
   16. some nutshell scripts (from the CAMEL book)/
   17. starting and handling PROCESSES and pipes with perl/

Any objections?

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Sat Dec 11 04:07:57 1993
Received: by feenix.metronet.com id AA11588
  (5.67a/feenix for wjm); Fri, 10 Dec 1993 22:08:05 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA11583
  (5.67a/feenix for <perl-packrats@feenix.metronet.com>); Fri, 10 Dec 1993 22:08:02 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id UAA09776; Fri, 10 Dec 1993 20:07:57 -0800
Date: Fri, 10 Dec 1993 20:07:57 -0800
Message-Id: <199312110407.UAA09776@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
Cc: perl-packrats@feenix.metronet.com
Subject: breaking bookmarks, metronet listing, discussion
References: <199312100738.AA04735@feenix.metronet.com>

Bill> My fear has been that this harmonic convergence we're having here
Bill> would result in a lot of folk's gopher bookmarks and links being
Bill> useless, due to our necessary conformity forcing me to change names.

Certainly a valid concern.  It would not be of great use to the community to
break everything while we dick around trying to get it right.  How many
people have a link deep into your structure?  I imagine most people would
link to the perl top page, the scripts top, or the faq.  Do you have reason
to believe changes other than those would hose large numbers of people?

Bill> More on this tomorrow, i welcome y'all to retrieve my listing in
Bill> /pub/perl, to kinda get an idea of how it works.  It's very crude,
Bill> still, and i hope Jared, Mark, and the rest of you can offer more
Bill> commentary and guidance.

What is the intended purpose of that listing?  While it is indeed an
overview, I would have trouble using it for much.  The file names aren't
always complete, and I would probably wouldn't use it to look for a relevant
package since the current structure makes it more likely I'll be able to
find it by following the structure via gopher.

Bill> [P]erhaps we should allow the discussion to broaden before devising
Bill> methods/plans.

Maybe.  It has been my experience that specific proposals lead quickly to
broad discussions if there is something fundamentally wrong with the
proposal.  I think we know what our general goal is ("Improve archiving
services for the Perl community"), now the question is, "How do we do that?"

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Sat Dec 11 05:53:12 1993
Received: by feenix.metronet.com id AA20311
  (5.67a/feenix); Fri, 10 Dec 1993 23:53:13 -0600
Return-Path: <perl-packrats>
Received: by feenix.metronet.com id AA20306
  (5.67a/feenix for perl-packrats); Fri, 10 Dec 1993 23:53:12 -0600
Message-Id: <199312110553.AA20306@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Fri, 10 Dec 1993 23:53:12 -0600
In-Reply-To: Jared_Rhine@hmc.edu
       "Re: plots, plans, dreams, and schemes" (Dec  9, 11:18pm)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: perl-packrats
Subject: Re: plots, plans, dreams, and schemes

On Dec 9, 11:18pm, Jared_Rhine@hmc.edu wrote:
} Subject: Re: plots, plans, dreams, and schemes
} I'd like to propose using the experimental Bunyip extension to Archie.  

Sounds good, what's a bunyip?  Can you please show us an example?
And how hard do you think it will it be to go
thru everything we eventually make "canonical" and create one
for it?  I'm personally interested in http, and WWW.
I don't have a clue how to run a server yet tho.  
We just barely got lynx going, and one of our users built it.

Perhaps some discussion on how to implement quality http docs,
along with that regarding general archive format?  I think
that both Stephen and Henk have implemented them.  To my
notions, there could be one type leading thru the archive by 
subject/topic and any number leading thru the archive in a
"tutorial" way.


} I think you've done a fairly good job with your server structure.  Could you
} consolidate it onto a page for examination?  

As you've noticed, this archive has a plethora of stuff that is organized
pretty much by what made sense to me at the time.  Mark's going thru it
now, and will have some comments later, i think.  Possibly even flames.  :)

} Could you answer a question about your WAIS indexing: how much space does it
} take?  As a percentage of the text indexed?

It's about 1/2 of whatever you index.  I've found it well worth
the space, and it's used often.  Can you create links to indexes with
WWW?

} site's information within yours.  Hmmm, I immediately forsee some problems;
} it is going to be difficult to integrate any sites which use a different
} subdirectory structure.  It would be easy to mirror a bunch of directories
} into a single directory, but single->multiple is impossible and
} multiple->multiple is very difficult.

  There are a number of problems with mirror schemes.  One i recall has
to do with the various implementations of ls used by ftpd.  I think
a daily listing of changes/updates/submissions fed to a client on the 
mirror sites might be less load on the network.  One could schedule
updates in a round-robin fashion, to maximize on throughput on different
sites, and connection distances. Directory archive structure might be 
less important with such a scheme too.  

  Another issue here is the specialized nature of some sites.  Tim
Bunce (not on our list yet) has the defacto DB archive, and others
have specialty scripts which i currently maintain only "pointers" to
in my archive.


} Perhaps mirror could be exteneded to view the directory structure from
} above, so to speak.  It would check a remote directory, and not mirror it if
} a local copy existed anywhere in the local structure, not just the directory
} that it was mirroring into.  It would place all changed files into a
} "Changed" directory, which the maintainer would have to sort manually.  Of
} course, if two site had the same directory structure, it would not be a
} problem.

Yes, i like this, and analogize it above.

} 
} Perhaps you can use this kind of system to mirror other sites.  How are you
} currently mirroring?  If we can get this kind of system setup, your
} maintainers could keep an eye on all the other archives, quickly turning
} your site into a canonical site, and yet everything will be sorted into your
} local structure.  Comments?
} 

Er, maintainers?  hehe.  you got the wrong site.  

Bill



From perl-packrats Sat Dec 11 08:19:15 1993
Received: by feenix.metronet.com id AA28295
  (5.67a/feenix); Sat, 11 Dec 1993 02:19:17 -0600
Return-Path: <perl-packrats>
Received: by feenix.metronet.com id AA28291
  (5.67a/feenix for perl-packrats); Sat, 11 Dec 1993 02:19:16 -0600
Message-Id: <199312110819.AA28291@feenix.metronet.com>
From: wjm (Bill Middleton)
Date: Sat, 11 Dec 1993 02:19:15 -0600
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: perl-packrats
Subject: one new subscriber 2day!


Greetings Merlyn!


So, now we are:

<henkp@cs.ruu.nl>                    Henk Penning
<jem@sunsite.unc.edu>                Jonathan Magid
<jhi@snakemail.hut.fi>               Jarkko Hietaniemi
<Jared_Rhine@hmc.edu>                Jared Rhine
<lmjm@doc.ic.ac.uk>                  Lee McLoughlin
<mark@blackplague.gmu.edu>           Mark Matthewson
<merlyn@ora.com>                     Randal Schwartz
<spp@cis.ufl.edu>                    Stephen Potter
<tchrist@wraeththu.cs.colorado.edu>  Tom Christiansen
<wjm@metronet.com>                   Bill Middleton


It would be nice to hear what the rest of you cats are thinkin...

Bill

From perl-packrats Sat Dec 11 12:32:53 1993
Received: by feenix.metronet.com id AA19115
  (5.67a/feenix); Sat, 11 Dec 1993 06:34:03 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA19111
  (5.67a/feenix for <perl-packrats@feenix.metronet.com>); Sat, 11 Dec 1993 06:33:54 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id EAA24475; Sat, 11 Dec 1993 04:32:53 -0800
Date: Sat, 11 Dec 1993 04:32:53 -0800
Message-Id: <199312111232.EAA24475@osiris.ac.hmc.edu>
From: Jared_Rhine@hmc.edu
To: perl-packrats@feenix.metronet.com, templates-info@bunyip.com,
        ftp-managers-l@hmc.edu, Mudd Infosys discussion <infosys-l@hmc.edu>
Subject: Indexing systems for the CPAN
References: <199312110553.AA20306@feenix.metronet.com>

The members of the Perl community are currently considering methods to
consolidate and coordinate the various Perl archive sites.  Among the topics
being discussed are the question of indexing and documenting the resources,
documents, and programs available via the archives.  Please feel free to
comment, if you think you can add anything relevant to the discussion.

Bill> - Indexing
Bill>     a. general index record format - naming, author?, version,
Bill>        master archive 
Bill>     b. descriptions - current one-liners or (possible) abstract format 
Bill>     c. more?

Jared> I'd like to propose using the experimental Bunyip extension to
Jared> Archie.

Bill> Sounds good, what's a bunyip?

Bunyip is Bunyip Informations Systems, the creators of the archie directory
system.

Bill> Can you please show us an example?

I've included the text of the original announcement (also available via my
web server as http://www.hmc.edu/www/internet/ftp/bunyip.announce.txt, which
includes an example.  (Note that there are couple of files in the .../ftp
url.)

Bill> And how hard do you think it will it be to go thru everything we
Bill> eventually make "canonical" and create one for it?

Compared to some other schemes, the Bunyip format is fairly long-winded, but
that gives it a great deal of flexibility and it holds all relevant
information.  It will require an initial investment of time, but that it
true regardless of the system we use.  Any decent editor should make the
editing of the files (plaintext) fairly trivial, and whatever fields you
don't feel like filling in can be left blank.  (Note that the use of "forms"
mode for Emacs would make maintenance _trivial") I imagine you could pretty
easily write a script to take whatever information you currently had in
.names files or scattered around and parse it into the Bunyip format to get
minimal functionality; more fields could be filled in later at your
convienence.

Bill> I'm personally interested in http, and WWW.  I don't have a clue how
Bill> to run a server yet tho.

Glad to hear it.  I used to be a gopher devotee, but it has many problems.
HTTP is much more mature and much more flexible.  If you would like some
help in setting up a server, let me know; I'd be more than happy to help
someone avoid the confusion I initially went through.  I highly recommend
the package called Plexus; it is written in Perl and is very easy to
configure.  (BTW, an _excellent_ piece of code by Tony Sanders.  Superior
coding techniques abound).  Feel free to browse my server at www.hmc.edu,
port 80, if you need a place from which to launch off.  (There Perl page is
at http://www.hmc.edu/www/computing/packages/perl, but it just pointers to
other sites).  There's also a native HTML gatewfay to my ftp server is on
there (http://www.hmc.edu/ftp/perl), although I haven't done much to develop
the view of the perl archives from Web.  (And my server consists solely of a
mirror of Ohio, Florida and Metronet; oh, well).

Something I was thinking would be very cool to do is to use Bunyip's
extension indexing system (herein called AFA) to build gopher and web views
of the archive automatically.  It should be very easy to have a couple of
(perl) scripts parse the AFA files and build the .names files for the gopher
server based upon the fields of the AFA files, then automatically update the
.cache files when that is done.  A similar thing could be done to build a
HTML file nightly for indexing.  The HTML indexes would be able to retain
all the advantages of Web, including having a link for which to mail the
author of the package, and the long description of the package (gopher could
probably only use a shorter, one lines description).  The same AFA files
could also be WAIS indexed, if that was desired.

Additionally, there should be other ways to access the indexes.  I was
already planning on building some sort of AFA -> HTML gateway, so that you
could search, via HTML forms, the various fields of the local AFA indexes
and it would return all the matching files.  A more limited gateway could be
put in for gopher, using the go4gw package, (again, written in perl).  The
AFA files are also being indexed by the Bunyip Archie project, so are
available via that method. (Bunyip people, how do you search the AFA files
via archie?  Is it currently implemented anywhere?  If there's a remote
server somewhere, writing a HTML gateway to that would also be extremely
cool; let me know if you are interested.)

Bill> Perhaps some discussion on how to implement quality http docs, along
Bill> with that regarding general archive format?

The main documents can be in just about any form.  If they are in plaintext,
I can parse them into HTML, or just leave them in plaintext (HTTP can serve
the MIME type "text/plain" just fine).  Or, they could be written in HTML,
and parsed into plaintext for the gopher to use.  Whatever floats your boat;
there are numerous ways of implementing this.  In any case, writing "quality
docs" consists primarily in knowing how to write.  I can provide general
style guidance if such becomes necessary.

Bill> I think that both Stephen and Henk have implemented them.  To my
Bill> notions, there could be one type leading thru the archive by
Bill> subject/topic and any number leading thru the archive in a "tutorial"
Bill> way.

I've looked at Stephen's server, and it is kind of weird.  All the links are
links to url's referenced as "ftp://blah..." even for files on the local
server.  This is _not_ the way to do Web.  Normal documents should (almost)
invariably be served via the url "http://blah...", even for plain text.  I
personally think a hypertext version of the man pages and the faq would be
really cool.  It could be parsed back into plain text very easily.  Note the
man page can also be converted straight via a troff->HTML gateway.

Bill> It's about 1/2 of whatever you index.  I've found it well worth the
Bill> space, and it's used often.  Can you create links to indexes with WWW?

Yes, you can use a url like "wais://blah" and you'll perform a Wais query,
getting back a list of documents which match.  You click on the appropriate
document, and off you go.

As you may have gathered by now, Web is extremely flexible.  Things can be
parsed to and from HTML, gateways are very easy to write, and it can access
most any other kind of information service (I forgot to note that there is
also a url for "gopher://blah..." if you want to access a gopher document or
directory.  The link to Metronet on my archive points to your gopher
server).

Bill> There are a number of problems with mirror schemes.  One i recall has
Bill> to do with the various implementations of ls used by ftpd.  I think a
Bill> daily listing of changes/updates/submissions fed to a client on the
Bill> mirror sites might be less load on the network.

I'd like to hear from someone with some experience as to how much load on
both ends mirroring creates.  Is it significant in any way?  Since the
majority of the server is static, I'd imagine you're rarely getting more
than a couple of files in any one sessions.  The new files have to get
across one way or another; the vast majority of the traffic will be the
actual files; any overhead created by mirroring should be dwarfed by that.

I really can't imagine updates being made via some customized bizarre
automated clients being significantly less load than mirroring.  All the
mirroring software has to do to get a list of new remote files is to do an
'ls -lR' in the top directory, and parse that for dates.  Even this load can
be eliminated, since the mirroring software can be told to use the remote
'ls -lR' file, if available.

Bill> Another issue here is the specialized nature of some sites.  Tim Bunce
Bill> (not on our list yet) has the defacto DB archive, and others have
Bill> specialty scripts which i currently maintain only "pointers" to in my
Bill> archive.

Pointers to other resources are very cool with Web, and even handy with
gopher.  It is not of much advantage to ftp users, though, and something I'm
trying to eliminate is having people run all over the world to get at what
they are looking for.  It also makes WAIS indexes less useful.  I think
having a canonical site is still very useful, and I would like to mirror as
many sites as possible, eliminating duplicates automatically via mirror, if
possible.  Such as scheme doesn't eliminate the usefulness of distribution
of maintenance.  Tim can still maintain the canonical DB tree, even on his
own site, but I don't see any reason why it should be mirrored and indexed
on those site who wish to implement a comprehensive archive.  At the very
least, I think a couple of sites should have all the indexing information
they can get their hands on, as a service to the community.

The size of the perl archives are not yet really big enough to make the
creation of a comprehensive archive impracticle.  I currently am mirroring
Ohio, Florida and Metronet, as mentioned above, so I effectively have three
copies of all the scripts (plus Metronet mirrors convex, coombs, demon, and
umanitoba, so I've copies of those too), and I'm currently a _total_ of
about 33 Mb.  This is pretty trivial, all things considered.

Bill> If we can get this kind of system setup, your maintainers could keep
Bill> an eye on all the other archives, quickly turning your site into a
Bill> canonical site, and yet everything will be sorted into your local
Bill> structure.

Bill> Er, maintainers?  hehe.  you got the wrong site.

I was pretty sure there wasn't anyone else, but I didn't want to insult any
hard workers doing their jobs in the background...

The Bunyip announcement follows:

-- begin Bunyip announcement

From: bajan@bunyip.com (Alan Emtage)
To: ftpadmins@bunyip.com
Cc: peterd@mocha.bunyip.com
Subject: Bunyip Internet Directory Project
Date: Thu, 28 Oct 1993 11:10:21 -0400

This message is being sent to the "postmaster" address at the archive
sites listed in the archie database of today (October 28, 1993) and is
intended for the FTP administrators at your site. If you are not
responsible for FTP administration, we would ask that you forward this
message to the appropriate person(s).

-------------------------

Over the next several months we at Bunyip Information Systems (creators
of the archie directory system), will be running a pilot project to
assess the viability of collecting additional data cataloging information
directly from the Internet. Ultimately, this information will be made
available to the Internet community via the distribution mechanisms now
in place for the familiar archie anonymous FTP database.

One of the biggest problems currently facing the use of the wealth of
information now available on the Internet is locating the files,
documents or services that you are looking for. While archie goes some
way to addressing this problem, "cataloging" information such as contact
information, abstracts, descriptions, keywords and access policies for
the files and services are simply not available through simple filename
searches. The problem of trying to keep this cataloging information up to
date manually in such a dynamic environment is well illustrated by the
current archie Software Description Database (or "whatis" database as it
is more commonly known).

We are asking you, the anonymous FTP archive site administrators and
operators of Internet services to cooperate with us by preparing your
information for automated gathering and serving. In doing so, we believe
you will be supporting a data distribution model that is well suited to
the Internet.  Control over your content will remain with you, and the
automatic preparation of additional information services will be made
possible. The data served to the community will be updated when you
change your own files. 

You need not tell us that you have these files: if you have asked to be
part of the archie service in the past, this information will be
automatically retrieved from your archive.

Our project will use an extended version of the archie system to collect
data "templates" on anonymous FTP archives. These templates will be
automatically retrieved, collated and indexed into a form easily
searchable by users on the network. Over the period of the pilot a number
of search and retrieval techniques will be assessed to determine their
usability.

The hope is that the general user will be able to query this database and
obtain descriptive information about freely available or Public Domain
documents, images, sounds and services on the network...worldwide. In
addition to the descriptions, the templates point the user to the source
of the original data or service which may then be accessed.

By filling out and making these templates available for retrieval, you
can advertise the files and services available from your department or
institution. The templates can be used to describe

      Services
   
        The archive can offer an overall description of each the various
        Internet services offered by your organization's systems, along
        with corresponding contact information.

        This description would then indicate whether the the parent
        organization offers such services as:

        o on-line library catalogues

        o Interactive online information services such as WAIS, gopher,
          Prospero, World Wide Web or archie

        o specialized information servers such as those providing
          weather, geographic information, newswire feeds etc.

        o Other information services

      People

        By putting non-sensitive personal contact information into the
        templates we hope to provide the start of the first truly global
        White Pages service: the ability to locate people anywhere on the
        network.

      Documents, Datasets, Mailing List Archives, USETNET Archives,
      Software Packages, Images and other Objects

        You might wish to make available a brief description of available
        software, documents, images, sounds, video, datasets, USENET
        archives and mailing list information through the anonymous FTP
        archive.

A typical "services" template might look like this (no offense to the
Census Bureau intended :-)

Template-Type:          SERVICES
Name:                   Census Bureau information server
Host-Name:              census.ispy.gov
Host-Port:              1234
Protocol:               telnet
Admin-Name:             Jay Bond
Admin-Postal:           PO Box. 42, A Street Washington DC, USA 20001
Admin-Work-Phone:       +1-202-222-3333
Admin-Work-Fax:         +1 202 444 5555
Admin-Email:            jb007@census.ispy.gov
Description:            This server provides information from the
                        latest USA Census Bureau statistics (1990)
                        Type "help" for more information.
Authentication:         Once connected type your email address at
                        the "login:" prompt. No password is
                        required.
Registration:           No formal registration is required
Charging-Policy:        There is no charge for the use of this service
Access-Times:           9:00 EST / 17:00 EST
Access-Policy:          This service may not be used by sites in
                        the Republic of the VTTS
Keywords:               census, population, 1990, statistics
Last-Modified-Name:     Miss Moneypenny
Last-Modified-Email:    m.moneypenny@census.ispy.gov
Last-Modified-Date:     Wed, 1 Jan 1970 12:00:00 GMT


Note that the machine on which the actual service or documents reside
need not be the same as that of the anonymous FTP archive: you may place
these templates on any anonymous FTP archive. The template contains the
date on where to locate the information itself.

The templates to be used will be those constructed by the Internet
Anonymous FTP Archives working group of the Internet Engineering Task
Force. These documents describing the templates are available as Internet
Drafts from the standard network repositories. They are also available
from archive.cc.mcgill.ca in the pub/Network/iafa directory. The two
files are 

draft-ietf-iafa-publish-00.txt

and

draft-ietf-iafa-templates-00.txt

the templates themselves are generally extensible and you may add fields
if you do not think that the ones provided suit your particular needs.

Please do not be put off by the size of the documents: the first contains
several examples explaining and illustrating how they are filled out; the
second provides empty templates for you to use. Also, do not feel that
you are obliged to complete one of every template or even every field in
each template. We ask only that you document current practice at your
site as completely as possible. 

If you have any questions about this project or suggestions about the
templates, we would be happy to answer them. Please send mail to

        templates-info@bunyip.com

Please feel free to redistribute this message to groups you think
appropriate.


-- 
-Alan

------------------------------------------------------------------------------
Alan Emtage,                            "The Left in Canada is more gauche
Bunyip Information Systems,              than sinister"
Montreal, CANADA                         -The Economist

bajan@bunyip.com
Voice: +1 (514) 875-8611                Fax: +1 (514) 875-8134

-- end Bunyip announcement

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Sat Dec 11 12:53:16 1993
Received: by feenix.metronet.com id AA19650
  (5.67a/feenix); Sat, 11 Dec 1993 06:53:27 -0600
Return-Path: <perl-packrats>
Received: from alpha.hut.fi by feenix.metronet.com with SMTP id AA19646
  (5.67a/feenix for <perl-packrats@feenix.metronet.com>); Sat, 11 Dec 1993 06:53:24 -0600
Received: by alpha.hut.fi; id AA08306; Sat, 11 Dec 1993 14:53:16 +0200
Date: Sat, 11 Dec 1993 14:53:16 +0200
From: Jarkko Hietaniemi <jhi@alpha.hut.fi>
Message-Id: <9312111253.AA08306@alpha.hut.fi>
To: perl-packrats@feenix.metronet.com
Subject: My thoughts about CPAN
Reply-To: Jarkko.Hietaniemi@hut.fi


I am all for unifying the structure.  In my opinion, _probably_ and
_maybe_ breaking some people's gopher bookmarks is a minor pain
compared to the greater glory of having a consistent and comprehensive,
replicated and cached Archive.

I myself do know zilch about gopher and WWW except as a user,
and ISO 9660 and the ilk are also Greek to me, so I won't comment
on those.  ftp.funet.fi is one of the largest ftp sites of the world
and there _are_ information retrieval hooks in and into it, for example
prospero.  I guess can found out how all this is done...

The script division by Bill looks fine to me.  Stand by for heated
discussion where does the new script "foo" belong :-)

To fuel the discussion, here is the structure of things I have been
archiving: (note that I am not proposing this structure as the goal,
it's more of a sorry example...)

amiga				# Amiga port(s) (2)
db				# The DBperl from ftp.demon.co.uk
info				# mainly from convex.com info:
				# FAQ, Gotchas, tchrist slides about Perl5,
				# YAPH collection, Wall quotes.
mac				# Mac port(s) (1)
msdos				# MSDOS port(s) (umpteen)
				# NOTE: this is actually maintained by
				# the MSDOS archivist folk, not by me:
				# it's is a direct mirror of some MSDOS
				# archive site Perl directory
os2				# OS/2 port(s) (1)
nt				# NT port(s) (1)
misc				# at the moment, I have here
				# Perl purity test, X bitmaps (like the Camel),
				# some slides, vgrindefs.perl
				# (most of this stuff from convex.com misc)
nutshell			# O'Reilly book example code
nutshell/llama			# either the beast...
nutshell/camel
nutshell/learning_perl		# ...or the proper name is the real dir,
nutshell/programming_perl	# ...the other one is a symlink.
				# (hmmm, should there be also a reference
				#  to this called "ora" (or "oreilly"))
refguide			# by Johan Vromans from ftp.cs.ruu.nl
scripts
scripts				# mirror of coombs (well, actually,
				# mirror of src.doc.ic.ac.uk, to ease
				# the Trans-Pacific load)
scripts/old-ftp.funet.fi	# this is a gross (lazy) hack:
				# when the coombs archive started,
				# I just threw all the old scripts
				# (mainly from convex.com) I had
				# been archiving here and forgot about 'em.
				# I am so ashamed...
snmp				# snmp perl
tutorial			# tchrist's handiwork, here
vms				# VMS port(s) (5)
wafe				# a symlink to the Wafe, resides in reality
				# under Tcl/Tk tree in ftp.funet.fi
				# (done because of WafePerl)
xenix				# Xenix port(s) (1)
				# (because allegedly Xenix is a toughie;
				#  maybe should also be archiving the
				#  Solaris port(s) for the same reason?)

Oh, where do I keep The Official Source?  At the top level.  There
is also the latest Perl5 alpha release kept, as is the Perl-byacc
source code tar.Z.

One lesson to learn from this is that there are "forces" outside us Perl
opossums gathering Perl stuff: the MSDOS people, the DBPerl project,
to name few.  The MSDOS ports are legion, there are at least three
Mac ports, etc.  Should we, could we, will we try to get all these
people in concert with CPAN?  "The Proper Way to Get Your Stuff
into CPAN?"  "Three Easy Steps to Get Your Fancy Script/Port for
Architecture Zool Canonised by CPAN?"  Or we going to be strictly
the large-cheeked critters we claim to be, doing all the foraging
ourselves and stuffing the CPAN?

Another point to remember: there are multitude of ports, even for
the "UNIX" one is not enough: Xenix and Solaris need their own hackery.
Ports might be patches, whole source distributions, just the binaries,
combinations of these, the binaries might come in many memory models,
w/ or w/o networking, et cetera.

++jhi;

From perl-packrats Mon Dec 13 18:14:55 1993
Received: by feenix.metronet.com id AA12031
  (5.67a/feenix); Mon, 13 Dec 1993 12:16:40 -0600
Return-Path: <perl-packrats>
Received: from swan.doc.ic.ac.uk by feenix.metronet.com with SMTP id AA12025
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 13 Dec 1993 12:16:33 -0600
Received: from doc.ic.ac.uk by swan.doc.ic.ac.uk id <14195-0@swan.doc.ic.ac.uk>;
          Mon, 13 Dec 1993 18:14:59 +0000
From: Lee McLoughlin <lmjm@doc.ic.ac.uk>
Date: Mon, 13 Dec 1993 18:14:55 +0000
In-Reply-To: Jared_Rhine@hmc.edu "Mirroring technology" (Dec 10, 5:05pm)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: Jared_Rhine@hmc.edu, perl-packrats@metronet.com
Subject: Re: Mirroring technology
Message-Id: <"swan.doc.i.209:13.11.93.18.15.02"@doc.ic.ac.uk>

} An easy way to handle this, I think, would be to periodically run a cron job
} which would build a directory full of links to all the files on the system.
} This would be the destination directory for mirror into on the local site.
} This would allow implementation of the "overview" view discussed above.  You
} would have mirror compare the remote directory with this "symlink-hell"
} directory, and it would be able to compare timestamps with the real files
} since that's where they would resolve to.  This would require a flat
} namespace, but that's probably a good idea in general for the perl archives.

complicated solutions generate complicated problems.  For the src.doc.ic.ac.uk
archive I have a heirarchal namespace with things arranged as sensibly as I
can manage.  Then there is a top level "packages" directory that I maintain
manual symlinks in to the interesting bits in the rest of the tree.

The symlinks need occasional pruneing and maintenance but overall it seems to
work pretty well.

-- 
--
Lee McLoughlin.                          Phone: +44 71 589 5111 X 5085
Dept of Computing, Imperial College,     Fax: +44 71 581 8024
180 Queens Gate, London, SW7 2BZ, UK.    Email: L.McLoughlin@doc.ic.ac.uk

From perl-packrats Mon Dec 13 18:29:59 1993
Received: by feenix.metronet.com id AA13674
  (5.67a/feenix); Mon, 13 Dec 1993 12:31:38 -0600
Return-Path: <perl-packrats>
Received: from swan.doc.ic.ac.uk by feenix.metronet.com with SMTP id AA13656
  (5.67a/feenix for <perl-packrats@feenix.metronet.com>); Mon, 13 Dec 1993 12:31:29 -0600
Received: from doc.ic.ac.uk by swan.doc.ic.ac.uk id <14645-0@swan.doc.ic.ac.uk>;
          Mon, 13 Dec 1993 18:30:04 +0000
From: Lee McLoughlin <lmjm@doc.ic.ac.uk>
Date: Mon, 13 Dec 1993 18:29:59 +0000
In-Reply-To: Jared_Rhine@hmc.edu "Indexing systems for the CPAN" (Dec 11, 4:32am)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: Jared_Rhine@hmc.edu, perl-packrats@feenix.metronet.com,
        templates-info@bunyip.com, ftp-managers-l@hmc.edu,
        Mudd Infosys discussion <infosys-l@hmc.edu>
Subject: Re: perl archiving
Message-Id: <"swan.doc.i.656:13.11.93.18.30.06"@doc.ic.ac.uk>

On Dec 11,  4:32am, Jared_Rhine@hmc.edu wrote:
} Subject: Indexing systems for the CPAN
} 

} Bill> There are a number of problems with mirror schemes.  One i recall has
} Bill> to do with the various implementations of ls used by ftpd.  I think a
} Bill> daily listing of changes/updates/submissions fed to a client on the
} Bill> mirror sites might be less load on the network.
} 
} I'd like to hear from someone with some experience as to how much load on
} both ends mirroring creates.  Is it significant in any way?  Since the
} majority of the server is static, I'd imagine you're rarely getting more
} than a couple of files in any one sessions.  The new files have to get
} across one way or another; the vast majority of the traffic will be the
} actual files; any overhead created by mirroring should be dwarfed by that.
} 
} I really can't imagine updates being made via some customized bizarre
} automated clients being significantly less load than mirroring.  All the
} mirroring software has to do to get a list of new remote files is to do an
} 'ls -lR' in the top directory, and parse that for dates.  Even this load can
} be eliminated, since the mirroring software can be told to use the remote
} 'ls -lR' file, if available.

Mirror is fairly cheap on both local and remote site unless you are trying
to process more than 10,000 files in a single run.  For the most part it
is no worse that doing the same steps manually (ls local, ls remote, compare,
get missing bits).

} Bill> Another issue here is the specialized nature of some sites.  Tim Bunce
} Bill> (not on our list yet) has the defacto DB archive, and others have
} Bill> specialty scripts which i currently maintain only "pointers" to in my
} Bill> archive.
} 
} Pointers to other resources are very cool with Web, and even handy with
} gopher.  It is not of much advantage to ftp users, though, and something I'm
} trying to eliminate is having people run all over the world to get at what
} they are looking for.  It also makes WAIS indexes less useful.  I think
} having a canonical site is still very useful, and I would like to mirror as
} many sites as possible, eliminating duplicates automatically via mirror, if
} possible.  Such as scheme doesn't eliminate the usefulness of distribution
} of maintenance.  Tim can still maintain the canonical DB tree, even on his
} own site, but I don't see any reason why it should be mirrored and indexed
} on those site who wish to implement a comprehensive archive.  At the very
} least, I think a couple of sites should have all the indexing information
} they can get their hands on, as a service to the community.

The best way to do this is to maintin it manually.  If a subsection is
already maintained then mirror it in.  Otherwise copy in the files once
manually, arrange them into a pretty structure, send out tons of
announcements about how to get stuff (including to FAQs) and add stuff
to the archive and then keep it up to date.  I've tried pulling in
collections from various different places and getting mirror to get rid
of the duplicates.  Mirror wasn't really designed for the task so its
a fair bit of work.  The archive then ends up with a bizarre layout
that is hard for people to browse.

-- 
--
Lee McLoughlin.                          Phone: +44 71 589 5111 X 5085
Dept of Computing, Imperial College,     Fax: +44 71 581 8024
180 Queens Gate, London, SW7 2BZ, UK.    Email: L.McLoughlin@doc.ic.ac.uk

From perl-packrats Fri Dec 17 06:35:19 1993
Received: by feenix.metronet.com id AA08112
  (5.67a/feenix); Fri, 17 Dec 1993 00:36:01 -0600
Return-Path: <perl-packrats>
Received: from Osiris.AC.HMC.Edu by feenix.metronet.com with SMTP id AA08107
  (5.67a/feenix for <perl-packrats@metronet.com>); Fri, 17 Dec 1993 00:35:50 -0600
Received: from localhost (jared@localhost) by osiris.ac.hmc.edu (8.6.4/8.6.4) id WAA20708; Thu, 16 Dec 1993 22:35:19 -0800
Date: Thu, 16 Dec 1993 22:35:19 -0800
Message-Id: <199312170635.WAA20708@osiris.ac.hmc.edu>
To: Lee McLoughlin <lmjm@doc.ic.ac.uk>
Cc: perl-packrats@metronet.com
Subject: Mirroring via symlinks not complicated
References: <"swan.doc.i.209:13.11.93.18.15.02"@doc.ic.ac.uk>
From: Jared_Rhine@hmc.edu
X-Attribution: JRR
X-Quote: To live is to war with trolls.  -- Ibsen

Jared> An easy way to handle this, I think, would be to periodically run a
Jared> cron job which would build a directory full of links to all the files
Jared> on the system.  This would be the destination directory for mirror
Jared> into on the local site.  This would allow implementation of the
Jared> "overview" view discussed above.  You would have mirror compare the
Jared> remote directory with this "symlink-hell" directory, and it would be
Jared> able to compare timestamps with the real files since that's where
Jared> they would resolve to.  This would require a flat namespace, but
Jared> that's probably a good idea in general for the perl archives.

Lee> complicated solutions generate complicated problems.  For the
Lee> src.doc.ic.ac.uk archive I have a heirarchal namespace with things
Lee> arranged as sensibly as I can manage.  Then there is a top level
Lee> "packages" directory that I maintain manual symlinks in to the
Lee> interesting bits in the rest of the tree.

Lee> The symlinks need occasional pruneing and maintenance but overall it
Lee> seems to work pretty well.

I don't think you understood what I was suggesting.  Your symlink system is
exactly what I proposed above.  If you mirrored other sites with that
directory as the destination, you'd end up retreiving updated files directly
into the proper place on the heiarchy (since it would follow the symlink (or
should, I haven't tested it)) and brand new files would be put in that
directory, awaiting your manual sorting into the proper spot.  I further
suggested that we automate the symlinking process so that you could place
new packages in their corresponding location in the "heirarchal namespace";
the symlink from the "packages" directory to that package would be
automatically created, but this is not a key part of the scheme.  You're
perfectly welcome to add them manually; I'd rather let a script do it -- to
each his own.

The only thing this system requires is a flat name space for the perl
scripts in the CPAN (someone come up with a better name, please!), which is
a good idea whether we implement my proposed mirroring scheme or not.

So much for "complicated solution"...

-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
Jared Rhine                              |  Remember, only users lose drugs.
Jared_Rhine@hmc.edu                      |-- -- -- -- -- -- -- -- -- -- -- --
wibstr - Harvey Mudd College             |   To live is to war with trolls.
http://www.hmc.edu/www/people/jared.html |-- -- -- -- -- -- -- -- -- -- -- --

From perl-packrats Mon Dec 20 08:53:46 1993
Received: by feenix.metronet.com id AA12054
  (5.67a/feenix); Mon, 20 Dec 1993 02:54:06 -0600
Return-Path: <perl-packrats>
Received: from wcarchive.cdrom.com by feenix.metronet.com with SMTP id AA12050
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 02:54:02 -0600
Received: from localhost (mark@localhost) by wcarchive.cdrom.com (8.3/8.3) id AAA03546; Mon, 20 Dec 1993 00:53:47 -0800
From: Mark <mark@wcarchive.cdrom.com>
Message-Id: <199312200853.AAA03546@wcarchive.cdrom.com>
Subject: Starting from scratch, the right way
To: perl-packrats@metronet.com (Perl Packrats)
Date: Mon, 20 Dec 1993 00:53:46 -0800 (PST)
X-Mailer: ELM [version 2.4 PL23beta2]
Content-Type: text
Content-Length: 892       

Hi All,

Well, Yet Another Perl Archive lives, tho it's rather empty (read not a byte)
at the moment. With all these good ideas and stuff floating around I was planning
on putting it to you guys to come up with the directory structure for the
ideal perl archive, (most likely similar to the metronet one but it dont hurt to
ask), and what services should serve it for users with lovely [gui] net search
clients. Bear in mind I havent used a www, html client and just a bit on gopher.
Basically we have the raw disk space.. now what rude things do we do to it?

On top of all this the guy kind enough to provide all this (Bob, rab@cdrom.com)
is planning a perl cdrom so we want to plan for as close to 600 megs of stuff
as possible. Im going to be working with Bill at metronet as he has one of the
better sites Ive seen. (not that ive gone thru them all)

Fire away people, Im all ears.
Mark

From perl-packrats Mon Dec 20 15:37:58 1993
Received: by feenix.metronet.com id AA12790
  (5.67a/feenix); Mon, 20 Dec 1993 09:38:31 -0600
Return-Path: <perl-packrats>
Received: from wraeththu.cs.colorado.edu by feenix.metronet.com with SMTP id AA12783
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 09:38:24 -0600
Received: from localhost by wraeththu.cs.colorado.edu with SMTP id AA09818
  (5.65c/IDA-1.4.4 for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 08:38:00 -0700
To: Mark <mark@wcarchive.cdrom.com>
Cc: perl-packrats@metronet.com (Perl Packrats)
Subject: Re: Starting from scratch, the right way 
In-Reply-To: Mark's message <199312200853.AAA03546@wcarchive.cdrom.com> of Mon, 20 Dec 93 00:53:46 PST.
References: <199312200853.AAA03546@wcarchive.cdrom.com> 
Reply-To: tchrist@wraeththu.cs.colorado.edu
Date: Mon, 20 Dec 93 08:37:58 MST
Message-Id: <9816.756401878@wraeththu.cs.colorado.edu>
From: Tom Christiansen <tchrist@wraeththu.cs.colorado.edu>

forgive my inattention, but has anyone given much thought to indexing
this stuff in such a way as to make retrieval useful?  as a bare
minimum, you want "site index" to work from ftp.  we probably 
want some kind of more sophisticated index.  i believe people
are working on gopherizing and webbifying the thing.  i don't
know whether wais comes into the picture, or even whether it's
useful to do so.  being able to fetch based on multiple request keys 
seems reasonable.  if each package has a reasonable description, that
would be easier.  for example,

    tchrist glob sort 

WAIS's weighting might be useful here, or not.  i've been playing with
dbase retrieval concepts of late.  one of these the LQ stuff myself.  it
was some nice features.  but a simple agrep does as well.  or even some
soundex matching.

--tom

From perl-packrats Mon Dec 20 15:38:39 1993
Received: by feenix.metronet.com id AA12846
  (5.67a/feenix); Mon, 20 Dec 1993 09:39:15 -0600
Return-Path: <perl-packrats>
Received: from wraeththu.cs.colorado.edu by feenix.metronet.com with SMTP id AA12840
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 09:39:04 -0600
Received: from localhost by wraeththu.cs.colorado.edu with SMTP id AA09832
  (5.65c/IDA-1.4.4 for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 08:38:40 -0700
To: Mark <mark@wcarchive.cdrom.com>
Cc: perl-packrats@metronet.com (Perl Packrats)
Subject: Re: Starting from scratch, the right way 
In-Reply-To: Mark's message <199312200853.AAA03546@wcarchive.cdrom.com> of Mon, 20 Dec 93 00:53:46 PST.
References: <199312200853.AAA03546@wcarchive.cdrom.com> 
Reply-To: tchrist@wraeththu.cs.colorado.edu
Date: Mon, 20 Dec 93 08:38:39 MST
Message-Id: <9830.756401919@wraeththu.cs.colorado.edu>
From: Tom Christiansen <tchrist@wraeththu.cs.colorado.edu>

someone else whose name for the nonce escapes me is also doing a perl
cdrom.

--tom

From perl-packrats Mon Dec 20 16:35:04 1993
Received: by feenix.metronet.com id AA01758
  (5.67a/feenix); Mon, 20 Dec 1993 10:32:55 -0600
Return-Path: <perl-packrats>
Received: from blackplague.gmu.edu by feenix.metronet.com with SMTP id AA01752
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 10:32:52 -0600
Received: by blackplague.gmu.edu (NX5.67d/NX3.0S)
	id AA21166; Mon, 20 Dec 93 11:35:05 -0500
From: Mark <mark@blackplague.gmu.edu>
Message-Id: <9312201635.AA21166@blackplague.gmu.edu>
Subject: cdroms...
To: perl-packrats@metronet.com
Date: Mon, 20 Dec 1993 11:35:04 -0500 (EST)
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 367       

The other person (lost somewhere in my coombs email) was doing *a*
cd but not particularly for perl.. they were doing a admin's cd or
somethign and they copied a snapshot of the coombs archive and ran off
and did the dirty with that. 

Thats my recollection, mebbe you know of another? I can dig up the
guys email if you want me to. More discs the better IMHO.

Mark

From perl-packrats Mon Dec 20 18:12:42 1993
Received: by feenix.metronet.com id AA06235
  (5.67a/feenix); Mon, 20 Dec 1993 12:13:14 -0600
Return-Path: <perl-packrats>
Received: from wraeththu.cs.colorado.edu by feenix.metronet.com with SMTP id AA06231
  (5.67a/feenix for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 12:13:09 -0600
Received: from localhost by wraeththu.cs.colorado.edu with SMTP id AA10791
  (5.65c/IDA-1.4.4 for <perl-packrats@metronet.com>); Mon, 20 Dec 1993 11:12:44 -0700
To: Mark <mark@blackplague.gmu.edu>
Cc: perl-packrats@metronet.com
Subject: Re: cdroms... 
In-Reply-To: Mark's message <9312201635.AA21166@blackplague.gmu.edu> of Mon, 20 Dec 93 11:35:04 EST.
References: <9312201635.AA21166@blackplague.gmu.edu> 
Reply-To: tchrist@wraeththu.cs.colorado.edu
Date: Mon, 20 Dec 93 11:12:42 MST
Message-Id: <10789.756411162@wraeththu.cs.colorado.edu>
From: Tom Christiansen <tchrist@wraeththu.cs.colorado.edu>

> The other person (lost somewhere in my coombs email) was doing *a*
> cd but not particularly for perl.. they were doing a admin's cd or
> somethign and they copied a snapshot of the coombs archive and ran off
> and did the dirty with that. 

> Thats my recollection, mebbe you know of another? I can dig up the
> guys email if you want me to. More discs the better IMHO.

rich morin

From perl-packrats Wed Dec 22 04:16:57 1993
Received: by feenix.metronet.com id AA00580
  (5.67a/feenix); Tue, 21 Dec 1993 22:54:45 -0600
Return-Path: <perl-packrats>
Received: from cairo.anu.edu.au by feenix.metronet.com with SMTP id AA00553
  (5.67a/feenix for <perl-packrats@metronet.com>); Tue, 21 Dec 1993 22:54:36 -0600
Received: from localhost (mark@localhost) by cairo.anu.edu.au (8.6.4/8.6.4) id PAA00605 for perl-packrats@metronet.com; Wed, 22 Dec 1993 15:16:58 +1100
From: Mark <mark@cairo.anu.edu.au>
Message-Id: <199312220416.PAA00605@cairo.anu.edu.au>
Subject: maps etc
To: perl-packrats@metronet.com
Date: Wed, 22 Dec 1993 15:16:57 +1100 (EST)
X-Mailer: ELM [version 2.4 PL21]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 495       

Hi all,

(blackplague is down so Im on this for a bit)
Could everyone who runs a perl archive send me the city and state they are
situated in and if possible what sort of link they have (t1, 56k, t3)
to their provider..

On the same idea does anyone have a good ascii map of the world and a
good ascii map of the usa. I think people would relate better to a 
graphic index for perl sites than site names. Then they can get the
best one for them, as they would have the same stuff on each.

Mark

From perl-packrats Thu Dec 23 04:48:04 1993
Received: by metronet.com id AA28390
  (5.67a/IDA1.5hp); Wed, 22 Dec 1993 22:48:07 -0600
Return-Path: <perl-packrats>
Received: by metronet.com id AA28386
  (5.67a/IDA1.5hp for perl-packrats); Wed, 22 Dec 1993 22:48:05 -0600
Message-Id: <199312230448.AA28386@metronet.com>
From: wjm (Bill Middleton)
Date: Wed, 22 Dec 1993 22:48:04 -0600
In-Reply-To: "Peter Deutsch" <peterd-mailing@mocha.bunyip.com>
       "New Information Collections for archie/Gopher" (Dec 22,  1:28am)
X-Mailer: Mail User's Shell (7.2.5 10/14/92)
To: perl-packrats
Subject: Re: New Information Collections for archie/Gopher


  Hello fellow packrats.  Merry Christmas, and let's make it a good one.

  I just got this from the bunyip people.   I'm thinking of telling them to
hold off on the perl area, til we get this settled.  I'm still kinda
confused about this bunyip thing.  Do we create the indexes for each
item?  This letter makes it sound like they're gonna do it.

  Going on short vacation.  Please, talk among yourselves.  :)


Bill

PS - Mark we're in Dallas, with 10MB bandwidth.

 
On Dec 22,  1:28am, "Peter Deutsch" wrote:
} Subject: New Information Collections for archie/Gopher
} Hi,
} 
} As you might know, the archie Internet service indexes the contents of
} anonymous FTP sites that wish to have their information advertised to the
} Internet community. Each archie server is owned and operated by individual
} volunteer sites around the Internet, using server software originally
} developed at McGill University in Montreal. There are already well over 30
} archie servers installed around the world and continued support for the
} service is now provided through Bunyip Information Systems, a company
} we've formed to ensure the availability of quality operational and
} software support for Internet services. 
} 
} Over the past year we have made a number of improvements to the archie
} system. In particular we've added support for a range of additional information
} collections. With these new capabilities, it will be possible in the coming
} months for archie server operators to offer a range of additional useful
} information.
} 
} Perhaps the single most important addition we have planned for the near
} future is a fully supported index of Gopher menu items. The code for this
} new collection has been written and tested and will soon be made available
} as a free upgrade to all existing archie servers. This upgrade will be
} accompanied by a new direct Gopher frontend onto both the original archie
} anonFTP database and the new Gopher menu item collection (and the other
} collections we are planning as they become available). It is our hope that
} a number of archie operators will choose to offer this new Gopher index
} collection, bringing a new level of availability and operational support
} to such services.
} 
} We are writing to you as the operator of a Gopher site because we would
} like your permission to include an index of your gopher tree in the new
} collection. It has been our policy with the anonFTP database to not
} include a site unless we have positive acknowlegement from the site
} operator, and we intend to continue this policy with this new
} collection. Rest assured that if you don't answer this note, we will not
} include your site in our new collection.
} 
} At the same time, we intend to follow existing practice in gopher
} indexing where possible. Thus, we will try to work with site
} administrators and tool builders to use existing indexing files where
} possible, rather than place any additional load on your server by scanning
} your entire tree. If you have such an index file available at your site
} please let us know and we will endeavour to use it.
} 
} With your permission, we will access your gopher site on a periodic basis
} to create an index of the menu items accessible from your site. The index
} will be built by one of the cooperating archie sites, that will in turn
} make it available to all the other archie sites.  Note that the archie
} system permits this sharing of collected information in a predigested
} format, lowering the load we place on the network and also
} minimizing the interactions needed at your site.
} 
} We hope that this new index will be of use both to gopher site
} administrators (in advertising the existence and contents of your site)
} and other members of the Internet community (in providing a localized,
} reliable directory of gopherspace which enjoys full operational support).
} Your comments and feedback on the service as we bring it up are most
} welcome.
} 
} 
} TO SAY YES (OR ASK QUESTIONS):
} -----------------------------
} 
} If you would like to have your site indexed, or would like more
} information, please send e-mail to 
} 
}     gopher-index-ack@bunyip.com
} 
} Also, please let us know if the information we have (as described below) is
} incorrect or in need of updating.
} 
} 
} 
} TO SAY NO:
} -----------
} 
} If we do not receive any e-mail from you, we will assume you would prefer
} not to have your site indexed, and we will not access it directly or
} through links from other gopher sites.  At the same time, if you have
} particular concerns with this concept, and would like to let us know why
} you don't want your gopher site indexed, please feel free to send comments
} to
} 
}    gopher-index-concern@bunyip.com
} 
} 
} 
} Thanks in advance!
} 
} Peter Deutsch, Alan Emtage and the gang at Bunyip Information Systems.
} 
} -----------------------------------------------------------------------------
}    
}    And a Happy Non-Denominational Seasonal Equinox Festivity 
}                       to you and yours!!
} 
}                           *
}                           #
}                          ###
}                         #*###
}                        #####*#
}                       ###*#####
}                      #*########*
}                     #####*###*###
}                          ###
}                          ###
}                   
} -----------------------------------------------------------------------------
}                   
}                   
} 
} 
} ================================
} 
} Our current information about your gopher tree(s):
} 
} 
} 
} 
}      Texas Metronet
}      Host: feenix.metronet.com
}      Port: 70
}      Path: 
}      Admin: Bill Middleton <wjm@feenix.metronet.com>
} 
}-- End of excerpt from "Peter Deutsch"


