[TUHS] UUCP Maps

[TUHS] UUCP Maps

[Lyndon Nerenberg]

The recent UUCP network conversation has me wondering ... is anyone collecting/curating the UUCP maps that represented the way we communicated (outside the ARPANET) from the time of Chesson's paper until the death of comp.mail.maps?  Brian Reid's postscript maps were a work of genius; the hand-drawn ASCII maps that predated those are even more wonderful bits of Internet history, let alone art.

--lyndon

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140825/6f265902/attachment.sig>

[TUHS] UUCP Maps

[Jaap Akkerhuis]

On Aug 25, 2014, at 22:55, Lyndon Nerenberg <lyndon at orthanc.ca> wrote:

>  Brian Reid's postscript maps were a work of genius; the hand-drawn ASCII maps that predated those are even more wonderful bits of Internet history, let alone art.

Brian told me the other day that he still as an archive of all the
maps.

	jaap


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 235 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140825/c275e8e7/attachment.sig>

[TUHS] UUCP Maps

[Lyndon Nerenberg]

On Aug 25, 2014, at 2:22 PM, Jaap Akkerhuis <jaapna at xs4all.nl> wrote:

> Brian told me the other day that he still as an archive of all the
> maps.

Then we should see about getting them to Warren for the archives.  They are a part of "internet" history that should never be lost. (Along with the code for pathalias.)


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140825/09815f04/attachment.sig>

[TUHS] UUCP Maps

[Lyndon Nerenberg]

On Aug 25, 2014, at 2:56 PM, Lyndon Nerenberg <lyndon at orthanc.ca> wrote:

> Then we should see about getting them to Warren for the archives.  They are a part of "internet" history that should never be lost. (Along with the code for pathalias.)

And perhaps a bit more importantly, does he still have the raw data that drove creating the maps?  It would be very interesting to see and compare those byte counts against today's internet. (I'm also very curious to see the relative flow between the hubs, and how data fanned out from them.)

I remember when the newsgroup traffic on Usenet hit 5 MB/day.  I think that was the third "death of the net" predicted event.  Then someone stepped up and explained (in a newsgroup) how to work your way through the entire mess, using very optimized applications of rn and kill files :-)  Where is Henry Spencer these days, anyway?!

--lyndon

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140825/168d9f1d/attachment.sig>

[TUHS] UUCP Maps

[John Cowan]

Lyndon Nerenberg scripsit:

> Then we should see about getting them to Warren for the archives.
> They are a part of "internet" history that should never be lost. (Along
> with the code for pathalias.)

Pathalias is distributed as part of Smail 3 in the pd/pathalias directory.
It is version 10 with a few changes.  Smail is currently available at
<ftp://ftp.planix.com/pub/Smail/smail-BETA.tar.gz>.  It was also in
Ubuntu up through 8.10 Intrepid, and as such should be available in
Ubuntu old release archives, probably accessible with some pain.

-- 
John Cowan          http://www.ccil.org/~cowan        cowan at ccil.org
Long-short-short, long-short-short / Dactyls in dimeter,
Verse form with choriambs / (Masculine rhyme):
One sentence (two stanzas) / Hexasyllabically
Challenges poets who / Don't have the time.     --robison who's at texas dot net

[TUHS] UUCP Maps

[Dave Horsfall]

On Mon, 25 Aug 2014, Lyndon Nerenberg wrote:

> I remember when the newsgroup traffic on Usenet hit 5 MB/day.  I think 
> that was the third "death of the net" predicted event.  Then someone 
> stepped up and explained (in a newsgroup) how to work your way through 
> the entire mess, using very optimized applications of rn and kill files 
> :-)  Where is Henry Spencer these days, anyway?!

For me, the death happened when thanks to the "cancel wars" the newsgroup 
ctl.cancel had the highest volume; at that moment I shut down our NNTP 
feed and went to mailing lists.

Knowing utzoo!henry (and I did meet him when he was visiting Australia) I 
have no doubt that he's around somewhere.

-- Dave

[TUHS] UUCP Maps

[Larry McVoy]

On Mon, Aug 25, 2014 at 06:36:17PM -0400, John Cowan wrote:
> Lyndon Nerenberg scripsit:
> 
> > Then we should see about getting them to Warren for the archives.
> > They are a part of "internet" history that should never be lost. (Along
> > with the code for pathalias.)
> 
> Pathalias is distributed as part of Smail 3 in the pd/pathalias directory.

Good old pathalias.  I've got a story for you there, Clem (he's Masscomp
right?) might get a grin out of it.

I was sys admin for 20 users on a Masscomp machine that had a 40MB disk.
We were at the end of a dog leg in UUCP, ...!uwvax!geophys!geowhiz!$user,
and could easily see the appeal in user at host.whatever.

When I ran pathalias it generated a 2MB (or more) sized file.  Way too big
for our disk.

So I walked across the street to talk to my alg prof (Udi Manber, he was
A9, ran search at google, he's a smart cookie, I was not).  He listened
to the problem and quickly said "You've got time best cast and space
worst case" and described how you would change the system to do a lookup
for each host in turn (the furthest host returns the next closer host
and so on).  That gets you space best case, time worst case.  I got it,
I saw how to write that code, and I say thanks and head out the door.
"Not so fast" says Udi.  "What would be really interesting is if you
could approximate both space and time best case".  Which lead to about
a 6 month (or more) programming effort on my part (it's a graph problem
and a dynamic programming problem) and a paper for IEEE.

The fun part about that project was that Udi was all theory and I was all
practice.  I was reading the maps in and building the graph on a microvax
with not a lot of ram, might have been 4MB, might have been less.  For
some reason that escapes me I had to sort them and I used qsort() and it
took overnight.  I went to Udi and told him about it and he asked what
sort I was using and I said qsort().  "Oh, that explains it, you need to
use a radix sort".  Stupid me slinks out to figure what a radix sort was,
did so, implemented it, it got slower!  WTF, right?

So I poked around, this was my first journey into performance debugging,
vmstat was a useful tool then (and now).  I eventually discovered that
the machine was swapping like crazy and I poked some more.  After much 
poking I decided I was fragmented and it was malloc()s fault.  Wrote my
own malloc that allocated 400K at a time and did all my strdup()s into
there.  Sort time, with qsort, when from overnight to 20 minutes.  Go
practice, eh?

I like to think that while Udi taught me all about the theory (and he did,
I flunked his class at least twice before I passed), I taught him about 
practice.  VM != real memory for example.

He's also the guy who watched me jump every time xclock chimed on the hour
and asked why?  I told him it sounded exactly the same as the beep on my
radar detector (I was a kid, I drove way too fast).  After that every time
in chimed and I jumped he'd laugh and say "you're hacking too fast" :)

--lm

P.S.  Is Honeyman still around?  I haven't seen him in years, we met in
person at usenix and he promptly dragged me outside to smoke a spliff.
Fun guy, last I heard he was running a research lab in Michigan.  Oh, yeah.
Googled, he's still there.

[TUHS] UUCP Maps

[Dave Horsfall]

On Mon, 25 Aug 2014, Larry McVoy wrote:

> P.S.  Is Honeyman still around?  I haven't seen him in years, we met in 
> person at usenix and he promptly dragged me outside to smoke a spliff. 
> Fun guy, last I heard he was running a research lab in Michigan.  Oh, 
> yeah. Googled, he's still there.

Please, not "Honeyman" but "honeyman."  He never could find his shift key, 
and must've used a custom keyboard for the "!" character.

-- Dave

[TUHS] UUCP Maps

[Larry McVoy]

> but...what was the final size of the file?

It was tiny.  I don't remember but instead of A!B!C!D!E!.... for every 
destination it was a backpointer for each node.  You compute it easily,
it's O(2N) where N is the number of nodes vs mucho bigger because of all
the repeated dog leg stuff.
-- 
---
Larry McVoy            	     lm at mcvoy.com             http://www.mcvoy.com/lm 

[TUHS] pathalias (was: UUCP Maps)

[Greg 'groggy' Lehey]

On Monday, 25 August 2014 at 18:36:17 -0400, John Cowan wrote:
> Lyndon Nerenberg scripsit:
>
>> Then we should see about getting them to Warren for the archives.
>> They are a part of "internet" history that should never be lost. (Along
>> with the code for pathalias.)
>
> Pathalias is distributed as part of Smail 3 in the pd/pathalias directory.

For those of you using FreeBSD (aren't you all?), it's in
/usr/ports/mail/pathalias.

Greg
--
Sent from my desktop computer.
Finger grog at FreeBSD.org for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed.  If your Microsoft MUA reports
problems, please read http://tinyurl.com/broken-mua
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140826/1e44366c/attachment.sig>

[TUHS] UUCP Maps

[A. P. Garcia]

On Aug 25, 2014 7:52 PM, "Larry McVoy" <lm at mcvoy.com> wrote:
>
> On Mon, Aug 25, 2014 at 06:36:17PM -0400, John Cowan wrote:
> > Lyndon Nerenberg scripsit:
> >
> > > Then we should see about getting them to Warren for the archives.
> > > They are a part of "internet" history that should never be lost.
(Along
> > > with the code for pathalias.)
> >
> > Pathalias is distributed as part of Smail 3 in the pd/pathalias
directory.
>
> Good old pathalias.  I've got a story for you there, Clem (he's Masscomp
> right?) might get a grin out of it.
>
> I was sys admin for 20 users on a Masscomp machine that had a 40MB disk.
> We were at the end of a dog leg in UUCP, ...!uwvax!geophys!geowhiz!$user,
> and could easily see the appeal in user at host.whatever.
>
> When I ran pathalias it generated a 2MB (or more) sized file.  Way too big
> for our disk.
>
> So I walked across the street to talk to my alg prof (Udi Manber, he was
> A9, ran search at google, he's a smart cookie, I was not).  He listened
> to the problem and quickly said "You've got time best cast and space
> worst case" and described how you would change the system to do a lookup
> for each host in turn (the furthest host returns the next closer host
> and so on).  That gets you space best case, time worst case.  I got it,
> I saw how to write that code, and I say thanks and head out the door.
> "Not so fast" says Udi.  "What would be really interesting is if you
> could approximate both space and time best case".  Which lead to about
> a 6 month (or more) programming effort on my part (it's a graph problem
> and a dynamic programming problem) and a paper for IEEE.
>
> The fun part about that project was that Udi was all theory and I was all
> practice.  I was reading the maps in and building the graph on a microvax
> with not a lot of ram, might have been 4MB, might have been less.  For
> some reason that escapes me I had to sort them and I used qsort() and it
> took overnight.  I went to Udi and told him about it and he asked what
> sort I was using and I said qsort().  "Oh, that explains it, you need to
> use a radix sort".  Stupid me slinks out to figure what a radix sort was,
> did so, implemented it, it got slower!  WTF, right?
>
> So I poked around, this was my first journey into performance debugging,
> vmstat was a useful tool then (and now).  I eventually discovered that
> the machine was swapping like crazy and I poked some more.  After much
> poking I decided I was fragmented and it was malloc()s fault.  Wrote my
> own malloc that allocated 400K at a time and did all my strdup()s into
> there.  Sort time, with qsort, when from overnight to 20 minutes.  Go
> practice, eh?
>
> I like to think that while Udi taught me all about the theory (and he did,
> I flunked his class at least twice before I passed), I taught him about
> practice.  VM != real memory for example.
>
> He's also the guy who watched me jump every time xclock chimed on the hour
> and asked why?  I told him it sounded exactly the same as the beep on my
> radar detector (I was a kid, I drove way too fast).  After that every time
> in chimed and I jumped he'd laugh and say "you're hacking too fast" :)
>
> --lm

but...what was the final size of the file?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140825/20a57bfe/attachment.html>

[TUHS] UUCP Maps

[Christopher Vance]

I looked after the Australia map long after it became boring and unused. I
think it was already an afterthought when I picked it up.

I guess that's what happens when you like collating lists of things.


On Tue, Aug 26, 2014 at 12:17 PM, Larry McVoy <lm at mcvoy.com> wrote:

> > but...what was the final size of the file?
>
> It was tiny.  I don't remember but instead of A!B!C!D!E!.... for every
> destination it was a backpointer for each node.  You compute it easily,
> it's O(2N) where N is the number of nodes vs mucho bigger because of all
> the repeated dog leg stuff.
> --
> ---
> Larry McVoy                  lm at mcvoy.com
> http://www.mcvoy.com/lm
> _______________________________________________
> TUHS mailing list
> TUHS at minnie.tuhs.org
> https://minnie.tuhs.org/mailman/listinfo/tuhs
>



-- 
Christopher Vance
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140826/f422cb55/attachment.html>

[TUHS] UUCP Maps

[Larry McVoy]

On Tue, Aug 26, 2014 at 11:40:14AM +1000, Dave Horsfall wrote:
> On Mon, 25 Aug 2014, Larry McVoy wrote:
> 
> > P.S.  Is Honeyman still around?  I haven't seen him in years, we met in 
> > person at usenix and he promptly dragged me outside to smoke a spliff. 
> > Fun guy, last I heard he was running a research lab in Michigan.  Oh, 
> > yeah. Googled, he's still there.
> 
> Please, not "Honeyman" but "honeyman."  He never could find his shift key, 
> and must've used a custom keyboard for the "!" character.
> 
> -- Dave

I thought he was honey at somewhere but whatever.  If he wants to nerd out about
the case in his name do I get nerd out that I was lm at sun.com and lm at sgi.com?
Doug forwarded some mail, Rob is r at google.com, I was shorter but that's 
a pretty cool email address.  
-- 
---
Larry McVoy            	     lm at mcvoy.com             http://www.mcvoy.com/lm 

[TUHS] UUCP Maps

[scj]

> On Mon, Aug 25, 2014 at 06:36:17PM -0400, John Cowan wrote:
>
> Good old pathalias.  I've got a story for you there, Clem (he's Masscomp
> right?) might get a grin out of it.
>

A great story.  It reminds me of a quote I first heard from Andy Koenig:
"What's the difference between theory and practice?  Well, in theory there
isn't any, and in practice there is."

[TUHS] UUCP Maps

[Aaron J. Grier]

On Mon, Aug 25, 2014 at 03:21:08PM -0700, Lyndon Nerenberg wrote:
> And perhaps a bit more importantly, does he still have the raw data
> that drove creating the maps?

I have a late snapshot from just before comp.maps.uucp was shut down.  I
also recall grovelling ftp sites for historic data, and finding some at
ftp.uu.net, now dead.  uucp.org also had map files and lookups
available, although it also seems to be dead now.

> It would be very interesting to see and compare those byte counts
> against today's internet. (I'm also very curious to see the relative
> flow between the hubs, and how data fanned out from them.)

the maps didn't contain byte counts, AFAIK.  the data rates were self-
reported, so were likely not terribly accurate.

by the time I got involved with UUCP, the mapping project was in its
twilight, and I recall numerous errors in my local (u.usa.or)
neighborhood.  I fixed up what I could, and even made contact with and
resurrected a few links.  (one of them actually used UUCP to pass email
traffic during a network changeover.)

the UUCP maps revealed a very coarse social topology, a lot of which I
think is still active today.  I don't think it was common to pass high
levels of traffic with strangers unless you were doing it on a
commercial basis.  :)

> I remember when the newsgroup traffic on Usenet hit 5 MB/day.  I think
> that was the third "death of the net" predicted event.

I still have an "imminent death of the net" button around here
somewhere.  :)
(http://photos.vec.com/Blog/2011-08/Blast-From-the-Past/i-9KKNPj3)

when I was still in my larval phase at my alma mater in the latter half
of the 90s, one of the two available T1s was constantly filled with
usenet traffic.  (we didn't carry binary groups.)  the feed itself was
handled by a sequent symmetry (dynix) with fujitsu supereagle(?) drives
for the spool, exported to ultrix decstations (and later OSF/1 alpha)
via NFS.  as I was exiting my larval phase, nntp was replaced with the
"on-demand" dnews running on the alphas.  (the sequents were also
retired, which subsequently broke the HVAC due to lack of load...)

I tried for a while in the early 00s to get diablo
(http://www.openusenet.org/diablo/) running under NetBSD, but never
figured out the right incantation to get it to pass socket descriptors
between processes.  (ObTUHS: what is the history of this ability?)

more recently (IE within the last few years), eternal september looked
promising, and while I could get trn to authenticate to read news, I
couldn't figure out how to get inews to handle authentication for
posting.

I digress...

-- 
  Aaron J. Grier | "Not your ordinary poofy goof." | agrier at poofygoof.com

[TUHS] UUCP Maps

[arnold]

"Aaron J. Grier" <agrier at poofygoof.com> wrote:

> more recently (IE within the last few years), eternal september looked
> promising, and while I could get trn to authenticate to read news, I
> couldn't figure out how to get inews to handle authentication for
> posting.
>
> I digress...

I use trn with eternal september for reading and posting all the
time. I think you just use the inews that comes with trn.

Arnold

[TUHS] UUCP Maps

[Mary Ann Horton]

I would be happy to curate them on stargatemuseum.org.  I can always 
download them from the Google archive of comp.mail.uucp but I wonder if 
someone has the final official tarball copy?

On 08/25/2014 01:55 PM, Lyndon Nerenberg wrote:
> The recent UUCP network conversation has me wondering ... is anyone collecting/curating the UUCP maps that represented the way we communicated (outside the ARPANET) from the time of Chesson's paper until the death of comp.mail.maps?  Brian Reid's postscript maps were a work of genius; the hand-drawn ASCII maps that predated those are even more wonderful bits of Internet history, let alone art.
>
> --lyndon
>
>
>
> _______________________________________________
> TUHS mailing list
> TUHS at minnie.tuhs.org
> https://minnie.tuhs.org/mailman/listinfo/tuhs

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://minnie.tuhs.org/pipermail/tuhs/attachments/20140914/d8380c2e/attachment.html>