27 Nov
2020
27 Nov
'20
6:59 p.m.
I'm trying to find out why compress(1) uses .Z as filename extension.
My theory is that it was inspired by pack(1), which uses the .z extension.
However, I haven't been able to find any info on why pack(1) uses that
extension. Does anyone here know?
Some searching led me to [1] which is a man page for pack from AUSAM.
It's written by Steve Zucker in 1975, so perhaps the extension is z for
Zucker?
Was Zucker's pack(1) the first, though? This message [2] talks about a
Bell version.
Does anyone here have any information about this?
Cheers,
Hans
1. https://minnie.tuhs.org/cgi-bin/utree.pl?file=AUSAM/doc/man/man1/pack.1
2. https://tech-insider.org/unix/research/1984/0319.html
27 Nov
27 Nov
9:57 p.m.
On Fri, Nov 27, 2020 at 8:07 AM Hans Wennborg <hans(a)hanshq.net> wrote:
I'm trying to find out why compress(1) uses .Z as filename extension.
My theory is that it was inspired by pack(1), which uses the .z extension.
I think that connection is plain. As for the .z extension, I'd guess (and
it's nothing better than that, just a conclusion I jumped to long ago) that
it makes compressed files sort after files that share the same root name.
IIRC, compress(1) used to be able to unpack files as well as uncompressing
them. I don't know if that's still true.
John Cowan http://vrici.lojban.org/~cowan cowan(a)ccil.org
May the hair on your toes never fall out! --Thorin Oakenshield (to Bilbo)
11:16 p.m.
On Fri, Nov 27, 2020 at 8:08 AM Hans Wennborg <hans(a)hanshq.net> wrote:
I'm trying to find out why compress(1) uses .Z as
filename extension.
My theory is that it was inspired by pack(1), which uses the .z extension.
Yes.
However, I haven't been able to find any info on why pack(1) uses that
extension. Does anyone here know?
No idea - but yes, Zucker used a .z at Rand when he wrote.
Some searching led me to [1] which is a man page for pack from AUSAM.
It's written by Steve Zucker in 1975, so perhaps the extension is z for
Zucker?
Was Zucker's pack(1) the first, though? This message [2] talks about a
Bell version.
Zucker wrote it at Rand - early/mid 1970s. IIRC, It was later included in
the original Harvard USENIX tape in the 'Rand' directory. I believe that
Rand Pipes (named pipes) are in the same directory. Although some of the
Rand stuff was being shared by folks on the ArpaNet before USENIX existed
and I think it made it to the wild before the first USENIX tape.
It was really important back in the day. Remember RK05's are only 2.5M
bytes - source archiving and packing files was pretty important given the
cost / byte of disk.
I think there may have been an early version @ BTL - PWB may have
distributed it also, but I'm fairly sure it was the Rand code that started
it. Noel might remember more than I. I'm 90% sure we had it at CMU before
we got either PWB 1.0 or UNIX/TS from Ted -- I want to say it we had it on
5th edition but maybe not.
One of the PDP-10 folks will need to chime in here. My memory is there was
something like pack(1) on the CMU PDP-10s and 20s that I saw before I saw
the UNIX tool [not sure why I think this, but it may have been SAIL program
- I remember looking at a number of simple tools when I learn SAIL years
and years ago - 74/75-ish]. IIRC they were not exactly the same format as
the 10's were 36-bit words, stored 5 chars in a word, but it was the same
idea.
28 Nov
28 Nov
1 a.m.
On Fri, Nov 27, 2020 at 10:17 AM Clem Cole <clemc(a)ccc.com> wrote:
On Fri, Nov 27, 2020 at 8:08 AM Hans Wennborg <hans(a)hanshq.net> wrote:
Regardless of the early history, this is why gzip files end in .gz and not
.z. The pack'd files were considered too prevalent at the time to crowd
into that name space, even though this happened maybe 12 years after
compress started to displace pack.
Warner
I'm trying to find out why compress(1) uses
.Z as filename extension.
My theory is that it was inspired by pack(1), which uses the .z extension.
Yes.
However, I haven't been able to find any info on why pack(1) uses that
extension. Does anyone here know?
No idea - but yes, Zucker used a .z at Rand when he wrote.
Some searching led me to [1] which is a man page for pack from AUSAM.
It's written by Steve Zucker in 1975, so perhaps the extension is z for
Zucker?
Was Zucker's pack(1) the first, though? This message [2] talks about a
Bell version.
Zucker wrote it at Rand - early/mid 1970s. IIRC, It was later included in
the original Harvard USENIX tape in the 'Rand' directory. I believe that
Rand Pipes (named pipes) are in the same directory. Although some of the
Rand stuff was being shared by folks on the ArpaNet before USENIX existed
and I think it made it to the wild before the first USENIX tape.
It was really important back in the day. Remember RK05's are only 2.5M
bytes - source archiving and packing files was pretty important given the
cost / byte of disk.
I think there may have been an early version @ BTL - PWB may have
distributed it also, but I'm fairly sure it was the Rand code that started
it. Noel might remember more than I. I'm 90% sure we had it at CMU before
we got either PWB 1.0 or UNIX/TS from Ted -- I want to say it we had it on
5th edition but maybe not.
One of the PDP-10 folks will need to chime in here. My memory is there
was something like pack(1) on the CMU PDP-10s and 20s that I saw before I
saw the UNIX tool [not sure why I think this, but it may have been SAIL
program - I remember looking at a number of simple tools when I learn SAIL
years and years ago - 74/75-ish]. IIRC they were not exactly the same
format as the 10's were 36-bit words, stored 5 chars in a word, but it was
the same idea.
1:17 a.m.
I thought the gz was because dos voilf not tell the difference between Z
and z which was strange of course since it stored it on 8 bits unlike rt11
which used RAD50 (5bits) sigh.
On Fri, Nov 27, 2020 at 2:00 PM Warner Losh <imp(a)bsdimp.com> wrote:
On Fri, Nov 27, 2020 at 10:17 AM Clem Cole <clemc(a)ccc.com> wrote:
--
Sent from a handheld expect more typos than usual
On Fri, Nov 27, 2020 at 8:08 AM Hans Wennborg <hans(a)hanshq.net> wrote:
Regardless of the early history, this is why gzip files end in .gz and not
.z. The pack'd files were considered too prevalent at the time to crowd
into that name space, even though this happened maybe 12 years after
compress started to displace pack.
Warner
I'm trying to find out why compress(1) uses
.Z as filename extension.
My theory is that it was inspired by pack(1), which uses the .z
extension.
Yes.
However, I haven't been able to find any info on why pack(1) uses that
extension. Does anyone here know?
No idea - but yes, Zucker used a .z at Rand when he wrote.
Some searching led me to [1] which is a man page for pack from AUSAM.
It's written by Steve Zucker in 1975, so perhaps the extension is z for
Zucker?
Was Zucker's pack(1) the first, though? This message [2] talks about a
Bell version.
Zucker wrote it at Rand - early/mid 1970s. IIRC, It was later included in
the original Harvard USENIX tape in the 'Rand' directory. I believe that
Rand Pipes (named pipes) are in the same directory. Although some of the
Rand stuff was being shared by folks on the ArpaNet before USENIX existed
and I think it made it to the wild before the first USENIX tape.
It was really important back in the day. Remember RK05's are only 2.5M
bytes - source archiving and packing files was pretty important given the
cost / byte of disk.
I think there may have been an early version @ BTL - PWB may have
distributed it also, but I'm fairly sure it was the Rand code that started
it. Noel might remember more than I. I'm 90% sure we had it at CMU before
we got either PWB 1.0 or UNIX/TS from Ted -- I want to say it we had it on
5th edition but maybe not.
One of the PDP-10 folks will need to chime in here. My memory is there
was something like pack(1) on the CMU PDP-10s and 20s that I saw before I
saw the UNIX tool [not sure why I think this, but it may have been SAIL
program - I remember looking at a number of simple tools when I learn SAIL
years and years ago - 74/75-ish]. IIRC they were not exactly the same
format as the 10's were 36-bit words, stored 5 chars in a word, but it was
the same idea.
1486
days inactive
1486
days old
4 comments
4 participants
participants (4)
-
Clem Cole -
Hans Wennborg -
John Cowan -
Warner Losh