Linux versioning file system

Peter Flass

unread,

Jun 15, 2009, 5:49:42 PM6/15/09

to

A few days ago someone posted the name of a filesystem that handles file
versions. I just spent 1/2 hour with Google today, and I'll be darned
if I could find the post. Could I impose on you to repost the name? Thanks.

Joe Pfeiffer

unread,

Jun 15, 2009, 6:25:12 PM6/15/09

to

Peter Flass <Peter...@Yahoo.com> writes:

In the linux world, I posted that I've got a student working on it...

Alec Kojaev

unread,

Jun 16, 2009, 5:15:42 AM6/16/09

to

There's also NILFS <http://www.nilfs.org>, version 2 recently accepted
into Linux kernel 2.6.30. Not exactly file versions, but regular
snapshotting of the filesystem with independent access to any snapshot.

--
Alec Kojaev
St.Petersburg, Russia [30E18 59N56]

Peter Flass

unread,

Jun 16, 2009, 6:50:09 AM6/16/09

to

Too bad there still isn't something like VMS. I'm starting to do some
development on Linux, and I'd live the simplicity of SET/FILE/VERSION=
for just the sources. I don't want to add RCS/CVS/whatever to the stuff
I'm learning to use.

Scott Lurndal

unread,

Jun 16, 2009, 1:08:45 PM6/16/09

to

http://publib.boulder.ibm.com/infocenter/cchelp/v7r0m1/index.jsp?topic=/com.ibm.rational.clearcase.books.cc_ms_install_ux.doc/cc_ms_install10.htm

scott

Peter Flass

unread,

Jun 16, 2009, 7:41:50 PM6/16/09

to

Interesting, but it sounds like another flavor of CVS and, being IBM, a
more complicated flavor.

Quadibloc

unread,

Jun 17, 2009, 7:10:50 AM6/17/09

to

I know that OpenVMS handles file versions quite nicely.

John Savard

Sarr J. Blumson

unread,

Jun 17, 2009, 1:49:22 PM6/17/09

to

Peter Flass <Peter...@yahoo.com> wrote:

Well, Clearcase is something IBM bought rather than built but yes, it is a more complicated equivalent
of CVS. But yes part of the added complication is that it does manage to present it's tree as a virtual file
system that (at least on Solaris, which is the only platform I've used it) does present itself to
users as just a file system.

--
--------
Sarr Blumson sarr.b...@alum.dartmouth.org
http://www-personal.umich.edu/~sarr/

Anne & Lynn Wheeler

unread,

Jun 17, 2009, 2:38:56 PM6/17/09

to

"Sarr J. Blumson" <sa...@umich.edu> writes:
> Well, Clearcase is something IBM bought rather than built but yes, it
> is a more complicated equivalent of CVS. But yes part of the added
> complication is that it does manage to present it's tree as a virtual
> file system that (at least on Solaris, which is the only platform I've
> used it) does present itself to users as just a file system.

we had developed our own source management system in CMS. It was used
for virtual machine related projects.

clear was used for os360 lineage ... although some number of those
products did majority of their stuff (using cms tools) and then were
forced to morph them into proscribed tools (JES was major example for a
long time).

this became less of a concern to customers after OCO-wars in the early
80s, where company stopped shipping source.

an old folklore from the period was a sensitive gov. agency asked the
corporation for the "exact" source corresponding to running MVS
product. supposedly there was a $5m investigation that eventually
concluded that it wasn't feasible or practical (this was when virtual
machine customers were still getting source and building their running
systems from the shipped source).

--
40+yrs virtualization experience (since Jan68), online at home since Mar1970

Louis Krupp

unread,

Jun 17, 2009, 4:07:43 PM6/17/09

to

Anne & Lynn Wheeler wrote:
<snip>

> clear was used for os360 lineage ... although some number of those
> products did majority of their stuff (using cms tools) and then were
> forced to morph them into proscribed tools (JES was major example for a
> long time).

"Proscribed" means "prohibited." "Prescribed" is closer to the word you
want.

Louis

Rich Alderson

unread,

Jun 17, 2009, 7:01:52 PM6/17/09

to

Quadibloc <jsa...@ecn.ab.ca> writes:

As does TOPS-20 and its ancestor TENEX. VMS simply followed its ancestor
RSX-11M.

--
Rich Alderson "You get what anybody gets. You get a lifetime."
ne...@alderson.users.panix.com --Death, of the Endless

Christopher C. Stacy

unread,

Jun 21, 2009, 2:21:11 AM6/21/09

to

Rich Alderson <ne...@alderson.users.panix.com> writes:

> Quadibloc <jsa...@ecn.ab.ca> writes:
>
>> On Jun 15, 3:49=A0pm, Peter Flass <Peter_Fl...@Yahoo.com> wrote:
>
>>> A few days ago someone posted the name of a filesystem that handles file
>>> versions. I just spent 1/2 hour with Google today, and I'll be darned
>>> if I could find the post. Could I impose on you to repost the name? Thanks.
>
>> I know that OpenVMS handles file versions quite nicely.
>
> As does TOPS-20 and its ancestor TENEX. VMS simply followed its ancestor
> RSX-11M.

And TENEX got it from ITS

Rich Alderson

unread,

Jun 22, 2009, 7:08:41 PM6/22/09

to

> Rich Alderson <ne...@alderson.users.panix.com> writes:

>> Quadibloc <jsa...@ecn.ab.ca> writes:

Good point. Was it original to ITS, or did someone somewhere publish a paper
that the ITS developers for certain read? There's always the possibility of
parallel developments, of course.

Christopher C. Stacy

unread,

Jul 5, 2009, 7:34:01 PM7/5/09

to

Rich Alderson <ne...@alderson.users.panix.com> writes:

> cst...@news.dtpq.com (Christopher C. Stacy) writes:
>
>> Rich Alderson <ne...@alderson.users.panix.com> writes:
>
>>> Quadibloc <jsa...@ecn.ab.ca> writes:
>
>>>> On Jun 15, 3:49=A0pm, Peter Flass <Peter_Fl...@Yahoo.com> wrote:
>
>>>>> A few days ago someone posted the name of a filesystem that handles file
>>>>> versions. I just spent 1/2 hour with Google today, and I'll be darned
>>>>> if I could find the post. Could I impose on you to repost the name? Thanks.
>
>>>> I know that OpenVMS handles file versions quite nicely.
>
>>> As does TOPS-20 and its ancestor TENEX. VMS simply followed its ancestor
>>> RSX-11M.
>
>> And TENEX got it from ITS
>
> Good point. Was it original to ITS, or did someone somewhere publish a paper
> that the ITS developers for certain read? There's always the possibility of
> parallel developments, of course.

I think it was original on ITS; not sure which individual was responsible for it.
It was before my time, but I believe the filesystem that they were looking at was CTSS.

Christopher C. Stacy

unread,

Jul 5, 2009, 7:40:43 PM7/5/09

to

Rich Alderson <ne...@alderson.users.panix.com> writes:

> cst...@news.dtpq.com (Christopher C. Stacy) writes:
>
>> Rich Alderson <ne...@alderson.users.panix.com> writes:
>
>>> Quadibloc <jsa...@ecn.ab.ca> writes:
>
>>>> On Jun 15, 3:49=A0pm, Peter Flass <Peter_Fl...@Yahoo.com> wrote:
>
>>>>> A few days ago someone posted the name of a filesystem that handles file
>>>>> versions. I just spent 1/2 hour with Google today, and I'll be darned
>>>>> if I could find the post. Could I impose on you to repost the name? Thanks.
>
>>>> I know that OpenVMS handles file versions quite nicely.
>
>>> As does TOPS-20 and its ancestor TENEX. VMS simply followed its ancestor
>>> RSX-11M.
>
>> And TENEX got it from ITS
>
> Good point.

Oh, and after TENEX gave it to TOPS-20 (well, and VMS), the Lisp Machine
picked it up for its native file system(s). (The Symbolics file system
also picked up "generation retention count", which ITS did not have.)

On an unrelated note (file system versions were a filesystem/kernel thing),
the Lisp Machine also took the TOPS-20 command line prompting/processing ideas
to a new level. The LispM took a number of ideas from ITS, TENEX/TOPS-20,
Multics, and other places. This happened not because of publications,
but because the team was composed of people who were familiar with all
those systems (as both users and often as developers).

jmfbahciv

unread,

Jul 6, 2009, 8:00:44 AM7/6/09

to

That (developers familiar with other OS features) is how a lot of
things got implemented on other systems. It's why DEC OSes
look similar (except for the one that NIH syndrome for its first
5 years).

/BAH