Let's start with what I'm not interested in. I'm not interested in a
debate on different version control systems or on the usefulness of
version control in general.

What I am interested in is what *existing* strategies may have been used
in the past (or are used now) for doing version control using Forths
that stored source using *true* blocks.

Note carefully that I make two qualifications here: First, I'm only
interested in *existing* strategies because I'm looking for something
that has been proven by others, not an abstract discussion about what
possibly could be done. Second, I'm asking about Forth systems that are
based on *true* blocks, not blocks implemented on top of files since the
trivial answer in that case would be to version control the files
holding blocks.

I appreciate that such a version control system might be ultra-minimal
and not offer features like branching, tagging, differencing, and
merging, etc.. It might only offer manual roll-back of a specific block
to a prior state, or possibly a grand snapshot of all blocks captured at
a particular point in time. Whateve

On Sep 19, 8:51 am, John Passaniti JapanIsShinto.com> wrote:

On Sep 19, 9:18 am, Jean-François Michaud comcast.net> wrote:

John Passaniti wrote:

Well, a quarter century ago I briefly set up a system that sent the
blocks out over a serial line to a UNIX server which stored them in a
file, which was then placed under RCS. Worked fine, but rapidly mutated
into an approach where the field systems just kept binary images in
their block devices, with source on the UNIX servers. Handier.

Jean-François Michaud wrote:

I thought I was clear. I'm looking for *existing* strategies. Those
can be from the past or currently in use.

Jean-François Michaud wrote:

Again, I'm not looking for a discussion on how to implement this. I am
interested in existing strategies that others may have used because I'm
looking for proven approaches, not an abstract discussion about how I
might choose to implement this.

The answer may simply be that nobody had ever implemented version
control in block-based systems. That's fine, and if so it tells me that
I'm not going to be considering proven approaches but something I invent
for the particular project I'm considering.

John Doty wrote:

I didn't want to discuss my application because it doesn't matter. I'm
interested in what others have done both to learn about past solutions
and from a simple computer history perspective.

Thanks for your comments.

John Passaniti wrote:

That's where your outlook differs fundamentally from mine. The
application *always* matters. It matters more than anything else. But we
also disagree on what "application" means. Applications are things like
selling soap and finding planets. The computer and its code are tools.

On Sep 19, 8:51 am, John Passaniti JapanIsShinto.com> wrote:

A block that contains the current and previous checksums for all
blocks,
a word to show changed blocks, and an archive of release versions by
date.

On Sep 19, 9:55 am, John Passaniti JapanIsShinto.com> wrote: