[Moin-user] RCS for historical page versions?

[Douglas Alan]

Antonios Christofides <anthony at itia.ntua.gr> wrote:

> Douglas Alan wrote:

> > I'm not sure what you mean by "RCS has problems".  I've used it for more
> > than 15 years without any problems whatsoever.

> I'm far from an expert, and I don't really remember.  They were
> locking-related error messages that I got in twiki.

That's a problem with twiki, then, not with RCS.  Twiki just isn't using
RCS properly.  Or it didn't give enough consideration to the case that
two people are working on a page at the same time.  But that's something
that all wikis have to worry about anyway.

> And the developer told me of a subtle difference between the way CVS
> and wikis work; I think it had to do with concurrency

Well, yes, a common way of working with RCS is to lock all of the files
that you are working on, and then unlock them when you are done working
on them.  This approach has plusses and minuses.  A huge minus is that
someone else then can't work on a file at the same time that you are
working on it.  With CVS, you never lock files -- instead you have to
perform a merge if there ends up being a conflict of two people having
worked on the same files at the same time.

The fact that CVS is layered on top of RCS, however, proves that one
doesn't have to use RCS the way that is commonly used.  RCS is a
versatile tool and can be used in many ways.

> (yes, you see I'm not an expert, and I only replied to your message
> because no-one else did.  In any case the answer to your question is
> "no, there's no rcs", and our discussion here is academic).

Ah, well, then, thanks for the info.  Alas, the info is rather
disappointing.

> > I don't know why you say that "the compressed sized must be about the
> > same as if RCS was used".  RCS doesn't do compression -- it stores old

> I mean that if you compress a document with all its history, and say
> that its history is kept with RCS and that the total size of the
> document and its history is 1 MB, the compressed result might be 300
> KB.  Now if instead of RCS you keep the document's history in one file
> per version, and the total size of these files is 50 MB, the
> compressed result shouldn't be much more than 300 KB (it should be the
> same order of magnitude), because redundancy is much larger in the
> second case.

What you say here is not true for the vast majority of compression
algorithms, as they are generally not quite as smart as you give them
credit for.  I think you'll find that that 50 MB of files might compress
down to 15 MB.

Also, I'm not so concerned about tape backup compression ratios, as I
don't compress backups -- I do disk-to-disk backups using rsync.  I'm
concerned about disk space usage.

|>oug