> On Wed, Oct 22, 2003 at 12:34:14AM +0200, Robert Simmons wrote:
> > > systems like Perforce that use a single repository wide changeset
> > > number, and they seem to get by. What's the big issue you have with it?
> > Every day some 10,000 people world wide get in their car and drive drunk.
> > doesnt mean its an especially good idea. Just because "others" do something
> > doesnt mean I wish to do it as well. I simply dislike it when I have low
> I think I understand your point, though I would probably not have used
> the example of drunk driving. Simply stated, driving drunk could not
> be considered typical behavior for the general driving population.
> Perforce, on the other hand, may be considered normative behavior for
> many developers.
> Yes, I agree that the repository-wide revision numbers require some
> changes to our behavior as developers.
> Yes, I agree that it is interesting and desirable to be able to ask
> the question: how stable is this file or this set of files.
> Speaking from the perspective of someone who has used CVS since the
> mid-nineties, I think this is more about changing our behavior and
> assumptions than about changing the tool.
That would be cool. =) Now you just need to convince 100 million corporate
drones that think they know it all. =) Sometimes you have to compromise and give
people a feel of the old as you introduce the new.
I like the idea of a file revision number for the reasons I illustrated. Others
will DEMAND it merely because they are used to it. If you use revision nubmers
as is in subversion and as an aside have also a file version number as well, you
will hook far more fish.
> Note that
> 1) The revision number for a file only changes when the file changes
Yes, but then one change skips potentially thousands of revision numbers. I
understand the current system and even like it. I just think this one little
extra thing would be 1) handy for the issues I mentioned and 2) handy for
breaking the ice against the corporate mindset.
> 2) A CVS-like revision number can be recovered from the log by
> counting the number of entries.
NICE. Then data migration wont be so hard. =)
> 3) Tagging is much more effective in subversion than CVS and should
> be used accordingly.
Thats good too, but Id hate to have to tag every time I changed a file.
To you too. =)
To unsubscribe, e-mail: firstname.lastname@example.org
For additional commands, e-mail: email@example.com
Received on Wed Oct 22 01:15:36 2003