Re: [Monotone-devel] monotone philosophy (was Re: Undo a commit)

[hendrik]

On Thu, Oct 09, 2008 at 05:24:46PM -0500, Timothy Brownawell wrote:
> On Thu, 2008-10-09 at 17:44 +0100, Bruce Stephens wrote:
> > Bruce Stephens <address@hidden> writes:
> > 
> > > Lapo Luchini <address@hidden> writes:
> > 
> > [...]
> > 
> > >> Uh, really? 0_o
> > >> I had thought for the past years that monotone meant just that…
> > >
> > > Yes, I think it was about monotonically non-decreasing quality, not
> > > about never removing stuff.
> > 
> > I was fairly sure that was the original motivation (and that all this
> > other stuff (better merging, etc.) was just an unfortunate
> > distraction), but I can't seem to find any confirmation of that.  So
> > maybe I'm misremembering?
> 
> 
> Feb 09 16:32:22 graydon       I initially wrote monotone in order to support
> the use-case of a "monotonically improving branch", where there's a view
> of a branch which never gets worse, constructed by selecting only the
> branch certs signed by an automatic tester.
> Feb 09 16:32:41 graydon       this way a customer or user who wanted "new, but
> never worse" software could view the branch that way
> Feb 09 16:33:22 graydon       or alternatively, a developer could configure
> their update rules to exclude stuff other people wrote which isn't
> working yet

This suggests that monotone would be ideal for a mathematics formal 
verification project, where mathematians could check in their attempts 
at formal proofs with duly deliberate gay abandon, but where only some 
of them would ever get certified -- by an automated proof-checker.

-- hendrik