ANN of new versions of Mercurial

[Paul Boddie]

On 27/12/12 04:46, Adrian Klaver wrote:
> On 12/26/2012 01:05 PM, Matt Mackall wrote:
>> And that's what the goal of announcing the release candidates instead of
>> the releases is: encourage Earl to be earlier so that his extra
>> enthusiasm actually helps the project.
>
> The fault I see in this is the assumption you can increase the pool of 
> early adopters. I yet to see that happen. There is a fairly fixed 
> proportion of software users that will take the plunge, the rest wait. 
> The early adopters tickle the low hanging bugs. It is not until the 
> final release is made and the pool of users expands do the tough bugs 
> get flushed out. Do not understand why a production release should be 
> hidden. It would seem to be something to be proud of.  In any case it 
> is not my project, so I will go along with the program.

Another problem with only announcing release candidates is that people 
may hold off on upgrading because they see various release candidates 
and nothing more and then think that the release has been blocked somehow.

I'd also agree that even with substantial encouragement, most people 
won't even look at prereleases and will only ever stumble across bugs 
when they finally upgrade to a proper release, which I suppose is the 
essence of the "Earl" character. The problem is that most instances of 
Earl will not be in a position to use release candidates, either because 
they are not really allowed to or because they depend on the software 
for their work [*]. And it is precisely the class of bugs associated 
with using the software "in anger" that won't get flushed out just 
through enthusiastic testing, although you can obviously work towards 
better test coverage.

I have to say that it's a bit weird to implicitly claim that Earl's 
enthusiasm doesn't actually help the project. We can all be disappointed 
that Earl didn't give the release candidates a test drive and eliminate 
bugs from an actual release, but having him report bugs is better than 
him and everyone else affected by a particular bug giving up and 
silently using something else.

Paul

[*] When asked to give Ubuntu releases a test to see if the bootloader 
works, to provide an extreme example, I often wonder whether Canonical 
thinks its demographic is a bunch of tinkerers who don't do anything 
more than try out the latest release, as opposed to people using the 
software to do actual work. I also wonder whether they have any testing 
infrastructure to eliminate basic deployment issues at all. There's only 
a limited potential in asking people to bear a significant cost 
(disruption and potential loss of work) in order to save a much lower 
cost to the project.