Release criterion proposal: upgrade methods

[Adam Williamson]

On Mon, 2012-09-24 at 21:52 +0000, "Jóhann B. Guðmundsson" wrote:
> On 09/24/2012 09:43 PM, Adam Williamson wrote:
> > Only 17-18: using the definite article 'the' rather than the indefinite
> > article 'a' implies this. It says 'the previous stable Fedora release' -
> > which strictly means "only the single preceding release" - not 'a
> > previous stable Fedora release' or 'any previous stable Fedora release'
> > or anything like that. I suppose it's a distinction which is clearer to
> > a native speaker, admittedly, it's a bit of a fine point in English.
> > That's definitely why it's written that way, though.
> 
> Arguably we should actually be covering both GA release. ( everyone I 
> know do it at EOL time )

You could certainly make the case, yeah. Given that our excuse for
people who say 'you have to upgrade every six months' is to say 'no you
don't, because we support releases for 12, you can just upgrade every
second release!', so we kinda do tell people that.

> Do you know if we keep log in our infrastructure that shows how many are 
> actually upgrading on which version they do it from?

I don't know that, no. I don't think we do. I suppose it might be
possible to infer such information from the yum records, with a careful
analysis, by looking at installations with reliable IP addresses and
seeing their upgrade patterns. That might actually be kinda interesting,
but I don't know if it's really possible. In general Fedora is pretty
conservative about logging user information. As a F/OSS project, you run
the risk of a bad case of Slashdotitis if you do anything else =)
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora
http://www.happyassassin.net