Re: Version numbers and an update from the UX team

On 20/08/11 23:10, Alex Faaborg wrote:
> == I personally believe the following things are true ==
> 
> -Making the about window primarily designed to achieve the task of checking
> "am I up to date" is great

I agree.

> -Some technical users are reacting in a jaded way to large version bumps
> like 5, 6 and 7 because it took us several years for both 3 or 4, the sudden
> change in interval comes off as cheating even though that wasn't the
> intention (just need to get changes out faster to further progress on the
> Web)

I think that a vague understanding of version numbers ("first number is
a big change, second number a small change") is much more widespread
than you think. This isn't quite the same as our previous technical
definition ("major number change == breaking things; minor == less or no
breakage) but it mapped reasonably. Now, it maps very badly, and this
confuses people.

Given the arbitrariness of version numbering from a geek point of view,
I still have no idea why we didn't decide to increment the major number
once per year, and the second number for all the other releases in a year.

It's a short step from that idea to "2011.4", etc. - but I think that
still causes cognitive dissonance, because version 2011 of a product
seems weird. Even Ubuntu knocked off the first two digits to get the
numbers down into a sane range.

> -If a user reads that 7.0.1 was released and they don't see this in the
> dialog box, they will be confused if we claim 7.0 is the most up to date
> version

But removing the version number from the About dialog has the potential
to cause all sorts of other press-report-related confusion as well, some
scenarios of which people have explained in this thread.

> -Multiple decimal places in version numbers is silly, and is pretty much
> only encountered in the realm of software

"Silly" is a value judgement :-) They aren't decimal places, they are
separators - like the dots in a domain name.

> -Version numbers are generally speaking technical jargon, and user
> experience designers generally speaking like to try to remove instances of
> jargon

On the contrary, they are unique identifiers. Humans understand unique
identifiers with multiple parts. "23-06-2001". "14:33:17". "17 1st Ave,
San Francisco, CA". Understanding that this is software version "4.1.3",
and that "5.2.7" is a later version and "3.0.0" is an earlier one, is
not beyond the wit of most mortals.

> -Some technical users really want to see the version number in an easy way
> for legitimate reasons (web developers, extension developers, IT, etc.)
> these people are a minority, but like the people who need to access view
> source or developer tools, they are also important. Since they are building
> stuff you could even argue that they are very disproportionately important

And because they are people that, just at the moment, we don't want to
annoy any more than we have already. And they are drivers of Firefox
adoption among family and friends, and are the people who make sure
websites support us well.

> == One possible proposed solution (also proposed by many other people
> through many channels ==
> 
> We de-jargonify the version number, and make it time based.
> 
> Format is 2011.1, 2011.2, etc. We refer to point releases simply as "a patch
> to 2011.7", and we avoid press statements that include multiple decimal
> places to limit all around user confusion.

If we could switch to 11.1 instead, that would be much better - as you
say, big numbers feel weird. But we need to do it quick, before or when
mozilla-central goes to 11.

> We refer to the versions as "the fourth release of 2012" as a long form in
> some public statements, instead of 2012.4, to further de-jargonify it and
> make it human readable, but they are otherwise the same thing.

I'm not sure I have much confidence that the press will use the long
form over the short form.

> Overall keeping the version number around in the about dialog doesn't really
> diminish the work to make the dialog primarily about confirming that you are
> running the newest version (or that you need to apply an update), and that
> is from my perspective the larger UI win here.

Amen, brother!

> == Why so much emotion? ==
> 
> I think the reason this debate became so emotional is that some people want
> to change client side software to behave like the Web (where the user has no
> control over version), and some people simply aren't comfortable with that
> model. 

I think that's a very pertinent insight. And we need to decide what to
do about it, and how we can make them more comfortable. If knowing the
version number gives people a feeling of control, we need to take that
into account.

Gerv