UI map so we can try to understand what 90% of our users are doing

acelists at atlas.sk acelists at atlas.sk
Thu Sep 20 13:54:01 UTC 2012


I do not understand why we'd need the 'command' approach here. We do not want to measure the command itself, as that may be accessible from more places. But we want to measure the individual call places.
Is there just some technical reason for this?


> Od: "Mike Conley" <mconley at mozilla.com>
> Komu: <tb-planning at mozilla.org>
> Dátum: 20.09.2012 15:48
> Predmet: Re: UI map so we can try to understand what 90% of our users are doing
>Yes, this is something Blake, Jim and a few others were interested in doing.
>This is, however, a minor roadblock:  if we're to adopt the same study 
>that Firefox did, we have to change the way most of our buttons and 
>menuitems work behind the scenes.
>It seems that Thunderbird does a lot of "cheating" with menuitems and 
>buttons. We have the element, and then we connect the action to the item 
>via an oncommand parameter.
>According to Jim, what we need to do is to follow the XUL Command 
>pattern (basically, adding 1 more level of abstraction), and connect 
>those items to command nodes, and the command nodes fire the actions.
>I started to do this in bug 767118, but quickly got sidetracked (IM...or 
>Address Book...or Filelink. Something came up.)
>The project isn't very sexy - it's a lot of grunt work, but it might be 
>something that a new contributor might feel safe doing.
>This is, of course, assuming that we *must* use the command pattern to 
>adopt the study. Attaching click handlers to every menuitem / button 
>might be feasible - but I'm not sure what kind of powers TestPilot has 
>for this kind of thing...and if Firefox didn't do it that way, it tells 
>me that there's a problem with that approach.
>On 20/09/2012 9:41 AM, Ludovic Hirlimann wrote:
>> Was it two years ago the UI team for Firefox did a UI map of what user
>> used and what they didn't.
>> I have no idea how they did it (but I think they did it using some
>> combination of code + Test-pilot). WOuld it be worth figuring that out
>> so we could do the same to our users and thus try to make decisions
>> backed-up by numbers ?
>> Ludo
>tb-planning mailing list
>tb-planning at mozilla.org

More information about the tb-planning mailing list