unity-design team mailing list archive

[Vincenzo Ciancia <ciancia@xxxxxxxxxxx>]

Martín Soto ha scritto:
> 2009/6/16 Vincenzo Ciancia <ciancia@xxxxxxxxxxx 
> <mailto:ciancia@xxxxxxxxxxx>>
>
>     Please, let's keep the "this is something that only power user
>     like/dislike" old argument out of this discussion. I see this is not
>     your intention, but as we are all power users this is an effective
>     dialectic technique to lower the value of our observations.
>
>
> I wouldn't want to lower the value of any observations you've made, as 
> long as these are observations about other people's behavior and not 
> about your own behavior, tastes or preferences. 

Ok, then let's summarize: I can be defined a "power user" but I am 
seriously concerned with usability and not so much with customizability 
or feature-loadness.  I think I gained enough experience in the past to 
know why the latter two are not as much important as the former. I 
started 12 years ago as an FVWM advocate who liked to take one week to 
write his own .fvwmrc that did a looney-tunes like introduction at 
login, and some snow close to xmas - when kde and gnome did not exist. 
Old times :)

This is not offtopic: I want to tell you how I was convinced to change 
mentality. I read the famous gnome usability studies for gnome 2.0. Then 
I stopped laughing at gnome 2.0 and its feature-less-ness. This is why I 
would suggest that usability studies get published, it didn't happen 
this time and we are still not convinced.

> I believe you that you and Scott are not the only guys who hate this 
> feature. Still, the problem with saying "there are 20 people in 
> Launchpad who hate it too" is that all of you conform a self-selected
> sample. 

The point is not if I hate or not something. The point is "how many" 
hate this. Now there are two ways to see this. Either you want to use 
launchpad features as a metric or not. In the first case you compare the 
number of duplicates, subscribers (not only comments!) with other bug 
reports. Comments come from vocal persons, but reports do not 
necessarily. So duplicates are perhaps better than comments.

But you need to know about statistical properties (what is the typical 
distribution of the number of duplicates???).

If you think that launchpad is biased as you explain, and I certainly 
AGREE WITH YOU (surprised??),  then you have to invent another way. In 
both cases, saying that 20 duplicates means 20 users is a very bad 
estimate. You'd need to know the average of users that suffer from a bug 
w.r.t. those who take the burden to try to report it. AFAIK we do not 
have ANY CLUE about that. Unless we start messaging a bit in the default 
distribution about bugs we will never have this metric, but  I guess 
that going out and talking to your friends can give _you_ a good idea on 
the impression of people about the popup. I did that and reactions are 
laughters or just "I don't care about updates anyway".

>     Now let's get to the point of which evidence we have that people do not
>     like popups in general. For update-notification, if you want evidence,
>     again, create a poll and find a way to gather the opinion of users. I
>     won't do that because I already have good experience.
>
>
> The risk of such a poll is the same: Self selection. Obviously, people 
> are much more likely to participate if the are bothered by the feature, 
> which will immediately introduce a strong bias. 

Then a good idea *may be* to put a general poll in the testing 
distribution that all users are strongly encouraged to take. There you 
can discuss the 4-5 big changes of every distribution (those that 
typically go in the release notes). Just an idea.

Even better: we can actually implement, or mockup in a clear way, the 
possible design choices, and then in the poll ask users to test each 
choice for some time, and in the end tell us what they preferred. After 
all testing is for testing and for reporting. Why must testers jailed 
into bug finding? Can't they also be involved in the decision process? I 
will propose this as a separate thread but is ayatana the good place? 
Perhaps ubuntu-devel-discuss would be better.

Although I'm a
> scientist, I'm not an expert in this kind of research, so I guess I'll 
> ask my poll-designing colleagues here at work what they would do in such 
> a situation and see if they have a better answer. 
>  

There must be one! Please report your findings.

> You are speaking about pop-ups here, but the update notifier is rather a 
> "pop-under". It remains discretely behind other windows until you select 
> it. The only way it can be intrusive, as you already pointed out, is by 
> getting in your way when you're trying to switch windows with Alt+Tab. 

Not only, also by closing the window which is over. Then you find this 
and wonder "hey, since when this thing is here?". And you are certainly 
not guaranteed that closing e.g. evolution means not being already late 
to start serious work :) In any case, finding a window already opened 
the user will probably not want to act *right now*, and then will close it.

> Achieving a less disturbing system is, of course, a valuable goal. The 
> problem here is that if your system is, for example, running an insecure 
> network stack or a file system module that may destroy all of your data, 
> you'd rather be disturbed about it. 

Yes this was a strongly advocated point in favour of the new system. I 
just still think that the red-border white triangle with an exclamation 
mark in the notification area would be as good.

>      > . I would expect most people to just confirm the updates and keep
>      > going with their lives
>
>     Are you saying that you really NEVER experienced an upgrade that creates
>     a problem? I use my computer to work. Sometimes I just can't afford the
>     risk that the thing breaks, even in minor aspects. E.g. when I am
>     preparing a presentation or I am under a deadline. For this reasons the
>     popup cannot be that frequent, it'd be annoying to people like me.
>
>
> I'm starting to wonder if we are speaking about the same piece of 
> software. In my case, the pop-under remains quietly in the background. I 
> can ignore it for as long as I need. I was never forced by my system to 
> apply updates when I didn't want to. Am I missing something here?

No come on. If I really am working I don't want other windows opened at 
all, I must be concentrated, so I close the window. For me it is 
important to work concentrated in a clean environment. I keep as less 
windows opened as possible because I must be able to find the right one 
shortly. Keeping a window that I don't need opened is annoying for me. 
Therefore I close it and then I forget to apply updates e.g. when I stop 
my work and can afford a 30 minutes ubuntu-fix-break :) And indeed 
usually things go well! Otherwise I wouldn't be using ubuntu at all.

> [...]
>
>     Clear evidence will be obtained only when studies will be published. All
>     of this **** is based on a study which has not been published. I do not
>     work in usability and do not have the resources to do a test, but if you
>     find any, and you need some cooperation, I'll be glad to help designing
>     some experiments.
>
>
> This sounds great. We seem to have enough scientists here to run a small 
> University, and still don't manage to get our act together and try some 
> actual usability studies. I'm not an expert either, but we surely can learn.

We could use the testing distribution as a testing environment only for 
those users who, prompted by a clear question (a popup window maybe :P) 
when they install karmic (or karmic+1, I do not think we are in time for 
k), decide to participate in usability tests, involving changing the 
system in various ways for a week or so and then report feedback in a 
number-crunchable way.

But experts on statistics and usability will be of great help for this 
to be actually useful. Clear to everybody I suppose.

Vincenzo