Re: [Libreoffice] [PATCH] Removes data sources button from standard toolbars in Calc and Writer

Hi Samuel,

thanks for your kind reply ...

Am Sonntag, den 12.06.2011, 08:38 -0700 schrieb Samuel Cantrell:
[...]
> I was thinking about some of the things you mentioned...I believe that
> Mozilla had some kind of testing where they (with the user's
> permission) took data on which buttons were used the most often in the
> various Firefox 4 testing builds. Maybe we could do something similar
> with LibreOffice, to find out what buttons are used the most and when?
> (The UX team might find that kind of data interesting.)

In OpenOffice.org, we had such a system called "OpenOffice.org
Improvement Program" (a.k.a. User Feedback Program, a.k.a. Usage
Tracking). You may have a look at (implementation) details [1], or
results [2]. I extensively used this data for the printer dialog
redesign [3].

The system was mainly visible on Windows platforms, since most Linux
distributions disabled it. And as you already mentioned - although it
was a bit buggy (some actions weren't tracked), or difficult to use
(required detective work), and data access limited (only raw data
available to the community), the overall system was (and is) still a
great resource and helpful for making good UI decisions.

The system consisted of both a sender (the code in OpenOffice.org/
LibreOffice) and a receiver (the server side to collect, transform and
store the data). Unfortunately, the server side is available within
Oracle only - so we don't have any access to recent and more complex
data.

To be honest, I'm (or let's say: the Design Team are) still searching
for people/developers who might be interested to help us to rebuilt this
excellent information source (the major issues besides legal stuff and
server resources).

[1]
http://wiki.services.openoffice.org/wiki/User_Experience/OpenOffice.org_User_Feedback_Program

[2]
http://wiki.services.openoffice.org/wiki/Tracking_results#Results_of_the_User_Feedback_Program

[3]
http://wiki.services.openoffice.org/wiki/Printerpullpages/Current_State_of_Printing_OOo31#Source_Data

> I'm sorry if I'm coming across as trying to force changes on everyone
> else. I was just looking for stuff to do that was within my skillset,
> saw that someone had mentioned this idea [...]

Oh, my impression was different - more like "Cool that you joined to
help us! But how to make sure that this change is perceived as an
improvement for the majority of users". That's something we still try to
figure out - and the reason for the libo-ux-advice list. So, please keep
up the good work ... we really need people who are interested in
improving LibO step-by-step.

> I'll send a message to the UX list later today asking what they think.

Hehe, I did some quick analysis with the data available...

Some notes:
      * Items:
              * AcceleratorExecute = keyboard shortcut
              * GenericToolbarController = application menu
              * MenuBarManager = context menu / menu bar
      * The absolute numbers are not that helpful, so we have to
        normalize them. --> Thus, the toolbar element "File Open" refers
        to 100%.

        *** Comparison Writer
        GenericToolbarController .uno:Open 110056 (100%)
        
        *** Writer (.uno:ViewDataSourceBrowser)
        AcceleratorExecute 18230 (16%)
        GenericToolbarController 5744 (5,2%)
        MenuBarManager 985 (0,9%)
        
        *** Calc (.uno:ViewDataSourceBrowser)
        AcceleratorExecute 10691 (9,7%)
        GenericToolbarController 2362 (2,1%)
        MenuBarManager 799 (0,7%)

So what can we see here? The data suggests that most people use the
keyboard shortcuts instead of the menu or the toolbar items. So we might
guess that the removal of the toolbar icon is okay ... But, the buttons
seem to be used more often (4 ... 5 times) than the corresponding
application menu items - still a lot. Sometimes this doesn't matter that
much - so it would be interesting how the use of these toolbar buttons
relate to all the toolbar items visible for the user.

Personally, this means a fifty/fifty chance to get an improvement - if
the item is removed. To me, it would make sense to have a more complete
analysis concerning the UI (like [2] for Impress).

My very personal view: I think the usual LibO users will benefit if the
item is removed. Instead, we should try to fix the position of the data
sources explorer (currently in the View menu), but there is no easy fix
(in the OOo UX team, we discussed such issues lengthly ...).

Did that help somehow, at least being a first step?

Cheers,
Christoph

> Samuel Cantrell
>
> On Sun, Jun 12, 2011 at 5:08 AM, Christoph Noack <christoph@dogmatux.com> wrote:
> > Hi Alex, hi Samuel!
> >
> > Am Sonntag, den 12.06.2011, 10:22 +0200 schrieb Alexander Thurgood:
> > > > The data sources explorer is still accessible from the View menu,
> > > and
> > > > my patch didn't affect that.
> > > >
> > > > I agree that data sources are useful. I don't necessarily agree that
> > > > it's necessary to have an extra button on the standard toolbar that
> > > > users have to decipher its meaning so they can find the buttons they
> > > > want.
> > >
> > > Often a clearer or more significant icon will do wonders, but that is
> > > just my opinion, I'm not an UI designer.
> >
> > I think what the author of the original idea (remove that button) meant
> > was, that most people simply don't use that functionality - in such
> > cases a more significant icon won't help (then, we would need approx.
> > 500 more significant buttons). Quite the contrary, since such changes
> > might interfere with the more often used features.
> >
> > Of course, this is not the only hypothesis - maybe the functionality
> > doesn't work well and people don't use it, maybe people don't know what
> > it is for and miss it, maybe ...
> >
> > > This is not a gripe at you in particular, but I have this terribly
> > > awful
> > > sinking feeling that Base is being slowly pushed back into a state
> > > where
> > > it will be soon of no use to anyone. UI changes are one of the first
> > > most visible signs of such a slide IMHO. First, we remove the buttons,
> > > then people think, oh, you can't do db stuff from the spreadsheet or
> > > wordprocessor, so just forget the db thing, etc, etc, well you get my
> > > drift. I apologise if I appear paranoid here, and like I say, this is
> > > not directed at you in person, but this is a real fear that I have,
> > > and
> > > so far my fears have not been allayed.
> >
> > I understand your "fears" - but this is the problem having a software
> > that intends to solve the "one size fits all" issue (which won't be
> > solved ever). Thus, its crucial to make wise decisions for the
> > LibreOffice core product - target users and data can help here.
> >
> > Concerning the recent change - I'm also a bit worried, since nobody
> > discussed it in advance (at least I wasn't aware of). That's the crux
> > for people like you Samual, really wanting to improve things ... but
> > having individual voices asking for changes. So, for such stuff, I
> > propose to ping the mailing list "libreoffice-ux-advice" [1] that has
> > been set up by Thorsten. Maybe we can sort this out together.
> >
> > Cheers,
> > Christoph
> >
> > [1] http://lists.freedesktop.org/mailman/listinfo/libreoffice-ux-advise
> >
> > _______________________________________________
> > LibreOffice mailing list
> > LibreOffice@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/libreoffice