Re: GUI design

[Daniel J Sebald]

On 03/27/2012 04:28 PM, Michael Goffioul wrote:
> On Tue, Mar 27, 2012 at 8:59 PM, John W. Eaton<address@hidden>  wrote:
> > The Octave IDE will be used by novices.  It is NOT the place for UI
> > innovation.  We should use GUI/IDE features that are proven, familiar,
> > and easy to work with.  We do not need dockable widgets.  The ability
> > to change the relative sizes of the command, workspace, and history
> > windows would be useful, but we do not need to be able to rearrange
> > their locations within the main window.  Tabbed interfaces will not be
> > used
>
> I'm not going to argue forever about this, so I'll just explain my
> opinion once...
>
> I'm not desperately pushing for dockable widgets, but from my
> experience I would definitely start with a flexible window arrangement
> system. This will define the overall framework for the entire IDE. If
> you don't get it right from the beginning, or go for the more
> straightforward solution of a static window arrangement, it'll be more
> difficult to move away from it later on. I think it's worth spending
> time on that from the start.
>
> The problem I see with a more static layout is that users will most
> likely want to change it. Everybody has his own opinion about how
> window should be arranged, what's important, what should be always
> visible, what is never used... A generic dockable system[1] provides
> the flexibility to make everybody happy, as it's entirely under the
> control of the user.

Guess I'm with Michael on this one.  I don't see the point of aiming for 
GUI Octave.  The author did a nice job, but the features are somewhat 
limited.  If one wants a simple IDE for Octave, they've got it--GUI 
Octave.  The only substance to John's argument is that Octave having 
it's own IDE will mean that it will automatically be bundled with the 
Octave package.

To me, an IDE has to bring something to the table.  It has to convey 
information and organize one's work in a way that makes the user more 
productive.  And the IDE is rather unrestricted.  Rather than limiting 
one's self to GUI Octave or even Mathworks' IDE, think more along the 
lines of not only common features, but new features beyond Mathworks IDE 
and the computer environment in which one will be working.  Here are 
some thoughts and this by no means is the limit of things of substance.

* As Michael says, people like to customize things to there own 
preferences.  Today monitors are much bigger and wider aspect ratio than 
in days past.  Or, two monitors is quite common these days.  So start 
out building the shell of the IDE, non function but with titles for what 
everything is.  Then imagine how people are going to use it.  Will they 
want to pull tabs off and move sections to a different part of the 
screen or onto the second monitor?

* Are there convenient ways to hide and organize windows?  Sometimes I 
get so many windows open on my desktop that I spend a lot of time 
surfing around trying to find the right plot or whatnot.  Could there be 
a feature that condenses the list of figures as a sort of table of 
contents or summary?  Say a menu item lists all the figures with titles 
and as the mouse is placed over the top of a figure name to the side 
pops up thumbnail of the figure.  Or instead of a list, maybe a grid 
layout of all the figures and clicking on the thumbnail pulls up the 
figure plot in full size.

* Honestly, the biggest feature I would guess that people (other than 
me) are looking for is a manually modifiable figure.  E.g., clicking on 
the figure elements, or having a table of the figure's properties that 
can be modified by hand.  I would argue that to make such a thing 
worthwhile, there needs to be a feature to save everything about that 
figure to a file and a way to retrieve that information from an existing 
file.  That at least makes the whole process a bit less ephemeral, so as 
to not spend two hours tweaking a plot just to lose all that work once 
the figure is closed.

* GUI Octave has a variable explorer.  I suppose that is worthwhile, but 
often just plotting the data is easier.  However, I could see the 
variable editor being extended to cells.  Exploring cells and structures 
can be kind of convoluted.  Maybe a tree graph is useful for cells and 
structures.

* A debugger is something a lot of people might like, especially for 
inexperienced sorts who might make rudimentary mistakes that produce 
results not making much sense.  When concentrating on the code in the 
debugger, it is a bit of extra motion to keep glancing over to a 
variable explorer.  What would be nice is stepping through the code and 
being able to put the mouse over the top of a variable in the editor 
window and a dialogue pops up showing the contents of that variable.

* If running under a two monitor system, is there a way a figure window 
could be enlarged to fill the screen, but only the screen as far as the 
first or second monitor.  That would be nice--a quick way of making the 
plot nice and big on the second monitor.

* How about a codec window?  I.e., some way of playing signals appearing 
in a plot codec window to the audio system.  Stop/start buttons, etc. 
All the common features.

* Same thing for video codecs.  Could there be a way of playing a series 
of processed images in a plot window?

* How about if one were to drag a picture image from another program 
(say GIMP or from a browser) into the IDE or paste an image into the IDE 
from a clipboard and that image is automatically loaded into the 
workspace?  Say upon doing that a little dialog window pops up with a 
default variable for that image, say x1, and the option is for the user 
to type in a new variable name for that image.

These are the sorts of things that make an IDE worth having: 
organization, shortcuts, displaying information in a quickly 
comprehendible way.

Dan


> Michael.
>
> [1]: by this, I really mean a system where windows can be docked,
> popped out, split horizontally or vertically or merged into tabs.