Re: [Gimp-developer] An update on the menu search

[peter sikking <peter mmiworks net>]

Srihari Sriraman wrote:

> A command system would really boost the speed of interaction.

now that would really depend a lot on how you design this.
I must admit there is a tiny opportunity to come out faster than
mousing the menus for a plugin without a shortcut key (ctrl-...).
where is comes to menu items with shortcut keys, I am not sure you
ever can.

so let's see, you have an opportunity to do something good for
a minority part of the menu structure, which by itself forms 1/3 of the
GIMP interaction system.

is that worth it? 

this of course apart form the help/explore/documentation system we
talked about here and on irc. I thought we had understand each other
during those talks.

> We're headed there.
> Blender's and Rhino's systems are great, so we'll be taking inputs from their design too.

that does not sound like a confident approach to designing this
for GIMP. think about the whole context of GIMP, its vision and
its core users. all the work on the canvas.

it starts with that realisation that you are working on what is
a minority part of  1/3 of GIMP's interaction.

> Currently we're thinking of using a command syntax similar to that of ImageMagick..
> Would be great if we could pool in ideas regarding the command system here...
> (Will probably start a new thread on that soon).
> 
> Just to note, Tito has this handy feature.
> If you enter 2 letters, it matches those actions with those 2 letters as the first letters of its words...
> Ex:
> You can "Apply Canvas" by typing in "AC" and hitting enter.
> Or say "Gaussian Blur" by just "GB".

have you run statistic on how many clashes (same 2 or 3 letters for
different commands) in english? and then on to other languages, where
the 2 or 3 words english needs to describe something (a peculiarity)
is localised with a single word.

there is 76 localisations of GIMP, I would not blame you if it can be
made to work for all languages apart from urdu (or something like that).
but if at the end it only works for 5 languages out of 76, then
it would of course go nowhere.

then on to non-latin input methods, on those standard latin keyboards
computers are sold with. how would that be fast?

    --ps

        founder + principal interaction architect
            man + machine interface works

        http://blog.mmiworks.net: on interaction architecture