Re: Interface localization ?

[Nicolas Roard]

On Nov 23, 2007 8:36 PM, David Chisnall <theraven@sucs.org> wrote:
> On 23 Nov 2007, at 20:00, Nicolas Roard wrote:
>
>
> > On Nov 23, 2007 7:45 PM, Pete French <pete@twisted.org.uk> wrote:
> >>> I wouldn't do that either. It's like writing a master thesis in
> >>> LATeX
> >>
> >> ttoally off topic, but LaTeX is prertty much a requirement in
> >> academia
> >> in the UK still (?!). I thought it had vanished, but my flatmate had
> >> a job editing papers for journals and it is all required to be
> >> submitted
> >> in LaTeX. erp!
> >
> > Frankly I wouldn't want to write a paper or thesis in anything else
> > than LaTeX...
> > Not that we couldn't have great graphical tools for that; but we don't
> > right now.
>
> I'd love to have written my thesis or my book in something better than
> LaTeX.  I fully intend to abandon it as soon as something better exists.

My point :) -- LaTeX is the worst of all, apart from all the others.
(and what a fantastic procrastination tool ! quite appealing to grad
students...)

> Back on-topic, something like LaTeX for GUI design would be great.
> Currently, we have two options for GNUstep:
>
> - Gorm, which is like FrameMaker.  You lay out your page (app) exactly
> as you want it to appear to the end user.
> - Renaissance, which is like HTML.  You lay out your app roughly with
> some markup and hope it will look how you wanted it.

As a side note... I believe that most if not everything you do with
Renaissance you can or should be able to do with Gorm; the main
advantage of Renaissance is the automatic resizing. You could easily
add that capacity to Gorm by adding a palette containing the [h|v]grid
widgets... that renaissance uses (gshbox, etc).

> The ideal solution would be something like LaTeX, where you define the
> interactions you need, and it pulls in 40 years of HCI research to
> design a beautiful UI that embodies these interactions, with a
> \documentclass equivalent so you could make it conform to a set of
> predefined platform HIGs.  Unfortunately, I haven't yet decided
> whether designing such a system is impossible or just really, really
> hard.

Writing such a system could be a challenge, but if for starting we had
precise HCI rules to define a UI... then we could try thinking about
automating them...

-- 
Nicolas Roard