Sebastian,

Thanks for sharing this. I found the thesis interesting.

There are several points that are worth take into account:

He says when analyzing the possible outcomes of his thesis that:

"Putting paradigms aside, another possible outcome is that the language��s,

the IDE��s and/or the toolkit��s role in GUI development is much more crucial

than the role of the paradigm."

Which I think is probably the case of Pharo.

Also in another part:

"The existence of GUI builders for nearly all popular GUI toolkits demonstrates

that there are reasons for avoiding manual layout specifications in

code. Of course, GUI builders have some inherent advantages9 such as easier

discovery of all the widgets, immediate design feedback for prototyping and

little need to learn the layout API which in principle allows non-programmers

to create the layout. Still, the question is whether the paradigm, language

(environment) or toolkit can reduce the demand for a GUI builder which

would lead to simpler tooling requirements or, at least, to a convenient

alternative for situations where for some reason or another a GUI builder

cannot be used."

And I think in the case if Pharo we should have a combination of both.

Also the discussion of whether the UI toolkit should have an implementation language different of the DSM or if it should have the same is interesting.

Again thanks for sharing.

cheers

Nacho

Lic. Ignacio Sniechowski, MBA

Prosavic SRL

Tel: (011) 4542-6714

On Sun, Feb 15, 2015 at 10:08 AM, Sebastian Sastre <sebastian@flowingconcept.com> wrote:

Hi guys,��

I saw the interest in GUI lately and today I come across this really well researched master thesis paper that can put some light on how good we can perform in the GUI spectrum in terms of practicality, expressivity and productivity

Introduction

Even though OOP was born with the SIMULA programming language in the 1960s [15], it was Smalltalk that coupled GUI development with the OOP paradigm [25]. Smalltalk��s inventor Alan Kay was inspired by SIMULA, among other influences, to create a new language for graphics-oriented applications leveraging the OOP paradigm �� and the rest is history [24]. Ever since Smalltalk successfully showed the promise of OOP for GUI development a great deal of subsequent OOP languages and, particularly, object-oriented GUI toolkits came into existence and eventually into the mainstream. Nearly all mainstream programming languages today directly support the OOP paradigm and there is hardly any widely used GUI toolkit that does not use OOP. Various popular examples of toolkits like Cocoa, WinForms, Qt, wxWidgets, Tk, Swing and GTK+1 all seem to validate the notion of OOP and GUI development being a good fit.��

continues:��http://www.eugenkiss.com/projects/thesis.pdf

And it seems it triggered this movement of people that are open to receive implementations of the seven dimensions of this benchmark in his repo:

https://github.com/eugenkiss/7guis/wiki

I beleive Pharo can score nicely there and if would be a Pharo implementation there a whole new audience (that we probably aren't reaching now) might take a serious try

from mobile