2014-10-21 15:20 GMT-03:00 kilon alios <kilon.alios@gmail.com>:
Why are you saying that. You can get halos on any morph, just halos do not show up when you do not want them.

I am saying because of my experience with halos , I have not seen anything new added to them or improved the last 2 years I am using Pharo.

Did you ever look at the code of Halos? Halos are not defined per Morph but in a global tables. So we started to clean them.

No why would I need to looks at the code of halos ? I am its user not its developer. It great you clean the code of Halos but what that means to me as user ? Not much I am afraid.

 Do you think that spec is nothing in terms of dev. Give me 1/1000 of the ressources put on one of this project and we will have a real UIBuilder.

Stef


Thats the problem I see many very good projects like Spec , PharoLauncher, Pillar, AsmJit, Nativeboost , Athens , ��Opal and the list goes on and on. But I see no interest in having a UIBuilder. Apparently GUI designers are very important for most IDEs out there and judging from the developers using them I would say also most developers but is not for Pharo. Why ?��

You say you want the resources of 1/1000 do you really say that you need more than 10 developers to make a UIBuilder ? Here examples of very small development teams that have produced capable IDEs with GUI designers included


Building a good GUI designer is one of those problems where you cannot "solve" with 1-2 unexperienced programmers (and I am not talking about of confident programmers, I mean real experience). Is not about just painting widgets, is about layered APIs, application integration, extending specialized widgets, how to integrate second party widgets, portability, usability, etc. It is a domain where you cannot support the whole learning curve of a post-doc, and expect to see successful results. In short: You have to hire experienced developers.

��



and the list goes on... I dont think that any of them had more than 1 -2 and certainly they were not full time devs. Gambas alone is one developer effort for the entire IDE, language, libraries, GUI designer etc etc.


I didn't check but I wonder how they perform with resizing layouts? Do they use spacers? Do they use constraints?
��
Why such small development teams do dare to implement features that huge IDEs do have ? Mainly because it usually comes down to small teams, take a look at Visual Basic for example, I was reading its history recently and found out that VB was no Microsoft invention (surprise - surprise) but there was actually a very small company that create it as a product that then they were hired from Microsoft. Delphi has a small development team as well, actually it was single developer that leaded the effort and was mainly responsible for its creation.��

So add this to one of my ranting and one of my complains. I love spec , I love nativeboost , I love Playground and new GT tools but none of these will ever start to compete with the usefulness of having a GUI designer when up to 50% of a modern GUI based apps code is creating and managing that GUI.��

The problem with Pharo's GUI designer is not that it sucks, its undocumented , it severely lacks in features. But that it simple does not exist. And without good GUI designer and an undocumented GUI API my remark that on the GUI front Pharo is not in a good state is as honest as it can be.��


Yes, and I prefer not having a GUI builder than having a bad one.

��
On the other hand on other fronts, like web interfacing, databases, calling c libraries , vector graphics, unicode , true type fonts, ��JIT VM even 64 bit Pharo has made a lot more significant progress. So apparently Pharo resources wise seems to do great. Actually I am impressed how fast it progress forward , its amazing.��