Here is the vision: we need it better and simpler with better widgets, better
UIBuilder and better tools.
if you give me some engineers I can build something clear, now this is not the
case so we are exploring and multiple paths
Now if you want to see the system moving faster just join and help
So here is a list
Frameworks
----------
First step
- Integrate polymorph (move class to the right packages, remove
overrides)
- Reduce complexity of morphic when possible
In parallel
- Clean SimpleMorphic (remove preferences, PopUpChoiceMenu)
- Based on SimpleMorphic, clean also Morphic
Alain told me that he want to take simpleMorphic and create a kernel that can
be run and debugged on the side
or morphic. Then he wants to add list and tree and see if he can add a better
version of polymorph.
Now he can fail too. so this is why the incremental points are important.
Fixing and improving announcements is important
- It was done during the last sprint
- We were discussing how to make anonucement faster to avoid to
climbing their parent tree - but no success so far.
- We evaluated signals (as used in HPI frameworks) and we do not
really like it. The use of thisContext is probably a problem among others.
Widgets
-------
We need better widgets
- accordion
- better text
- grid
We should evaluate your widgets to integrate them. But first we should
clean what is there.
Stacking on top of bad foundation just make the system more complex and
difficult to maintain.
Low level: a new Canvas and clean event
-----------------------------------
Igor started to design a new canvas called Athens and we will see where
it will go
Again if people want to see this coming faster, they should help.
The idea is to have a canvas for
- opengl
- cairo
- bitbl
We are evaluating the event hierarchy designed by Mickael Rueger and we
would like
to integrate it: eliminate the floating of arrays btween eventfetcher
and HandMorph
the idea is that event should be emitted by eventSensor not raw array
I'm starting to clean Sensor references that are using the polling
behavior because we should not have polling anymore
the Windows virtual machine should be fixed and get the enhancement
that Igor sent more than a year ago for the input semaphore
and so that all the vm are aligned.
Tools
-----
We are rewriting from scratch the basic tools
- Browser (Nautilus soon to be announced - with groups, package
browsers, refactorings, may be plugin architecture)
- Finder
- TestRunner (soon)
- MCBrowser
- Debugger (waiting for a debugger model)
The idea is that we want to kill the StringHolder hierarchy alltogther
so in a first phase we will probably have to keep it
for the debugger but only for it.
We will remove toolBuilder (we are waiting to finish the TestRunner
rewriting).
Glamour will probably be used as a default super IDE of the future but
it depends on people
We are currently working on the fall back (the tools when you have
nothing).
So I hope it clarifies the picture and you are welcome to help. Now an
important point
This is not because something is under preparation that the other should not
get clean and improve.
We have no problem throwing away our effort if something better come up but we
should be prepared that nothing come up or is delayed
this is why we have this parallel strategy.
Stef
> Many new things is happens this days on Pharo, and I don´t understand the
> status of all.
>
> What is the future of GUI in Pharo? what changes will happen? which is the
> roadmap? Now I see SimpleMorphic. Pharo will work with it? Pharo will works
> with another way of draw/render the morphs?
> We will reduce the number of morph classes and hierarchy? The Morphic UI
> designer will be incorporated in dev image?
>
> Thanks for your answers.
