Hi martin
Dear all,
A few weeks ago, I read about the dispute on Spec, which has escalated, and as
a result of that, Ben has left the Pharo/Smalltalk community. It's always sad
when a long term member leaves, but live is going on.
Indeed.
From my (somewhat business minded) point of view one question is currently
unanswered:
Is there a future of Spec (without Ben)?
Yes. Definitively. :) It will move slower but we will. I started to
clean Spec double unnecessary interpretation loop
Ben was not alone behind Spec. The idea is coming from my experience
writing application in VW. Johan also helped. So we do not stress much.
Esteban has a new Spec interpreter on hold since a long long time. We
cleaned with igor many methods to improve the GC stress when we
introduced new version of LayoutFrame. And Esteban always wanted to do a
pass on Spec. So we will plan that.
If you look we are continuously making sure that the system will evolve
nicely.
some issues are planned on fogzbugs
https://pharo.fogbugz.com/default.asp?13888
https://pharo.fogbugz.com/default.asp?13479
https://pharo.fogbugz.com/default.asp?13851#107599
I'm asking this because I'm the author of a web frontend engine which can be
bound to a programming language (like Pharo) to create desktop-like web
applications.
Super can you tell us more?
Did you look at Phobos?
Currently there is an implementation for the J programming language as well
as a rough (but working) implementation for Pharo. My plans for the (far)
future is to bind the frontend engine to Pharo in a more elegant way probably
by extending Spec.
Excellent. We need clients and we will pay attention to you.
So one may write a desktop application in Spec, and a corresponding web GUI is
automatically created as well.
We want the same for Cocoa Widgets :)
Esteban has a first draft of a brigde so we could define our interface
in Spec and run the with Mars (the bridge) and get Cocoa widgets.
So I'm currently a little unsure whether to go with Spec or going back to the
good old Polymorph.
In fact it depends of your needs.
Glenn started to use spec and went directly to morphic because some
aspects of widgets are not covered by Spec (we should improve the
coverage of the Morph
API - color, selection....).
Now the API of polymorph is not really good because widget creation api
should be widget class and not UITheme. So I started to see how we could
improve the
situation and the first step is that every widgets with labels should be
font parametrized and access a theme. Then the API will be migrated from
UI to the widgets.
I will restart to work on that after 15 of october (or before if I
finish my current proposal writing).
Is somebody of the core team able to clarify the situation?
Let me know if this is ok for you.
Many thanks in advance.
Martin
