The future of UI, especially the grid
Locked
 
|
I , Arden,
thanks, these are good news for me, As the one fro the new bugfix release model (with store preloaded) ciao Giorgio
On Tue, Dec 6, 2011 at 11:10 PM, Arden Thomas <[hidden email]> wrote:
_______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: The future of UI, especially the grid
|
|
In reply to this post by Arden-8
Hmm, I'm not sure I have a
lot to add, but a couple of points. I think that the GUI in particular
has suffered from a lot of technical debt, something we recognized in
looking at doing Widgetry as a clean break from that. Unfortunately,
that didn't work out, and in the meantime the GUI continued to suffer
neglect. We've attempted to address that since then, but we do end up
spending a lot of time on debt reduction. I think sometimes it can be
difficult to distinguish "nibbling around the edges" from working on the
foundations. A grid-type widget in particular is quite a complex thing,
and most of the existing solutions spend a lot of time working around
issues. Things like getting rid of things that go around the
invalidation mechanism or making keyboard processing be handled
dynamically rather than requiring complex state maintenance are not just
peripheral tasks but things that are important for complex widgets.
We've not been happy enough with our existing Grid efforts to promote
them to production, and tend towards the opinion that producing a
monolithic grid out of duct tape isn't really the right thing, but
rather that there are a number of independent things that need doing and
combine to make a good Grid widget work naturally. And we are working
on them.
However, I'd like to emphatically agree with the point that there are a great many things to be done, and not enough people to do them, and this should be pointed out, preferably repeatedly and forcefully to management :-) Not that this is new. Quite a few years ago at a conference, Vassili Bykov, the lead at the time, showed a diagram describing the workflow. You have three houses, arranged in an equilateral triangle. All of them are on fire. There is a fire hydrant in the centre of the triangle, but it takes time to drag the fire hose over from one house to the other, so you have to make a judgement about which one is burning worst at the moment. And this applies not just within GUI, but of course we have a number of other major things that we have been doing over the past releases with the VM, web services, Store, security, internationalization, and numerous other areas of the product, all of which are quite important for different people.
_______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: The future of UI, especially the grid
|
|
Dear Alan,
> However, I'd like to emphatically agree with the point that there are > a great many things to be done, and not enough people to do them, and > this should be pointed out, preferably repeatedly and forcefully to > management :-) Not that this is new. Quite a few years ago at a > conference, Vassili Bykov, the lead at the time, showed a diagram > describing the workflow. ... I recall Vassili at ESUG 2003 also mentioning his then-manager's remarks to him about "the eight hours you waste every night sleeping" :-) Eliot, by contrast, proposed cloning Vassili. :-) Vassili objected to both these approaches. :-) So please be judicious in exactly how you phrase any attempts to call management's attention to these things. :-) Yours not very seriously Niall Ross _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: The future of UI, especially the grid
|
|
In reply to this post by Arden-8
I have to apologize for being too unfair and harsh in my post. I apologize. I was so angry when I tried to replace the Aragon DataSet with the VW one. I could not believe that it cannot act as drag source. I looked at porting that feature, saw that it would take some time to understand and do, thought “hmmm this is only one needed feature what else is missing?” and reluctantly reinstalled the Aragon DataSet… I do agree with Reinout that there has been progress. In the right direction. The redesign of the tabbing order is great and simplifies the code significantly. I am porting this to Widgetry and it also has a very good effect there. About Panels, I am not quite sure yet. The new tools with them are a big improvement, but it is unrelated to the current UI framework. They implement yet another system of widgets for some kind of specialized tree views without using TreeView… The event refurbishing seems to be a good thing to get away from explicit controllers. I did not take much notice, since event configuration has not been much of a problem for me. Skins… we will see… Alan, your post seems to honestly describe the picture. I can understand the difficulty of the situation. Captain Kirk would, in a desperate and completely unsurvivable situations (like getting sucked into a black hole or recalibrating the sensors J) say: “Scotty, more energy”. This has always worked! The pile of debt the UI has accumulated is indeed huge. Keyboard and events are a start of paying back. This house is burning since 15 years… and still no relief in sight for existing UIs/widgets. I wish Cincom would give the UI more attention and resources. It should be possible to build modern UIs with Smalltalk. Travis, thanks for posting! Christian Von: [hidden email] [mailto:[hidden email]] Im Auftrag von Alan Knight Hmm, I'm not sure I have a lot to add, but a couple of points. I think that the GUI in particular has suffered from a lot of technical debt, something we recognized in looking at doing Widgetry as a clean break from that. Unfortunately, that didn't work out, and in the meantime the GUI continued to suffer neglect. We've attempted to address that since then, but we do end up spending a lot of time on debt reduction. I think sometimes it can be difficult to distinguish "nibbling around the edges" from working on the foundations. A grid-type widget in particular is quite a complex thing, and most of the existing solutions spend a lot of time working around issues. Things like getting rid of things that go around the invalidation mechanism or making keyboard processing be handled dynamically rather than requiring complex state maintenance are not just peripheral tasks but things that are important for complex widgets. We've not been happy enough with our existing Grid efforts to promote them to production, and tend towards the opinion that producing a monolithic grid out of duct tape isn't really the right thing, but rather that there are a number of independent things that need doing and combine to make a good Grid widget work naturally. And we are working on them. [hidden email] Dear folks; We have a number of UI improvements on the way; - Skins (custom widgets with native rendering) - "fluid positioning" layout tools for interface construction - infrastructure updates: we try to constantly incrementally improve the ui infrastructure, so that we update it, but not so drastically that one has to look at an onerous porting effort. Grid has been tough. While maybe not apparent we have had internal progress on the grids, but I'll let Alan comment more when he can. Thank you for the feedback. We know grid is important, and we want to get it right. Regards Arden Arden Thomas Cincom Smalltalk Product Manager 845 296 0686 Cincom Smalltalk - It makes hard things easier, the impossible, possible "Simplicity is the Ultimate Sophistication" - Leonardo Da Vinci _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
| Free forum by Nabble | Edit this page |