+1 to Greg's points.  The use case of low-end systems seems hypothetical.
Thus far, we've prioritized features based on real-world use cases as
pointed out by our users (either on the user list or ourselves in building
real Pivot apps), and I haven't heard any users complaining about the
gradients or transitions.  When coupled with this being a very
de-stabilizing change, it makes me think that we should wait until we build
a new theme to think about this functionality.  Thus, I vote to push this
feature to 2.0.

-T

On Thu, Oct 1, 2009 at 10:40 AM, Greg Brown <[email protected]> wrote:

> To me, this sounds a bit too comprehensive for Pivot 1.x. It would touch a
> lot of code that currently works very well, and I'm not sure that there is a
> strong enough use case to justify the risk of introducing bugs into that
> code. Is poor graphics hardware really a concern for our target users? Even
> if it is, have we actually verified that Pivot runs poorly on such hardware?
>
> The question in my mind is - why might we want to invest time in this, as
> opposed to some of the other issues we have on our to-do list? Personally, I
> see a lot of other features on the road map that would offer a lot more
> value to me than this. I could see potentially considering something like
> this for Pivot 2.0, when we start building out the new theme classes, but it
> seems like a very low priority item right now.
>
> Honestly, I don't think we should consider taking any action on this item
> until we learn more about how our users are using Pivot, and what their pain
> points are. If this isn't one of them, we definitely shouldn't spend any
> time on it. If it is, then we can discuss whether this is the right solution
> or if another approach would be better (maybe for Pivot 2.0, when we start
> working on the new theme classes).
>
> G
>
> On Thursday, October 01, 2009, at 10:12AM, "Sandro Martini" <
> [email protected]> wrote:
> >Ok, I'll try to explain the idea in short terms (some point is already
> >explained in the JIRA ticket):
> >
> >The default behavior will be full effects ON, like currently.
> >
> >This proposal is to disable groups of effects by switching them OFF,
> >depending on what is (not) wanted.
> >We have to see if providing an enum with some common levels (like ALL,
> >MEDIUM, LOW, NONE) but in this case we have to see in what category an
> >effect is ... but this is a detail.
> >
> >The main use case for this is to exclude unwanted graphics effects, to
> >speedup execution of Pivot Applications / Applets, think for example
> >at Corporate (Intranet) Applications where usually the Graphics aspect
> >if less relevant, and usually there are many PCs (old, or at least
> >some year old) with not-so-good Graphics Boards ... like many Intel
> >boards, or with poor drivers.
> >GNome desktop has a feature like this, in the Graphic Configuration,
> >but this is based on the Graphics Board hardware functions ...
> >
> >So, if i could let developers customize (by code and/or in wtkx files)
> >what effects exclude (and maybe disable and re-enable later) to
> >speedup all.
> >
> >With "effects" here I'm thinking on:
> >- media
> >- transitions
> >- transparencies
> >- gradients
> >- backgrounds
> >- watermarks
> >- drawings
> >- maybe others ... like background colors, custom styles, etc ...
> >
> >
> >So a refactoring of some parts of graphics generation code should be
> >done (and this could be useful, in any case) in common methods, and
> >see if make if statements there to see if not process the code to
> >draw, ok ?
> >
> >As a first example, I'd like to try with the gradients part ...
> >probably on modern PC with good Graphics Boards the speedup is
> >minimal, but i think this is another useful feature to have.
> >
> >Or for example think on transitions: if i don't want them, i could
> >disable all them (skipping the related code in new common methods,
> >depending on the related flag), so users doesn't have to wait for
> >transitions to complete (also if is can set their duration as little
> >as possible, ok).
> >
> >
> >Comments ?
> >
> >Bye,
> >Sandro
> >
> >
>

Reply via email to