[OpenLayers-Dev] transition sandbox - request for comment

[Pierre GIRAUD]

Hi Paul,

I would like to get you informed that I started working on a global
"transition effect manager" for the google-like popups purpose
(specifically for the popup resize and the animated panning). It's
available as a new Tween class and Easing (.Linear, .Exponential,
etc...) classes.

I think this is generic enough to be used in other contexts and as I
was writing this down, I thought that it could be helpful for your
transition sandbox code as it would let you simplify much of it
(startTransition() for example).

You can find :
 - a working example in
http://dev.openlayers.org/sandbox/camptocamp/panto/examples/tween.html
 - tween and easing classes in
http://svn.openlayers.org/sandbox/camptocamp/panto/lib/OpenLayers/Tween.js,
 - example usage in :
http://svn.openlayers.org/sandbox/camptocamp/panto/lib/OpenLayers/Map.js
(new panTo method).

Don't hesitate to ask if you need more details.

Regards,
Pierre

On Jan 15, 2008 2:11 AM, Paul Spencer <pspencer at dmsolutions.ca> wrote:
> Hi all,
>
> I've been reminded a couple of times recently of the work I have been
> doing in the pagameba/transition sandbox.  I would like to complete
> that work in the next few weeks and present it for integration into
> trunk.
>
> In reviewing the current status, I realized that one of the reasons
> why I stopped working on it was that I got blocked by some challenges
> in implementing the animation effect that required more invasive
> changes than I wanted to make at the time.  I'll give a brief overview
> of what changes have been made first, then ask my question.
>
> The purpose of the changes in the sandbox is to address one issue and
> introduce a new feature into OpenLayers.  The issue that I wanted to
> address was the 'flash' on pan of singleTile layers and on zoom of any
> layers.  The new feature that I wanted to create was the ability to
> create transition effects on zoom.  In discussing these issues with
> Erik, we came up with a way of introducing this new transition
> capability in such a way as the default behaviour fixed my first issue
> while providing a framework for implementing the new features.
>
> The basic change that we architected was to create the concept of a
> 'back buffer tile' that sits behind every tile.  The back buffer tile
> is a clone of the tile that is updated through separate logic.  When
> the foreground tile is updated, the back buffer tile is made visible
> as a place holder until the tile has loaded its new image.  Actually,
> it stays visible until the layer has finished loading.
>
> We then scoped out two transition effect, 'resize' and 'animate'.  The
> 'resize' effect simply resizes the back buffer tile image so that its
> geographic extent is correct within the new zoom level.  This
> stretches or compresses the size of the image and you get a visual
> place holder that is more or less correct until the new images
> arrive.  The 'animate' effect does the same thing, except it scales
> the back buffer tile gradually between its old size/location and the
> new size/location.
>
> The current status of the sandbox is that the base architecture and
> the 'resize' effect have been implemented and seem to work well.  The
> 'animate' effect is implemented but it suffers from a couple of issues:
>
> 1) if the layer is panned before zooming, the transition starts where
> the layer was before panning.  This is due to a decision I made to
> update the back buffer tile position when the transition completes
> rather than just before it starts.  I'm sure I had a good reason for
> that, but I can't think of it right now.  This is fixable and I have
> toyed with a couple of ways of fixing it but neither struck me as the
> right way so I am still thinking on this one.
>
> 2) if not all the tiles in the grid are being used for a given zoom
> level, then the back buffer tiles are being left visible and can be
> seen during zooms.  I had originally intended to hide the back buffer
> tile when the main tile was hidden, but doesn't work since the back
> buffer tile needs to be visible while the main tile is loading and
> potentially after it has loaded if the surrounding tiles haven't
> finished loading.
>
> 3) there is some potential for problems if the map is zoomed while the
> map is still animating, depending on the state of the tile loading.
>
> In order to solve these issues, I feel like I have to make some
> intrusive changes that I need some time to consider carefully, and
> some review from those more knowledgeable than me about the
> implications of some of the changes.
>
> I would like your feedback on whether I should prepare a patch with
> the basic back buffer tile support and resize effect (minimizing
> changes to the current code base), and do a separate animation effect
> in a later patch, or wait and do a single patch with everything in it
> (potentially introducing changes with wider impact)?
>
> Cheers,
>
> Paul
> ___________________________________________
>
>     Paul Spencer
>     Chief Technology Officer
>     DM Solutions Group Inc
>     http://www.dmsolutions.ca/
>
> _______________________________________________
> Dev mailing list
> Dev at openlayers.org
> http://openlayers.org/mailman/listinfo/dev
>