[Sugar-devel] modified Home View

[Gary Martin]

On 9 Aug 2010, at 15:37, Christian Marc Schmidt <christianmarc at gmail.com> wrote:

> 
> (1) Can we reach consensus re spiraling in from the MAXIMUM radius
> once the Ring is full or spiraling out from the MINIMUM radius once
> the Ring is full?
> 
> I would say spiraling in from max radius. That way we can maximize the efficiency of the ring, before transitioning to the spiral.

Oh well I guess I just got out voted (think max inwards looks ugly, backwards, unexpected for a spiral) ;)

> Walter, looking at your mockups I'd try to come up with an algorithm that gives us a looser spiral with more space between each segment of the spiral, more along the lines of what Gary mocked up. The screenshots on the wiki look very dense. Gary's mockup really proved to me that this can work!

Yes, maybe that will help remove all the dead inner white space when using the max inward spiral.

>  (2) Is is OK to add an INTERMEDIATE icon size between STANDARD and SMALL?
> 
> I'd even go further and suggest that we could have icons scale dynamically within the ring/spiral, to achieve maximum balance between the available space and icon legibility.

FWIW I was testing Walters latest patch today and noted that the existing original ring code did much more with icon scaling than the current patch. In the low icon number count, icons start large and then gradually reduce in size before filling the minimum radius. Roughly the first 13 icons form the min radius ring are at their largest size, then up to about 19 icons the they slowly reduce in size down to the next preset standard, at which point the ring starts growing in radius. Once the radius hits max, I believe that's when the icons start to scale down further (not sure if this is a sudden size change to another small default or some gradual reduction). 

Walter, your latest patch does show a big visual jump once the large icons reach the maximum radius, as that point it triggers a large icon size step down and the ring goes from full height to about half height with on extra icon.

Am I correct in assuming that the style.STANDARD_ICON_SIZE, MEDIUM, and SMALL are there to improve icon rendering/cache efficiency? I vaguely remember some pre-rendering to set sizes discussion/work some time back. Not sure if we mess that up by using arbitrary values?

>  My only concern here is that we'd need to find the right balance so as not to interrupt the general zoom metaphor, going from large to small icons (home to neighborhood). This means we probably need to put a cap on the bottom end of the scale, not allowing icons to become too small so that they could reach the size of icons in the groups view...
> 
> All this will take lots of exploration I think before getting it right. I can work on mockups if that would help...

I'd say apply the patch and/or tinker with the current code (it's a single file, and just one short method* that pretty much fits in one page of source), I found trying to layout icons accurately in a spiral for a mockup quite an art in itself ;)

* _calculate_radius_and_icon_size is the method, and it's in /usr/lib/python2.6/site-packages/jarabe/desktop/favoriteslayout.py

>  
> (3) In regard to Q1, I could trigger the spiral before the Ring hits
> the MAXIMUM radius, perhaps at MAXIMUM-icon_size? (I've not
> illustrated this yet.)
> 
> Yes, I think that probably makes sense. We should play through all the possibilities and then make a decision based on what works best...

I was looking into growing the ring up from the min radius (large icons) up to the half way point between max and min radius. Icons would then smoothly reduce to the next standard size down. The spiral would then trigger, and grow both outwards and inwards. Once max/min are reached, icons are then gradually down sized again to fit. Might just be easier/close enough to use the old ring code and trigger something close to Walter's v1 spiral code at the halfway between max and min radius. 

--Gary  

>  
> -walter
> 
> --
> Walter Bender
> Sugar Labs
> http://www.sugarlabs.org
> 
> 
> 
> -- 
> anything at christianmarcschmidt.com
> 917/ 575 0013
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sugarlabs.org/archive/sugar-devel/attachments/20100809/ead1d56a/attachment.htm