On Wed, 28 Nov 2012 10:18:20 -0200 Bruno Dilly <[email protected]> said:
> On Tue, Nov 27, 2012 at 6:37 AM, Carsten Haitzler <[email protected]> > wrote: > > On Mon, 26 Nov 2012 21:34:13 -0200 Bruno Dilly <[email protected]> said: > > > >> Hi, > >> > >> I've just posted about it. > >> http://brunodilly.org/blog/ephysics_update/ > >> > >> Raster, I would be glad if you could take a look. I know you are very > >> busy, but would be great to hear what do you think about these new > >> features. > >> We have implemented all the tests / examples you have asked when > >> Barbieri and you discussed at Samsung a few months ago. > > > > i'm just loving watching the video. very awesome. :) i like all of it. cloth > > and mesh deformations are a bit of a problem in that they produce a tonne of > > objects so they probably need to always have controls on how much to divide > > - and maybe even "automatic" division based on detected framerates. (if > > things are too slow re-divide with less divisions etc. - is this doable > > given that the current mesh has a "state"? if not then auto mode may need > > some system setting hinting at the general speed vs quality tradeoff to > > use). > > Right now, when creating the cloth you defines the mesh size (number > of rows / cols). > It will generate a triangular mesh of 2 * r * c elements. Dividing > this mesh later isn't possible. > > When a evas object is set to this body it will be "sliced". We'll > create a proxy for each triangle. > Maybe this could be changed depending on framerates. If things goes > very slow it will > create slices again mapping many triangles of the mesh for a single > proxy instead of 1:1. > But this process of re-slicing would be visible for the user and won't > be cheap, it would > delete 2*r*c objects and create all the n proxies again. > > Anyway I would suggest to leave this kind of optimization to be done > after we realize it's required. > Example: "Somebody is working on a application for a smartphone and > things are slow". Even in this > case probably the developer will be interested in simply creating a > mesh with less triangles instead > of having a huge mesh with few proxy evas objects. sure - but for when an app has to span platforms and speed ranges - even if not a phone - a laptop.. it may be a basic pentium or a core i7.. who knows until it actually runs there. :) having an "auto" mode at least should be good where the # of slices is figured out by ephysics - maybe a fixed number. maybe a configureable one system-wide. maybe its calculated based on a combination of factors like fixed speed value AND size of object etc. > > > >> Best regards, > >> > >> -- > >> Bruno Dilly > >> Lead Developer > >> ProFUSION embedded systems > >> http://profusion.mobi > >> > >> ------------------------------------------------------------------------------ > >> Monitor your physical, virtual and cloud infrastructure from a single > >> web console. Get in-depth insight into apps, servers, databases, vmware, > >> SAP, cloud infrastructure, etc. Download 30-day Free Trial. > >> Pricing starts from $795 for 25 servers or applications! > >> http://p.sf.net/sfu/zoho_dev2dev_nov > >> _______________________________________________ > >> enlightenment-devel mailing list > >> [email protected] > >> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > >> > > > > > > -- > > ------------- Codito, ergo sum - "I code, therefore I am" -------------- > > The Rasterman (Carsten Haitzler) [email protected] > > > > > > -- > Bruno Dilly > Lead Developer > ProFUSION embedded systems > http://profusion.mobi > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) [email protected] ------------------------------------------------------------------------------ Keep yourself connected to Go Parallel: INSIGHTS What's next for parallel hardware, programming and related areas? Interviews and blogs by thought leaders keep you ahead of the curve. http://goparallel.sourceforge.net _______________________________________________ enlightenment-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
