Hello!
I have submitted my proposal on Melange. Do take a look whenever possible.
I have not yet understood how exactly am I going to code this thing while
the idea is pretty clear to me. I'm not able to visualize how the code
would look like. I'm working on it and will get back to you about it soon.
With Regards,
Kalpit Thakkar
On Sun, Mar 22, 2015 at 5:07 PM, Kalpit Thakkar <ceasy...@gmail.com> wrote:
> Hey!
> I have been looking at the files "viewweight.c" and "gqa.c" since
> yesterday. So, according to what I understand, I have to implement one
> function (or maybe two) for each of the classes in coreinterface, which
> would use the "rtweight methodology" to calculate the volume and centroid
> of the Database object using sampling and raytracing -- keeping in mind
> that while running the algorithm after getting the object from the tree
> (actually a DAG), I don't mutate the object as it will mutate the tree and
> would create problems (maybe I can make them const, as I would have to
> access their properties only and not perform any mutating operations).
> After this is done, I can add the corresponding virtual functions in
> Object class, enabling runtime polymorphism.
>
> I hope I'm following this correctly. Please correct me if I'm not
> following.
>
> One more thing, does doing this mean I would need a density file for each
> Database Object whose features are to be found? Because, while finding the
> centroid, it uses the weight calculated it seems and we would need a
> density file when we need to find the centroid.
>
> With Regards,
> Kalpit Thakkar
>
> PS : Do you think it would be possible to complete this rtweight part and
> implement two new classes as well or it would be too much to ask given the
> GSoC timeframe?
>
> On Sat, Mar 21, 2015 at 11:56 PM, Kalpit Thakkar <ceasy...@gmail.com>
> wrote:
>
>> Hello Daniel!
>> Yeah, that sounds like a good idea. Essentially, we wouldn't have
>> different solution prototypes for primitives (as combinations use the
>> rtweight methodology for an analytic solution), which would make it easier
>> to add the table functions in the Object class, and the code would also be
>> understood better. Nice! :D
>> However, I will have to take a look at it in detail, because I don't
>> exactly know how the code works in the rtweight methodology.
>> Muchas Gracias :'D
>>
>> Yes, I'll be meticulous in my approach. I would make sure I don't mutate
>> the objects in the tree. :)
>>
>> With Regards,
>> Kalpit Thakkar
>>
>> PS : Whenever you take a look at the proposal, please do tell me if I
>> need to make changes / add something. :)
>>
>> On Sat, Mar 21, 2015 at 11:01 PM, Daniel Roßberg <
>> danielmrossb...@gmail.com> wrote:
>>
>>> Kalpit,
>>>
>>> Implementing the missing functions for centroid, volume and surface in
>>> librt for the primitives can be highly non-trivial. How about
>>> implementing the "rtweight methodology" etc. fall-backs for every
>>> element in librt where an analytic solution is missing? After doing
>>> so you only had to add the appropriate table functions to Object in
>>> the C++ interface.
>>>
>>> However, you should be careful with your algorithm. It shouldn't
>>> change the elements selected with rt_gettree(). This would cause a
>>> dangerous side effect.
>>>
>>>
>>> Regards,
>>> Daniel
>>>
>>>
>>>
>>> 2015-03-21 12:32 GMT+01:00 Kalpit Thakkar <ceasy...@gmail.com>:
>>> > Hello Sean!
>>> > I figured it would be better to get your comments on the doc itself.
>>> So,
>>> > here is the Google Doc link :
>>> >
>>> >
>>> https://docs.google.com/document/d/1RgUDxU3x3IC1r9lba49IlKP9-wJ2PCgAbcwlj7wTlTo/edit?usp=sharing
>>> >
>>> > With Regards,
>>> > Kalpit Thakkar
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Dive into the World of Parallel Programming The Go Parallel Website,
>>> sponsored
>>> by Intel and developed in partnership with Slashdot Media, is your hub
>>> for all
>>> things parallel software development, from weekly thought leadership
>>> blogs to
>>> news, videos, case studies, tutorials and more. Take a look and join the
>>> conversation now. http://goparallel.sourceforge.net/
>>> _______________________________________________
>>> BRL-CAD Developer mailing list
>>> brlcad-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/brlcad-devel
>>>
>>
>>
>
------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the
conversation now. http://goparallel.sourceforge.net/
_______________________________________________
BRL-CAD Developer mailing list
brlcad-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-devel
