[Soc-2014-dev] Weekly Report #4: NURBS Modernization

Thu Jun 19 11:12:08 CEST 2014

For sure survey wouldn't lead to immediate results. But currently it's more
like you've been mostly offline for the week of survey and then a week when
you've tried to write the code. It's not good and all the difficulties
you're having and stuff you're working on are to be communicated better.

Since there were no commits to either of the repos (neither blender nor
your github repo) the question would be what you're working at this moment,
which issues you're having and what kind of help you need in order to move
further with the tessellator?

On Tue, Jun 17, 2014 at 10:22 AM, Jonathan deWerd <jjoonathan at gmail.com>
wrote:

> Last week was a low-productivity week, there's no denying that. It
> happened due to a handful of factors. A few things to keep in mind:
> * The "chalk-cart" logic is actually still in the display routine, not
> poly.cpp.
> * Learning about Rhino's trim tools didn't produce an immediate
> deliverable, but it almost certainly will be helpful or necessary for the
> midterm.
> * Time spent looking at raster literature hasn't filtered through the
> pipeline yet. Not until tomorrow.
> * Subdivision/tessellation and trimming require different algorithms,
> which meant more reading before I could start writing code. My intent was
> to produce an implementation, not a literature survey, so you don't see a
> pile of text, just a relatively small bit of code. Code that someone with a
> proper computational geometry background could have written and debugged in
> an afternoon rather than days. But I don't come from a computational
> geometry background, I come from a numerics background, so I have to learn.
> * I didn't find IsectLLPt2Df(), so I wound up re-implementing it. I must
> remember to search for these things more thoroughly. It represents a
> sizable fraction of the deliverable but not as large a fraction of the time
> spent producing the deliverable. LL was easy to find, easy to understand,
> easy to write, and didn't need debugging.
> * External: a cold during the later part of the week/weekend. Better now.
>
> I have to scramble to get UGC working before the midterm, but I'm still
> confident I can get it done in time.
>
> Regards,
> Jon
>
>
> On Jun 16, 2014, at 5:21 PM, Sergey Sharybin <sergey.vfx at gmail.com> wrote:
>
> I'm not really sure why it's taking so much time. That's like rather
> simple OGL visualization file and the file which is assumed to have
> trimming code only contains few of the functions. And half of the functions
> does exist in BLI, remaining part doesn't seem to do the actual business.
>
> Getting into account you've spent whole previous week in the research
> looking into algorithm details i'm not really sure what's causing
> difficulties now.
>
>
> On Sat, Jun 14, 2014 at 10:56 PM, Jonathan deWerd <jjoonathan at gmail.com>
> wrote:
>
>> That would be great! Send me 3dm files and I'll try to include them in my
>> midterm. I can't make promises this early: it's possible you make heavy use
>> of a feature that I won't get around to implementing until after the
>> midterm, but I'll do what I can. Either way production 3dm files will
>> provide valuable information regarding what work I should prioritize.
>>
>> Thanks for your interest,
>> Jon
>>
>>
>> On Jun 14, 2014, at 10:59 AM, claas kuhnen <info at ckbrd.de> wrote:
>>
>> I can provide you think models from my design work including meshes from
>> example moi so you can compare if you are interested.
>>
>> I am very excited about this project.
>>
>> Claas
>> On Jun 14, 2014 5:45 AM, "Jonathan deWerd" <jjoonathan at gmail.com> wrote:
>>
>>> This week's progress:
>>> - About 75% of the way to the trimming code for the UGC tessellator.
>>> -- Details:
>>> http://wiki.blender.org/index.php/User:Jjoonathan/Trim_Algorithms
>>> -- Code: https://github.com/jjoonathan/PolyTest/tree/master/PolyTest
>>> - Not sure how I should write up my midterm proposal, but I stated my
>>> intention on the wiki page to put up a gallery of imported 3dm files. I
>>> anticipate being able to show at least one from an interested artist, a
>>> handful from grabcad.com, and a few of my own that really show off
>>> trimming (I've gone through the lynda.com tutorials for the
>>> corresponding parts of Rhino)
>>>
>>> What I want to have done by next friday:
>>> - Finish the trimming code
>>> - Put the UGC+trimming code in blender
>>> - Demos for the midterm to show it off
>>>
>>> On Jun 7, 2014, at 5:50 AM, Jonathan deWerd <jjoonathan at gmail.com>
>>> wrote:
>>>
>>> > This week's progress:
>>> > - Finished integrating opennurbs into cmake
>>> > - Wiki page detailing the literature search and
>>> existing-implementation-review I did for the tessellator
>>> >
>>> http://wiki.blender.org/index.php/User:Jjoonathan/NURBS_Tessellation_Survey
>>> > - Extensive comments in Nurbana (Uniform Grid Cut) and Mesa (Recursive
>>> Subdivision) tessellation code.
>>> >
>>> > What I want to have done by next friday:
>>> > - Surface import with Uniform Grid Cut trimming (yep, slid another
>>> week, as expected last week)
>>> > - Midterm proposal (I really need to get on that!)
>>> >
>>> > On May 31, 2014, at 3:56 AM, Jonathan deWerd <jjoonathan at gmail.com>
>>> wrote:
>>> >
>>> >> This week's progress:
>>> >> - I can successfully load order 2, 3, and 4 NURBS curves! To be
>>> clear: no surfaces yet and I haven't added support for a bunch of .3dm
>>> features, from small ones (not-open curves with irregular knots) to big
>>> ones (trimming). But I'm not held up on these, I'm hung up on...
>>> >> - Integrating opennurbs into cmake. I got my first result by manually
>>> tweaking the Xcode files that came out of cmake but this is no good in
>>> general. I've put opennurbs in the "extern" folder and I'm still working
>>> out how to persuade cmake to generate the appropriate build files.
>>> >> - Got a stakeholder .3dm testfile. It's not production data but it
>>> tests for a specific cut that was needed in production. I have lots of
>>> .3dms now: I have 3dms that I made, 3dms from stakeholders, and lots of
>>> 3dms from the Rhino example libraries.
>>> >>
>>> >> Things that were done this week that weren't progress:
>>> >> - Reading trimming literature
>>> >>
>>> >> What I want to have done by next Friday:
>>> >> - Integrating opennurbs into cmake
>>> >> - Surface import (ideally including trimming although I suspect it
>>> will slide another week)
>>> >> - Midterm proposal (didn't really work on it this week)
>>> >>
>>> >>
>>> >> Cheers,
>>> >> Jon
>>> >
>>>
>>> _______________________________________________
>>> Soc-2014-dev mailing list
>>> Soc-2014-dev at blender.org
>>> http://lists.blender.org/mailman/listinfo/soc-2014-dev
>>>
>> _______________________________________________
>> Soc-2014-dev mailing list
>> Soc-2014-dev at blender.org
>> http://lists.blender.org/mailman/listinfo/soc-2014-dev
>>
>>
>>
>> _______________________________________________
>> Soc-2014-dev mailing list
>> Soc-2014-dev at blender.org
>> http://lists.blender.org/mailman/listinfo/soc-2014-dev
>>
>>
>
>
> --
> With best regards, Sergey Sharybin
>  _______________________________________________
> Soc-2014-dev mailing list
> Soc-2014-dev at blender.org
> http://lists.blender.org/mailman/listinfo/soc-2014-dev
>
>
>
> _______________________________________________
> Soc-2014-dev mailing list
> Soc-2014-dev at blender.org
> http://lists.blender.org/mailman/listinfo/soc-2014-dev
>
>

-- 
With best regards, Sergey Sharybin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.blender.org/pipermail/soc-2014-dev/attachments/20140619/cee0826c/attachment.htm 