User:Crdueck

Personal Info

Hello, my name is Chris. I'm a first year mathematics student at the University of Waterloo (Canada) with a strong interest in computer science and programming.

Contact Info

email: crdueck@uwaterloo.ca

irc: crdueck (find me on freenode, I'll be in #brlcad)

Programming Experience

I have experience with C (as well as Scheme and Python), and would be more than happy to have an opportunity to learn some C++ while working on my project.

Most of my experience so far has come from school assignments, projecteuler.net (programming challenges with a mathematical twist, check it out if you haven't heard of it before), and small personal pet projects.

I have knowledge of data structures and algorithms from my courses at school.

I have a solid foundation in basic university level mathematics (one of my primary reasons for wanting to participate in GSoC is to further my mathematical knowledge in the subject of computer generated models)

Project Summary

Primitive Volume and Centroid Fuctions

Currently, most primitives in BRLCAD do not have volume, centroid, or surface area functions. My goal is to implement volume and centroid functions for many of these primitives.

The primitives I have chosen are mostly simple solids of revolution or other simple shapes and thus it will be easy to construct explicit formulas for their volumes and centroids.

As well, I will be refactoring major sections of mged's analyze command so that it makes use of the functions I will be adding. I also plan on cleaning up the comments found in the source code that I will be working on throughout my project.

Proposal

My primary references for this project will be found in src/librt/primitives. I will be adding the new callback functions to the existing source code for each primitive. The relevant parameters that I will use for computing the volume and centroids will come from the rb_*_internal struct. I will make use of the existing API functions to manipulate this struct and its parameters as much as possible. Using my knowledge of solids of revolution, I will then derive formulas for the volume and centroid for each primitive and implement these as readably, concisely and efficiently as possible.

My reasoning behind choosing to implement both a volume and centroid for some of the primitives as opposed to just one function for all primitives is that I believe it will be easier to implement the second function after I've already spent time becoming familiar with the primitive through my work in implementing the first. In this way I can implement twice the number of new functions for the cost of becoming acquainted with one new primitive.

I plan to test my new functions using mged's analyze command (as this is how the volume and centroid information will primarily be presented to the user. However, separate mged commands could be created to specifically display either the volume or centroid of a primitive. This is outside the scope of my project, but a potential focus for after GSoC). To this end, I will be rewriting substantial portions of libged/analyze.c in order to make use of my newly added functions. There are many opportunities for refactoring in the analyze_{primitive} functions and some opportunities to write new analyze_* functions (there is currently no analyze_pipe). My work will make the source code for analyze considerably clearer and more concise.

I will develop a series of test cases for each primitive using analyze to ensure that my functions are working as expected. This will include verifying that my volume/centroid functions are returning correct values, and verifying that the other outputs of analyze are correct as well.

While I am working on the source code for a primitive, I will take time to clean up the comments and move them from the *.c source code file to the appropriate *.h API header file. I will make changes to the formatting if needed, check for typos and other such documentation related tasks. This is a simple task that will not detract from my main focus of implementing volume and centroid functions.

During the summer I expect to devote much of my time to fulfilling this project. I have no major prior engagements and will be able to spend a full 40+ hours per week on my commitments to BRLCAD. During this time I will make a strong effort to remain in close contact with my mentor, providing updates on my progress, as well as any obstacles I come across or questions I may have related to the project. I will be active on the #brlcad channel and mailing list in an effort to hopefully contribute to the growth of BRLCAD outside the scope of my project.

Deliverables

• Volume and centroid functions for the following:
  • ell (and sph)
  • tor
  • rec
  • rcc
  • hyp
  • arb8
  • arbn
  • pipe

• Improved documentation for each .{c,cpp} file in:
  • ell
  • tor
  • rec
  • tgc
  • hyp
  • arb8
  • arbn
  • pipe

• Refactored analyze_* functions in libged/analyze.c for:
  • ell
  • tor
  • tgc
  • hyp
  • arb
  • pipe

Timeline

• Present - May 21
  • compile BRLCAD from SVN source, ensure build environment is working
  • become familiar with submitting patches to SVN
  • become familiar with source code in librt/primitives and libged/analyze.c
  • get comfortable with using mged to generate and test primitives
  • participate with the BRLCAD community on irc, mailing list etc

• Week 1 (May 22nd to 27th)
  • implement a volume and centroid function for ell
  • rewrite sections of analyze_ell
  • test implementation for correctness

• Week 2 (May 28th to June 3rd)
  • implement a volume and centroid function for tor
  • rewrite sections of analyze_tor
  • test implementation for correctness

• Week 3 (June 4th to 10th)
  • implement a volume and centroid function for rec and rcc
  • rewrite sections of analyze_tgc
  • test implementation for correctness

• Week 4/5 (June 11th to 24th)
  • implement a volume and centroid function for arb8 and arbn
  • rewrite sections of analyze_arb
  • test implementation for correctness

• Week 6 (June 25th to July 1st)
  • move comments from each of the primitives visited so far to the appropriate header files

• Week 7 (July 2nd to 8th)
  • time to review code, more testing

• Week 8 (July 9th to 15th)
  • prepare midterm evaluation
  • discuss with my mentor which primitives I have chosen to implement surface area functions for

• Week 9/10 (July 16th to 22nd)
  • implement a volume and centroid function for pipe (lin_pipe and bend_pipe)
  • write analyze_pipe
  • test implementation for correctness

• Week 11/12 (July 23rd to August 5th)
  • implement surface area functions for chosen primitives
  • ~to be updated with list of primitives~
  • test implementations for correctness

• Week 13 (August 6th to 12th)
  • move comments from additional primitives to the appropriate header files

• Week 14/15 (August 13th to 26th)
  • final testing, code review to ensure all functions are working as expected
  • prepare final evaluation

Me & GSoC

I want to participate in GSoC 2012 for many reasons. Firstly, to develop a stronger understanding of how mathematics is implemented through programming in the real world. My courses at school consist of a lot of theory, and I haven't seen much of the application side of things yet. When searching for a potential organization to apply to, I knew I wanted to work on mathematics related software, and BRLCAD caught my eye right away.

I also hope to greatly expand my programming skills in a real world development setting. I feel that I can improve myself as a programmer by leaps and bounds this summer by working with the devs at BRLCAD. This is also a great chance to get involved with the open-source community. As an avid user of open-source software, contributing back to an open-source project has been something I have wished to pursue for a while now, but am just recently confident enough in my skills to be able to meaningfully contribute.

If accepted, I plan on continuing my involvement with BRLCAD after GSoC ends, moving on to more challenging tasks. In particular this project interests me as I have a good deal of experience working with trees from my Data Structures course at school. Having already invested some time familiarizing myself with BRLCAD's source library, I'm considering completing some of my proposed goals this summer even if my application is unsuccessful.

My Patches

As a warm up to gain knowledge about the internals of BRLCAD, I decided to implement a volume function for a simple primitive. After some discussion on #brlcad, I chose a ell for its simplicity. You can find my patch notes here.

After gaining some familiarity with how primitives are handled, I decided to implement a surface area function for ell as well. Patch notes can be found here.

If you have any comments on these patches please let me know :)