Difference between revisions of "User:Mitchdraft/GSoC15/Proposal"
(→Why I am interested in this project and why I am the best person to do it:)
(→Demonstration of competency and preparedness)
|Line 122:||Line 122:|
Revision as of 00:31, 27 March 2015
- 1 Personal Information
- 2 Project Information: OpenSCAD Standard Library
- 2.1 Project Description:
- 2.2 Why I am interested in this project and why I am the best person to do it:
- 2.3 Tasks
- 3 Reqs
- Name: Mitch Kelley
- Email-address: email@example.com
- IRC username: mitch
- Contact Number: 508-838-1960
- Blog: https://github.com/mitchdraft/scaddergories/wiki
- Background Information
- Candidate for Master's of Science in Mechanical Engineering at Georgia Tech, currently conducting thesis in improving CAD interfaces
- Bachelors of Science in Mechanical Engineering from MIT, 2011
- 4 Years (and counting) of mechanical engineering experience as a product development engineer (practicing user of Solidworks and other CAD tools in the engineering field)
- Budding software developer - the tools in a software developer's toolbox are extremely valuable in all aspects of life, I am eagerly integrating these tools into my workflow. I have extensive experience in Matlab/Scilab, Processing/OpenFrameworks, and arduino. I have a good understanding of programming principles and am not intimidated by new languages or development environments.
Project Information: OpenSCAD Standard Library
Title: OpenSCAD Standard Library
The OpenSCAD Standard Library requires three main things:
- Alignment of existing library items for ease of use while maintaining backward compatibility
- Definition of a framework for library expansion, involving:
- Identifying the items that should be included
- Cataloging these items in an easily searchable form
- Defining these items in the most expressive way.
- Production of new orthogonal content.
OpenSCAD comes with a number of basic built-in primitives and operators. When creating 3D designs it would be useful to have access to a library of commonly used parts and techniques. One way of improving the current situation would be to define an OpenSCAD Standard Library. This would include well-defined, orthogonal, well-tested and well-documented components. All components would be user-space OpenSCAD code.
Some initial attempts, as well as ongoing work is being on in scad-utils and the dev branch of MCAD.
Difficulty: Medium. The main challenge is to have good enough structure, documentation and testing in order to have trustworthy and useful components.
Customized and augmented by student:
This would include: identifying a base set of mechanical components, identifying variants in an organized hierarchy/network, establishing a module format and documentation style, seeding the framework with detailed representative modules, generating several instances of parts and assemblies to serve as usage examples, documenting how to contribute and soliciting/coaching community involvement.
Documentation will be very important. At minimum, I would have a photo of a rendered part for each module, these photos linked to source code and "neighboring" parts. A key challenge will be making a framework for sorting and finding parts.
Why I am interested in this project and why I am the best person to do it:
I have a professional interest in this project:
In my work as a product development engineer at a consulting company we often start projects by doing an orthogonal technology review. It's often one of the most important parts of the project, setting up a framework for analysis. I have gone through the process many times, mentored by industry experts. I thoroughly enjoy and have become adept at features identification, analysis, and implementation. I would like to apply this process to creating an OpenSCAD parts library. I have seen the way good frameworks can enhance productivity and creativity. I know a library like this will be useful to my colleagues and the growing maker community.
I have an academic interest in this project:
In my research as a master's candidate under Professor Rosen at Georgia Tech I am working to improve CAD funcationality and useability. My specific thesis topic is establishing a framework for and evaluating the utility of Human-Computer Interface (HCI) devices for CAD and other technical, creative work. This work involves developing a CAD environment to represent common practice user interactions and test new interactions and features enabled by the integration of software and existing and new interface devices. Professor Rosen is an authority on the topic of modern CAD systems. He has been a great mentor, identifying the state of the art and avenues towards the improvement of CAD systems, focusing the work towards areas of maximum benefit. In this project I have the opportunity to integrate insights of experts, community members, and an analytical framework into a fully functional open source platform.
I have a personal interest in this project:
My personal objective is to help people identify and achieve their goals. I believe that Computer Aided Design (including 3D geometry modeling and many other computer-facilitated organization and evaluation systems) is the best tool for expressing the imagination. I enthusiastically believe that it is true that "if you can dream it you can do it." And since dreaming is difficult and important I want to do whatever I can to help people realize their personal simple and fantastic dreams.
There are three main tasks in this project:
- Integrate Existing Libraries
- Include complete fastener kit
- Abstract the library framework
Integrate Existing Libraries
Hyperair has done a good job streamlining the existing MCAD library. The contents of the updated MCAD library feature 25 directories. There are many parts in these libraries, and many more on Thingiverse. It is difficult for the average user to find and deploy parts from this library. I will make an interactive catalog that will make it easy for users to scan through the available parts and manipulate their parameters in real time.
I have made a prototype of this tool using OpenSCAD's live render preview feature. You can see the tool in action here.
I will be able to move quickly in this task because Hyperair has done a good job collecting the various parts in the existing library and I have already identified the key features of the catalog tool and proven the concept. By the time I finish this task (~1 month) I will have an intimate knowledge of the current library. I will have gained insights into how to wrap it into a more holistic framework and identified areas for improvement.
- Develop an interactive tool for rapidly navigating and selecting customized parts from the parts library
- Add support for all parts in the current MCAD library
- Languages used:
- Processing - to make the library interface which will output a text file for rendering by OpenSCAD.
- OpenSCAD - for modifying existing MCAD code
- Expected time:
- 1 Month
Include complete fastener kit
Fasteners are a key part to most mechanical designs. Mechanical engineers spend a considerable amount of time selecting and integrating fasteners into designs. This can significantly impede the prototyping process. Fortunately, fasteners are well standardized and cataloged, so OpenSCAD is a perfect platform for automated fastener selection and integration.
While it can be interesting and, at times necessary, to generate complete models of thread geometry (ex: for 3D printing a model of a screw), in most cases, mechanical designers only care about the fastener's mechanical properties and volume occupancy. Complex thread representation can slow render rates. Many professional CAD tools like Solidworks offer "cosmetic threads" which replace thread meshes with textures mapped to simple geometry. A few cylinders are capable of capturing the key dimensions of most fasteners.
Fastening hardware is well standardized. Most hardware vendors have a common core of fastener offerings. Many vendors offer CAD models of these parts too. It is helpful for the mechanical designer to find a part in an online catalog and download the CAD model to test for interference. Often these vendors only provide the complicated full geometry representations. Many instances of these parts can make model manipulations sluggish. Furthermore, the CAD designer often has to specify the fastener mounting location, clearance holes, and mating hardware. It can take an adept designer several minutes to define the parameters for simple fastener application.
I would like to develop a tool for calling a fastener by the desired parameters (strength, material spec, screw driver type, etc.). The tool would filter the the fastener list and prompt the user to select an item or provide more filters. This would be similar to the parts filtering interface on McMaster.
Additionally, I would like to develop a function for specifying a fastener assembly (nut, bolt, washers) by position and fastener name. The function would make the necessary clearance holes in the existing assembly and add the fastener assembly geometry to the model.
- Create a simple hardware selection tool (build into the interactive library tool)
- Create .scad representations for all nuts, bolts, and washers in a leading hardware supplier's catalog
- Languages used:
- Processing - to make the fastener selection tool and integrate with the library tool.
- OpenSCAD - for generating hardware templates and
- SQLite(?) - for storing hardware catalog (could use .scad instead, will discuss with mentors)
- Data mining (TBD) - for collecting hardware parameters
- TBD - for converting available, verbose, CAD files to .scad hardware representations
- Expected time:
- 6 weeks
Abstract the library framework
There is a good amount of design work to be done to generate an effective standard library. A good way to approach this type of project is with iterative prototypes. I view the current MCAD library as a strong initial prototype. I hope to improve MCAD with my contribution but I would be naive to believe that the best library can be produced in a three month project and I would be negligent to discard lessons learned. I intend to sketch-out/mock-up a larger framework for the standard library based on my experience developing the features of Tasks 1 and 2.
- Integrate and abstract lessons learned to a larger library framework draft.
- Languages used:
- Processing - for extending the Task 1 & 2 library interface.
- OpenSCAD - for generating seed models for the abstracted library framework.
- Expected time:
- 2 weeks
- Personal Information
- E-mail address
- IRC username
- Brief background info
- (optional) Link to resumé
- Project Information
- Project Title
- Brief project summary (<500 words)
- Detailed project description (>500 words)
- Links to any code or algorithms you intend to use
- Deliverables (specific, measurable goals)
- Development schedule
- List at least three milestones
- Describe time availability (40+ hours/week assumed)
- List all commitments (e.g., exams, vacations)
- (optional) Why us?
- (optional) Why you?
- (optional) Anything else?