Here is the draft of the first of my GSoC applications. I've had a little feedback on IRC, and would like to get some in depth feedback.
EBo -- ===================================================================== Name and Email Address Name: John (EBo) David Email: [email protected] IRC: ebo on freenode phone: (h)+1-337-237-8688 (c)+1-860-753-1506 Institution and location School: University of Connecticut Department: Department of Computer Science and Engineering Address: 371 Fairfield Road, Unit 2155 Storrs, CT 06269-2155 Note: I typically only sleep 4 hours a night, so I can adjust my schedule to accommodate my mentors preferences. Also, I have completed all my coursework and am working on my research and writing my thesis while living in Louisiana (CDT = UTC−06). Bio, Résumé, or C.V. My research focus is on the interplay between ecological modeling and applied restoration ecology. Hierarchical Patch Dynamics (Wu and Loucks, 1995) provides a theoretical framework in which to approach a number of real world problems such as the effects of climate change on species distribution, biological invasion, functional and structural reorganization due to species assemblage, and understanding the role of disturbance on natural and anthropogenic regimes on ecological systems. I am actively working to use these theoretical concepts to develop tools to aid in problems found in restoration and conservation ecology. I am currently looking to reimplement some prior work in spatially explicit multiscale modeling (David 2002; Wu and David 2002). I am currently investigating using Plan 9 and/or Inferno as they provide an elegant functional alternatives to MPI and Globus. My Résumé can be found at http://pages.swcp.com/~ebo/papers/John_David_CV_1003.pdf Why are Plan 9 and related technologies interesting to you? My curiosity in Plan 9 and Inferno was first piqued after reading Eric van Hensbergen et al's paper on "Petascale Plan 9 on Blue Gene". There I learned of using Plan 9 as a functional replacement for MPI and GLOBUS. I was further excited to learn of Limbo which built concurrency primitives into the language, and projects like Chris Locke's Styx-on-a-Brick which built distributed file sharing primitives in an embedded system with only 32K of memory. Together these imply an efficiency and elegance which would greatly help my spatially explicit multiscale modeling efforts. Proof of Installation I have posted my email address on tcp!9srv.net!17037 in the file ebo. I have provided patches to plan9port and glendix-sources ebuild for Gentoo's portage system (See glendix-sources on the Sunrise overlay Bugzilla #303153 and #273890). I have also built a kernel to test Erik Quanstrom patches to etheryuk, and begun tracing through to discover why the device on my primary development machine will not work. Title Plan 9: Support tools for third-party per-package installs Abstract: As Plan 9 continues to grow there has been ongoing debates on how to manage applications both in Plan 9 and Plan 9 third-party applications. These debates have included not only maintenance within Plan 9 and Inferno, but its applicability to hosted operating systems and virtual machines as well. The overall objective of this project is to create a proof of concept per-package installation in Plan 9, and to reimplement this as part of a host OS package maintenance system (namely plan9port on Gentoo using the portage ebuild tools). Long-term benefits to the community include maintaining ebuilds for plan9port, 9vx, inferno, Glendix, as well as those that are Plan 9 specific. It is likely that this can be implemented on top of FGB's contrib, however some features may require significant changes or rewrites. These changes would also have important implications on hosted OS support and long term maintenance. Schedule Bonding period: continue discussions regarding per-package install process on Plan 9 and Inferno, Familiarize myself with Catalyst, UNetbootin, Plan9 disk partitioning, and 9load which are necessary to create a bootable Plan9 distro. Week 1: look into per-package build suite for Plan 9, and implement a singe canonical example. Week 2: finalize per-package build test on Plan 9. While this is not strictly necessary for a Gentoo/Glendix distribution, providing a new tool which changes the way packages are installed in Plan 9 could have serious implications to any Gentoo ebuild maintenance. If the test is successful, then support eclass functions would be provided to facilitate the per-package install using portage. Request initial Plan9/Inferno community review. Week 3: update all Gentoo ebuilds for hosted OS glendix-sources, plan9port and 9vx. Request initial Gentoo community review. Week 4: build a standard Plan9 distro. This would be intended as a proof that the tools and process works for building distro's. Week 5: integrate initial packages into the test distro. Week 6: port more plan9 libraries and applications to per-package install process. Week 7: (mid-term) build a test per-package managed ISO. Week 8: solicit formal Plan 9 and Gentoo community reviews. Week 9: incorporate bug fixes and changes. Week 10: build final project release of the per-package install scripts, the test distro, and Gentoo ebuilds. Notes: The primary focus of this proposal is NOT to make a new distro; that is just a method of performing extensive testing. The core intent is to provide per-package installation and compatible hosted OS support. In addition to the above, the following issues will be address if time permits: *) obtaining Gentoo developer status -- I have recently found a Gentoo mentor and am negotiating taking over the maintenance of the plan9port ebuilds which currently have a status of maintainer-wanted. This will facilitate long term maintenance. *) testing building a clean mini distro on some small embedded machine. Prior Work: I have written ebuilds for glendix-sources, and have updated those for plan9port, and these have gone through a limited community review. Availability I am free for most, if not all, of the summer, and expect to be free to work full time (40 hrs/wk). Since this work is related to my thesis research, and is intended to be my summer employment, it is realistic for me to spend full time on this project over the summer. My only potential conflict is an abstract which was submitted to the International Meeting of the Society of Wetlands Scientists (June 27-July 2). I have not yet heard if this has been accepted, but if this is a problem I have already arranged to have one of my coauthors present the talk. Mentor communication I prefer primary communication via email for technical details or phone for brainstorming (which I am willing to cover any long distant charges). I find IRC to disruptive while focusing on programming for long periods of time. I do use IRC, however, for quick immediate questions. I would use blogs and/or mailing lists for status reports and documentation. I often work with little supervision for weeks or months at a time once I have a clear picture of where my work is going. That can usually be negotiated in the first couple of weeks. During this "bonding" period I expect that I will be communicating with my mentor once a day or more for the first week or two. Past then I should be able to get by with a couple times a week as appropriate for work status updates, or code reviews. If my mentor or Google wishes daily reports I can accommodate that too. If I lost contact with my mentor for an extended period of time and needed help, I would contact the backup mentor or the GSoC Plan9 admin as directed. If I could not get in contact with any of them for another week I would then try to contact other Plan9 mentors. After another week I would seriously consider contacting GSoC directly as a last-ditch effort. Suggested Mentor Anthony Sorace Footnote A: [A] The initial paragraph was taken from Gentoo's GSoF 2010 Ideas wiki: http://en.gentoo-wiki.com/wiki/Google_Summer_of_Code_2009_ideas#Glendix:_Create_a_lean_distro_based_on_Gentoo_and_Plan_9_.28Glentoo.3F.29 -- You received this message because you are subscribed to the Google Groups "Plan 9 Google Summer of Code" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/plan9-gsoc?hl=en.
