Project Tracking

Fall 2005 - COSI for Credit

Hello, my name is Rouslan Solomakhin. Along with Mike McCabe, I am a co-director of Clarkson Open Source Institute. My main task is to maintain Linux build in COSI and ITL labs. It must be stable and accessible to students at all times. Also, Mike and I hold meetings and work-days for benefit of COSI and its members.

Semester Goals

• I would like to follow in the footsteps of the recent TuxMaster Invitational winners, and participate in the challenging contest.
• I am going to be a mentor for Martin. This involves teaching him everything I know about Linux, and showing him around the campus.
• This semester more than ever, I teach all willing both basic Linux skills and advanced concepts. Also, I teach people how to get involved in the community by triaging, submitting, and solving bugs.
• I constantly work on improving both the COSI Linux build, and the system to deliver it to workstations. I am also working on documentation of this matter, as I am not going to be here forever.

I plan to make everything I do this semester useful to me and both global and local open source community.

Current Projects

• Working on OSDL project: Universal Software Management (USM).
• Maintaining the lab build. See COSI Bugs page for status information.

Past Activities

• Held a workshop on Bash scripting for about 10 people. There were a few questions, and I noticed the general usage of Bash going up in the lab.
• Showed Martin around campus, and taught him a few tricks in programming.
• Prof. Hesse, Pat, Chris, and I went to ACM Programming Competition in Montreal. Although we did not win, it was a very good experience in terms of practice for future programming competitions. I would love to practice problems from such competitions in my free time, and participate again next year.

Accomplished

• I would rate my success of mentoring Martin as 60%, mainly because I lacked the skills to motivate him to be in the lab more often. When he did show up, he was doing school work, which is understandable with his busy schedule.
  
  In order to improve in this area, I would need to spend more time with my future mentees.
  
  
• I would rate my success in the maintaining the lab build as 90%. The reason I did not choose 100% lies in limitations of a source-based distribution running on a virtual machine. Gentoo requires compilation constantly, and resources devoted to it on Xen were simply not enough to complete some compilations as seen on bugs page.
  
  In order to improve in this area, we need to move away from a source-based distribution and change the way we distribute it onto the lab machines. Experience has shown that rsync and bash have their limitations and our method of maintaining the lab is error-prone and, consequently, time-consuming. If we are going to switch to Fedora, we have customized rpms for Clarkson's AFS scheme, and expertise in building new ones. If we are to use something else, we'd need to (a) learn how to make packages in the other distribution, and (b) most likely still use files from http://mirror.clarkson.edu/pub/cosi/afs/.
  
  Almost off-topic: many of the lab-build scripts have been converted to Java for robustness, but that is going to be thrown away when moving away from Gentoo. The only benefit here is in my increased skills in Java.
  
  
• As USM project is not done, and is not due until the end of January, I hesitate rating my success in it. Although C# is an interesting language to use in a project like this, I think that it would go a little faster if I used Java.

Overall, I am happy with results of this semester. There is room for improvement, which keeps me motivated.

Last edited by Rouslan Solomakhin on December 8, 2005.