Week 6 Summer 2014 (6/24/2014)

Week 6 Summer 2014 (6/24/2014)

We are in the dog days of Summer (past Summer Solstice on June 21st). Despite the heat 9actually it is not that hot), sudents are working hard and the projects are coming along. I need to get in touch with mentors to get more information about the progress of each project (I am still bewildered by the large number of students - my own fault!)

This week we had six talks:

1)  Kit Hammer and Noelle Todd - (Pantry Client) Client DB http://rcos.rpi.edu/projects/food-pantry-client-database/

2) Huiye Liu, Xitu Chen, Eunkyoung Lee, Peiyang Ding, Yiding Zhang - ShengGuanTu  http://rcos.rpi.edu/projects/sheng-guan-tu/  (Please add all group member names in the observatory project page)

3)-Maia Marchetti -  Milkyway@home Expansion http://rcos.rpi.edu/projects/milkywayhome-server-expansion/

4)  Nicholas Pitt - OpenLab  http://rcos.rpi.edu/projects/openlab/

5) Joe Allen and Joe Eduardo - Jojo http://rcos.rpi.edu/projects/jojo/
 

6) Ezra Dowd and Tahsin Loqman - RPI Walk   http://rcos.rpi.edu/projects/rpi-walk/

Kit and Noelle are working on a Food pantry database for a local food shelter. Their software is wanted by at least one client. Both Kit and Noelle know about the customer's needs. Noelle is working on the backend using SQL and PostgressSql and Kit is working on the front end. They have done a preliminary design and implemented portions of it. Since the client has an old PC, the team is designing a sosftware system that can run on any machine. They have also posted their slides. here is a link to it. http://goo.gl/wyreG2  Hope oters follow their footsteps and share their presentations.

Huiye, Xitu, Eunkyoung, Peiyang and Yiding are working on an ancient Chinese game . Here is a techincal paper http://www.jstor.org/stable/4132278  Acccording to the author of this paper, "It can be argued that board games reflect the society in which they are conceived. Thus Monopoly reflects the entrepreneurial spirit of a capitalist society, while the Chinese game of Shengguan tu  or "Promoting Officials" reflects the obstacles confronting officials throughout their careers.' On the scaled-down reality of a Monopoly board, players can  display their entrepreneurial talents by acquiring desirable properties. By increasing the number of buildings they construct on these plots, owners charge ever higher rents until the other players run out of funds. By contrast, success in Shengguan tu is measured not only by the amount of funds that a player accumulates on his way to high office, but also by the prestige that the office confers on the holder." 

 To a novice's eye (that is me), it looked like a snakes and ladder game - but a lot more complicated. This game is also based on Ancient Chinese History and a strateg based game. It is a multiplayer game. They are developing this game in unity. There were a number of suggestions. One of them is to scale down theor game so that tye can build a prototype before the end of the summer. The group has done a number of sceneries involved in the game.

Maia talked about Milkyway@Home Expansion. This is part of the Milkyway@Home project. She has  modified a lua file so that it will simulate several dwarf galaxies at once. She showed a video of this phenomenon having two dwarf galaxies. She has been working on the server code using an alpha project sets up on the milky way@home server.

Nick has been working on an Open Lab Project - open source version of LabVIEW - of course not the full version - but  with only a few features and a few innovations. H ehas serial input device working from both input and output side. He has written a lot of code using C sharp. There were a number of suggestion from different students about which packages Nick could use to further improve/develop his program.

Joe and Joe talked about their online gambling/gaming platform for competitive players. The games they are considering are deterministic games. It is based purely on strategy and not on luck. It is also a perfect information game (I think).  They are building a siple game of a shooter - The amount ammunition one gets is based on the amount he invests. In future the team wants other people to develop games for their platform. There were a large number of questions and suggestions from the audience. 

Finally Ezra and Tahsin talked about their android application RPI Walk. Tahsin is doing the front end of user interface and Ezra is working on the back end. They are using Open Street Map for their map of buildings and the open source route finder to find routes. Their route finder and the time it takes part of their program is done. They are working on the notificaion of wen to leave the building so as to arrive in a particular building (called notifier) and multiple routes. They are making great progress and they are making weekly blog posts too.

As usual I learnt a lot by attending these talks. Hope many of these projects succeed and make the students and RPI and RCOS proud!

Week 5 Summer 2014 (6/17/2014)

We are close to the midpoint of summer session. Students and the projects are picking up momentum. We had three scheduled talks (One of the four talks changed their mind)

We had presentations by
1)  Dr. Goldschmidt Stress test of Grapefruit -  https://github.com/grahamcracker/grapefruit
2)  Chenrui Cao, Kai Kang, Xi Xi, Juntao Zhuang - 2 W (When and Where)http://rcos.rpi.edu/projects/2w-when-and-where/
3)  Robert Rouhani - Data Oriented design - He has already posted his talk slides here http://slides.com/robertrouhani/data-oriented-design#/

Dr. Goldschmidt gave a brief description of Grapefruit -MOOC software that he spearheaded in building it. It was written in Ruby on Rails. Around 40 students logged on to the system (created accounts) and tried to use the video and other facilities simulatenously. Many students gave useful feedback (Grapefruit video does not seem to appear in mobile devices) - This software s going to used this summer for the bridge calculus program. Already Calculus content is placed here. We will hope that Grapefruit is a great success. Much is riding on its success.

Chenrui, Kai, Xi and Juntao talked about when and where. They are buildin android application to know the building hours, locations, maps, office hours and offic elocations as an app for Android. Eventually they want to merge it with RPI Mobile and RPIphone. Many suggestions are given about the scrapping of the data, have an automatic update. They provide both searchable as well as browsable interface.

Robert talked how to speed graphics application by judiciously placing data together. In just object oriented programming one may not worry about where data reside inside the object.  Lack of a clever organization may incur cache misses and a lot of disk/memory reads. Since processor speeds are much faster memory access speeds Robert outlined how to achieve the speed-ups. He gave some bench mark data which utilizes a judicious data organization. He also talked how such an organization will be helpful in multithreaded applications. His talk slikes may be found here http://slides.com/robertrouhani/data-oriented-design#/

Since we had only three talks, I tried to give an inspirational talk. I talked abot I lucked out in getting a sequence in Online Encyclopedia of Integer Sequences http://oeis.org . I told them about the simple python program I wrote http://arxiv.org/abs/1402.2878 and how Allen Lavoie, a former RCOS student sped up the program using threads. Allen's code is here. https://github.com/allenlavoie/path-counting . Our sequence is here http://oeis.org/search?q=1%2C3%2C12%2C28%2C48&language=english&go=Search
Finally I told them how I satisfied one of the items in my bucket list and advised all the students to keep a bucket list. I advised them to accomplish the items in the bucket list one by one.

Week 4 Summer 2014 (6/10/2014)

Weeks are passing by under a cool/warm summer breeze. Students are working hard. The blogs are getting posted and the codes are getting updated - All of them are good signs. Hope this trend continues.

This week we had four talks.
1)  Jonathan Bergman, Jake Lowey, and Erik Vink-Lainas - 3D HCI http://rcos.rpi.edu/projects/3d-hci/
2)  Alex white - Why functional programming?
3) Jorel Lalicki - Hackodex http://rcos.rpi.edu/projects/hackodex/
4) Daniel Baek and Sunil Kumar - OpenIRT http://rcos.rpi.edu/projects/openirt/  and Sodexo Menu Parser - http://rcos.rpi.edu/projects/sodexo-menu-parser/



Jonathan, Jake and Erik talked about a 3D HCI meant for Lazy people to watch and change channels in TV and veg out - ideal for many/most people in the world! They are using ardunio and handgoves to get sensor input and have the desired effect. They have a very basic program running. They have ardunio and they are using golf gloves for sensors. Looks like a very nice project with lots of potentials for use (including for handicapped people) Robert Rouhani made some helpful suggestions.

Alex talked about Functional Programming - its advantages and feature. Alex did not like much os Apple's  swife http://guardianlv.com/2014/06/apple-releases-new-programming-language-swift/  He gave other pure functional language, how they are thread safe, programs are succinct. Alex based on his talk stemming from his OCaml implementation for Natural languae parsing for his project. There were a number of interesting questions.

Jorel talked about Hackodex which provides project starters and head hunters about the capabilities of various persons and projects using a simple search. What I liked about his presentation was the number of releases (each incorporating some features). He is doing in Python and Django. For the first release, he is planning to use GitHub API. Then the will create a data base for the next release. The third release will have complex queries (those fields are not present in the API )

Daniel and Sunil talked about their joint  project on OpenIRT - Creating libraries for Infra Red sensors This will be on going project into next semester. By the end of this semester, they hope to have some open source libraries for IR sensors. They have the hardware ready.Moorthy suggested looking at this project (RFID) http://rcos.rpi.edu/projects/rensselaer-phrequency-identification/ and Jorel also had suggestions. Daniel is also working on Sodexco Menu Scrapper. he is cleaning up the code to be used in RPIMobile project. Daniel seems to be very enthusiastic and prompt. I hope some good things come out of this project.

As usual I am impressed with  very nice presentations and students' participations. The talks are uploaded in RCOS youtube channel http://www.youtube.com/rcosrpi1 (The last talk seems to be not loaded)   My major complaint is the students have not posted their presentation slides to every one nor posted them in their blog (I have to keep repeating/nagging like a nasty person :( )

Week 3 Summer 2014 (6/3/2014)

I am back from my hiatus. We have a lot of students working this summer (thanks to the efforts of Dr. Goldschmidt).  In terms of Pizza order, it is almost doubled!   We have able lead mentors Jorel and  Robert and a team of mentors to help guide the process. We are in a pressure to get more funding and make this summer a success. Hope things will turn out to be good.

This week we had four presentations.

1) Robert Rouhani  GIT tutorial  http://slides.com/robertrouhani/git
2)   Aaron Gunderson and John Yannou 4xB  http://rcos.rpi.edu/projects/4xb/
3) Jake Weiss and Roland Judd  MilkyWay@home  http://rcos.rpi.edu/projects/milkywayhome/
4) Alexander White Avi Weinstock Hayleigh SCM  related to http://rcos.rpi.edu/projects/mgi/ ( MARKOV-CHAIN Probabilistic Word Classification Based on Variable Order Markov Chains )

Robert gave a nice and detailed instructions on how to use Git. He has posted his slides to the group.

Aaron and John talked 4xB (or a whiteboard as against Blackboard) for a co-operative (crowd sourced) platform for sharing data on homeworks, tests, duedates and efforts needed. They are building using Django and python. They showed a demo of their mockup system.

Jake and Roland talked about their current work on Milkyway@home project. The two tasks they were handling are on workflow assignments and improving the threading library. Since this project has a well established code base, they are onto a good start. Jake and Roland gave the motivation (and the ultimate goal of n body simulation)behind their project.

Alexander, Avi and Hayleigh presented their SCM/MGI project. They are working on "A markov chain based grammatical checker which works by deciding likelihood of words occurring in certain contexts." They have started some preliminary implementation. Hope they post their slides on line (There was technical difficulty with their computer). 

All the talks have been uploaded in RCOS youtube channel https://www.youtube.com/user/rcosrpi1

Please continue the good work and be role models for others to follow!

Had a Blast - Last Blog Post for a while Summer 2013 7-29-2013

When I joined RCOS 6 years ago, little did I know it will have such a profound effect on me. Being associated with such bright, helpful and motivated students was the highlight of my professional career (on n years) at RPI. The diversity, originality and creativity among students have to be seen to be believed.

The hackathon at The center of gravity was a blast. Here was a octocopter demo in Uncle Sam Garage

I could cite achievements and other noteworthy adaptation of our software in various places. If I were to do that, I would be yet another bean counter! The peer mentors, group decisions about etiquette, healthy competition among projects were some of the reasons RCOS made a difference in every one.

We all have to thank Mr. Sean O'Sullivan '85 for his generous donation that made all this possible. I am also thankful to Dr. Raj Raghuram, Dr. TV Raman and Dr. Robert Ingalls who encouraged me when I had doubts. Dr. Janaki Krishnamoorthy and Raju Krishnamoorthy kept me sane (or tried their best) when I was going the deep end. Support by the administrative staff is also highly appreciated.

Hope to return to see a better and improved RCOS in Fall 2014.

Memories! Memories!! Memories!!!

Week 15 (5/7/2013) and (5/10/2013) Spring Semester

Week 15 (5/7/2013) and (5/10/2013) Spring Semester

This is the last week of the semester. We had two meetings this week - on Tuesday and on Friday.

On Tuesday Rob Escriva '10 (and a current graduate student at Cornell University) talked about HyperDex  http://hyperdex.org/ and his Github repsotiry is in  https://github.com/rescrv/HyperDex Hyperdex is a nosql key value store which supports atomic transactions, consistency, atomicity and durability. Rob gave a lucid presentation and topped with a demo. Rob also demonstrated the high performance of HyperDex on sample data sets and queries. Rob's talk was well attended and students asked a number of pertinent questions.

On Friday we had student presentations giving their second rounds of talks.

  

1) Jordan Yamada, Bharath Santhosh - Alfred - http://rcos.rpi.edu/projects/alfred/

2)  Andrew Karnani, Vera Axelrod , Colin Steifel - YAExS - Yet Another Exam Scheduler http://rcos.rpi.edu/projects/yet-another-exam-scheduler-yaexs/

3) Alex Freska, Ian McLoughlin - Flowur http://rcos.rpi.edu/projects/flowur/

4)  Joshua Makinen, Steven Hammar - Resistance Color Scanner - http://rcos.rpi.edu/projects/resistance-color-scanner/

5) Austin McGuire, Kevin Fung, Tianyi Chen, Daniel Fellzardo - Rate My Stream

http://rcos.rpi.edu/projects/rate-my-stream/

Jordan Bharath and Kyle talked about the current status of Alfred and their progress on it. One of the group members will be working during Summer to complete.

Andrew, Vera and Colin talked about their progress and problems in YAExS. One of the difficulties was with the library they used and the other was getting the relevant data.

Alex and Ian talked about Flowur. They had some minor problem with Layout. Alex is planning to work during summer to release a version. Ian will be working with RCOS on ThreeD layout.

Joshua and Steven talked about resistance color scanner. They spent time in learning development tools for Andoid. They also developed a prototype Python program to get the resistance color scanner working. They learnt about openCV (thanks to Bharath) and they plan to develop Android application during summer.

Austin, Kevin Daniel (and) Tianyi talked about RateMyStream project. Since they divided their project they resulted both useful ad hurt their progress. Resolving an authentication bug took some time. They learned a lot in doing this project.

We had a fantastic semester. I might have learned more than what students might have learned.

It was a great semester and a learning expereince.

Week 12 (4/16/2013) Spring 2013

Week 12 (4/16/2013) Spring 2013

With the Grand Marshal week behind, we are in the final stretch of the semester. We had our usual set of fine speakers.

1)  Stephen Silber and Steven Perkins RPIMoble  http://rcos.rpi.edu/projects/rpi-mobile/

2) Christian Johnson RPI Directory http://rcos.rpi.edu/projects/rpi-directory-app/

3) Ethan Bond, Raymond Jacobson Instant Dropbox Syracuse Hackathon

4)  Diogo Moitinho de Almeida, Bharath Santosh, Zach Jablons,  Puckmen in Kaggle Compettion

5) Diogo Moitinho de Almeida, Bharath Santosh ProtoML http://rcos.rpi.edu/projects/protoml/

6) Tim McMullen, Awesome Wav http://rcos.rpi.edu/projects/awesome-wav/

Stephen and Steve talked about their ongoing RPIMobile project. They have been progressing along. With UPE's server, they are planning to have their version in AppStore. Recent improvements include a nice event notifier, and an improved campus map with pins. They showed a demo of their system.

Christian resolved/helped fix the bug with google app-engine API. RPIDirectory is functioning again. Christian has incorporated a cool PhotoMap feature. You can check out here in http://rpidirectory.appspot.com/photomap

Ethan and Raymond talked about their experiences in participating in a hackathon in Syracuse. They won the second prize and a raspberry pi.  Their project involved an implementation of a fast drop box like feature for sharing (one need not download drop box) and just share an url. They implemented the whole system in a day (using ruby on rails)

Diogo, Bharath and Zach talked about their experiences in kaggle http://www.kaggle.com/ competition (machine learning) recently. They have decided the blue book value of bulldozers based the past data. they talked about their different techniques, what caused problems, and about bad data. Despite all this, the team puckman finished 101th among 478 teams  

http://www.kaggle.com/c/bluebook-for-bulldozers/leaderboard  (They slipped from the previous position of 27)

Diogo, Bharath also talked about their RCOS project ProtoML - They used many of the features of PortoML for their Kaggles competition. They showed a demo of their system.

Tim talked about his successful efforts in covering his steganography program Awesome Wav to work using threads. He showed a demo of his system. Tim has to resort to using PThreads for threads. Tim elaborated  his difficulties with the file i/o and how he resolved most of them.

As usual there were many interesting questions and suggestions. It was a great learning expereience for me as usual. I will be missing this next two weeks.