Tuesday, March 30, 2010

Video Hack update

Ok, today I added a little patch to the hack, now it works... most of the time.

This hack needs some work. We need to play with the gstreamer settings a bit. In it's current state, it runs around 85% cpu and the frame rate is at best 1 frame a second. Next task is to implement the system using farsight and telepahty.

Remy also showed us some of the civx system. It is an interesting piece of software, and a great idea!

Monday, March 29, 2010

A day of Research and a Hack

Well today was mostly a day of chatting and research. We were actually getting some decent feedback from the IRC channels which is great! While testing out different gstreammer pipeline strings, we decided to add our hack string to the activity. So now our activity is using a udp sink to send and receive the video. This is not a good solution, but it is an interesting hack for now until we get a chance to do it the correct/preferred day.

Another great note, Episode 3 was finally posted!

Saturday, March 27, 2010

Week 3 Review: Chat system

Fix Webcam on 1.5

  • autovideosink seems to be broke
  • Currently using xvideosink

Modularize the code

  • Split up the code into classes that allows easier editing

Documentation

Even More Hardware

  • 5 Linx Donated 12 of their video phones (10 old and 2 of the new model)
  • Their hardware supposedly runs linux, so some cool hacking might be possible in the near future
  • We are planning on bringing a couple of them to Innerlock Rochester and see what they can do with them

Basic Chat

  • We now have a working chat system
  • Basic join/leave messages, basic chat
  • Need to implement names with the messages, but otherwise things are working well.

Community Help

  • When asking for the community for help, we got a few RTFM's. The manual they were pointing us to had "to be documented" pages in the video section.
  • trose, had a conversation that may be helpful (we will see next week).

Storytelling

  • We recoded episode 3, but as of this moment it is still not online. I'll post a link as soon as I see it

Social Networking

  • Trose also started a facebook fan page for open video chat.

Wednesday, March 24, 2010

Working Chat System

Well today, Trose and I sat down and wrote a prototype front end for our chat system. It still needs some work, but is completely functional. I was helping him edit the existing GUI code I wrote the other day at the same time I was learning some more GIT skills. We are now going to test our code by using a dev_test branch to pass our code on our xo's. Once it is stable, it is copied over to master. We will be copying the commits over to master to keep the history intact. we will need to play wit this a little bit more in the future. If anyone has any git suggestions to follow this kind of development, let me know.

Tomorrow, I hope to get some work on the video system. We plan on heading to InnerlocRoc again, maybe we will be able to get something working.

Tuesday, March 23, 2010

Start of Chat System, Documentation

Well today we wrote up a basic design document and posted it to our git repository.

We also got a basic implementation of our chat system. It still needs some work, but the proof of concept is there in the protocol. The next part will be to get the GUI to display sent messages better then an alert message. Later on, it will be interesting to see if we can get it to work using existing protocols.

Today, 5links donated ten of their old and two of their new model video phones. We were able to power them up and awe over the new hardware to play with.

Week 3: Hacking

Well today we started hacking on code. Our fist task was to get the Webcam working on the 1.5's again. It seems they changed something in gstreamer where the autovideosink was no longer working. Thanks to Fran, we were able to get it working with xvideosink. It is not as nice as the auto sink, but it will do the trick.

Next task was to start splitting up the activity into modules like the network stack and gui. That is where we are right now, getting the networking stack to behave properly with shared activities.

Tomorrow, our task is to get up some documentation for our implementation plan.

I am also probably going to re-flash one of the xo's as it seemed to unload half of its kernel modules and doesn't want to auto load any of the ones needed for the wireless. I figure it will be quicker to do that then to screw around in console to figure it out. (its flash based harddrive takes less then 10 min to reinstall)

Friday, March 19, 2010

Week 2: Hardware Summery

It seems this week was focused in getting our hardware setup and tested for our project. We also did some work on our online infrastructure like setup trac.

New XO's

  • 6 New OLPC Xo's arrived Monday
  • Naming theme: Gunslingers of the West: DocHolliday, BillyTheKid, JesseJames, ...
  • Mac Addresses were sent to ITS and now waiting for them to allow them access to the RIT network. (T24)
  • Smoke Tested Them (T25)
  • Requested Developer Keys for the XO's (T26)

Design

We started some basic design documents at the end of the week. Expect to see some code and documents flying out next week.
  • Preliminary design discussion with our technical mentor Luke Macken
  • Feature List
  • Draft Design
    • Collabora python telepathy stack
    • Dual head design to allow non-sugar and sugar versions to work with same base code
    • Hope that non-sugar stack could be used in sugar version to connect with others using services (like an external jabber server or even popular protocols that major chat systems use) instead of point to point mesh

Story Telling

  • Published OVC Update: Episode 2 (T21)
  • We are still trying to get the story telling box running. This is a box that we will use to stream video to access grid. (T9)

Team Update

We would like to welcome Fran Rogers to the OVC Dev team. He was apart of the Blocku Project but decided to move over to the OVC team after some team restructuring. Blocku appears to have been benched at this time.

Story Telling, Design Docs, Xo Smoke and Dev

Last night, it looks like Episode 2 has been posted.

Open Video Chat Update, Episode 2


Today we accomplished a few things.  We had our second ASL class which was interesting.  I have never been very good at learning new languages, but the professor is really entertaining.  We also started some design docs.  Our feature list can be found in our wiki a basic design chart which can be found on our mailing list.

Aside from that, we also got our xo's smoke tested and requested dev keys.  Still waiting for ITS to get them on the network, but we brought them to InterlockROC and used their network to get them updated and tested.  We also got some video from there, so expect to see that soon!

Wednesday, March 17, 2010

XO Smoke Test


Today we conducted most of the smoke tests for the XO's. All xo's passed the following tests: Camera, Suspend/Resume, File to USB, Activity launch from USB, play ogg in browser, Mesh Network and Chat. We were unable to test the following until they have access to RIT's wireless: Connect to RIT Network, Download/Launch IRC Activity, Connect to School Server. So at the moment we are stalled by ITS. We waiting to figure out an acceptable solution as they are hesitant about adding our machines like they did the last ones. Until our xo's are registered, we can not complete smoke tests or test any code on them that requires a structured network.

We also have a new member joining our team which is great. We hope to have Fran officially added to our team in our wiki/track, and on foss.rit.edu soon.

Tomorrow, we are hoping to start development and design docs for our project.

Well here are some pictures of our new xo's.

Tuesday, March 16, 2010

XOs Mesh Test and Design Discussions

Well today was just wrapping up some lose ends with rit and paperwork for the co-op. After that we pulled some mac addresses off of the new XO's so that we may get them on the RIT network in the next few days. Just need to wait on ITS to get them on the wireless white list.

We also decided to change the name scheme for the new XO's. Instead of the seven dwarfs, we decided to just use names of the famous gunslingers of the west like Jesse James. We were able to get the new xo's to mesh and share a chat instance. The network screen was temperamental, but it eventually allowed us to connect them and share the activity. I feel that it should more reliable then it is. Oh well, it eventually worked after a few minutes of tinkering.

Tomorrow, we will be conducting smoke tests on the new XO's as well.

Some other good news, we had a chance to talk with our technical mentor Luke Macken about how we are planning to go about this project. Basically we will be planning to use the Collabora python telepathy stack. We are trying to avoid using the current implementation of the video chat program because of the binary telepathy and gstreamer libraries included in the xo. We also hope to code it in a way that will allow it to be run from outside the sugar environment as well.

One last fun note, today we had our first ASL (American Sign Language) class.

Monday, March 15, 2010

Start of Week 2

This morning we started off getting our hands on 6 brand new 1.5 XO's. Very nice. Our naming scheme... The 7 dwarfs (we had 1 1.5 before, so now we have 7 of them): Happy, Sleepy, Grumpy, Doc, Bashful, Dopey, and Sneezy. So we will have to get them all registered with RIT so that they may access the network.

Today we have been working mostly with our track page. All of our last week's tasks were added to the system and started added tickets for this week as well.

We also started playing with ohloh.

Saturday, March 13, 2010

Week 1: Bootstrap Summery

This week's task was to Bootstrap this co-op. I think we have been successful in this regard. We started off by settling down in the innovation center at RIT. We basically claimed a random room by force! We figured that since we are doing a fulltime co-op for RIT, we might as well get an office. So now we have one! We also moved out the small table that was in there and commandeered four of the trapezoid shaped tables and placed them as one large trapezoid table. We also grabbed a few of the nicest chairs in the innovation center (as trose mentioned in his blog post) We also managed to grab a computer that will be hopefully used in the future to broadcast over access grid.

Foss.rit.edu

  • Installed Drupal
  • Configured Permissions
  • Ported and added content
  • News Aggregater

IRC

  • #openinnovation @ irc.freenode.net
  • foss_bot chat logger. (View Logs)

Websites

Code Repository

Mailing Lists

Story Telling

Research

Friday, March 12, 2010

Final Touches to Foss@rit.edu and ovc research

Today was some more drupal tweaks on foss.rit.edu.  Some more content, permission tweaking, and some other administrative tasks.  At this point, I think that we are almost set up and ready to go!

After dealing with the website, I was able to conduct some more research into how to write this Open Video Chat activity.  At this point, I think that we may be able to get away with having the gstreamer pipeline deal with most of the hard work of capturing and transmitting the video over the network.  I seem to have found out that the pipeline has the ability to open up a tcp or udp port to stream the feed.  If this works as I currently understand, we may be able to just start the feed and let the sugar network api deal with the tubes.  Once that is working, the task will be to optimize the video feed for clarity as well as the size for transfer over the mesh network.  Today we visited InnerlockROC and got some good feedback on some possible optimizations and and possible paths this project can take.

On a fun note, our first story/interview has been posted to youtube!



Oh and I found a pic of us sitting in on a Lighting Talk at InnerlockROC on one of their blog posts.

Tomorrow we will be wrapping up our first week. We plan on setting up our wiki and track pages on fedorahosted.org/OpenVideoChat as well as hope to get a little work on the prototype activity as well.

Wednesday, March 10, 2010

Foss.rit.edu, OVC on fedorahosted.org

Today we have had some luck getting Drupal installed for foss.rit.edu.  In the end we just needed to disable SELinux long enough to get the system installed and running and then re-enable it.  Once Drupal was up, we spent a good portion of the day configuring the system as well as porting the old html site over.  We also set up a news aggregator to aggregate rss feeds from project blogs, personal blogs, news feeds, and even repository commits.  So that was a huge success.

We got some good news with the Open Video Chat, today we got hosting setup on fedorahosting.org.  This gives us a lot of powerful tools in one place.  We have a wiki, track, and git all in one place: https://fedorahosted.org/OpenVideoChat.  It also gave us a few mailing lists.  Tomorrow, our task will be to work on getting content up on our new wiki.

Screen, SELinux, FOSS@RIT, Drupal and Activity Concept update

Today we experimented with different collaboration techniques and programs that could be useful in the future.  We experimented with a screen capture program called recordmydesktop that could be used to record code development, useful commands, and program demos.  Another program we were playing around with is using the screen command to share a terminal between multiple users.  That was pretty neat, I can see that screen sharing could be helpful if not entertaining.

We also started working on foss.rit.edu website which will be the landing page of RIT's open source and innovation programs .  Our goal is to have a powerful content management system like Drupal running the website.  After a few hours of trying to set it up, I was reminded why I dislike working on servers running SELinux.  We were unable to get the Drupal system to install correctly because of permission problems.  Tomorrow, we will hopefully get the system working as now we know that SELinux was causing this problem.

On the lines of the Video Chat Activity, today I have a basic webcam viewer running in the sugar environment.  Next step will be to get the video displaying on another computer screen.  I can tell already that we would like to get in contact with someone who understands the gstreamer pipeline.  At this point we will probably be able to get it working, but it would be nice to get someone to help optimize the video so that we can pump out the best picture and frame rate while still having a small enough data stream that will not max out the mesh network.

On another note, I hope that the webcams on the new 1.5 xo's works.  The webcam does not work on my current 1.5, that could be a firmware issue that may have already been fixed.  I guess we will find out sometime next week when we get our new hardware.

Monday, March 8, 2010

Open Video Chat Welcome

Well today I have started working on a Video Chat Activity for the XO.  This is through an initiative PEN International.  The goal of this project is to improve the existing (or create a new) Video Chat Activity that will allow it to be used by deaf and hard of hearing to communicate with sign language.  The activity's current frame rate is too choppy and blurry which is problematic.

We have set up some websites to help with this project.  Our source code will be hosted under open-video-chat on Gitorious.  I have copied over the git repo from the old video chat activity and then patched it with the latest packaged code that was four months ahead of the repo.  It appears that when the project was abandoned, someone had updated it to get a working prototype, but did not use the repository.  This annoyed me for two reasons. Some of the changes that were not committed included adding binary libraries.  Another problem is the version in git does not successfully run on the XO and the bundle does.

More info to come.