Wednesday, October 20, 2010 Linux Education Project

Interesting intuitive by Jason Kridner.
You can find mail about " Linux Education Project" & webpage.
Also you can read more details about this meeting in this mail with subject " Linux Education Project meeting minutes 10/19/2010".
Check these following points:

Linux education project meeting minutes - 10/19/2010

* Introductions & roll call
* Project goals
* ECE597 requirements
* UT labs
* Tool usage
* Upstream push
* Web based native applications

Attendees (on phone)
* Jason Kridner (jkridner, jkridner|work) - Texas Instruments
* Gerald Coley - Texas Instruments
* Chase Maupin (chase, _chase_) - Texas Instruments
* Koen Kooi (koen, _koen_) - TI student
* Mark Yoder - Taught ECE597 at Rose-Hulman
* Mike Fontes - MTI Instruments
* Steve Kipisz - Texas Instruments
* Roger Monk - Texas Instruments

Project goals
* How was the bootable USB machine image created?
* Do we need multiple architecture support?  Can we stick to just
32-bit i386 support?
* Why a bootable USB vs. a virtual machine?
 - Concern is the amount of RAM on machines.
 - Also the lack of needing to configure the VM.
 - Quick start-up in classroom situation.
* Why a virtual machine?
 - Performance might not be absolutely critical, so a VM might be OK.
 - USB I/O performance is poor.
* What is our minimal machine target?
 - 2GB
 - 32-bit machine
 - USB 2.0 high-speed with USB boot??
* Can we just baseline on Ubuntu 10.04?
* Would include source for Rowboat (Android) and Angstrom, as well as
including tools for Android (Eclipse).
* A place to purchase a complete kit with the USB drive, cables, SD
card, power supply, USB-to-serial adapter, etc. for the
BeagleBoard-xM, except for monitor, keyboard, and mouse, but
pre-configured for VNC over Ethernet via DHCP in case you don’t have a
monitor and discovery with mDNS (with encryption handled for a Mac and
Ubuntu image that utilizes mDNS to discover the BeagleBoard).  This
could be the “student kit”.
* Need to include a secondary kit for a touchscreen monitor.  Needs a
bit more thought here.  First approximation is to use already
available monitors, keyboards, and mice.  Touchscreen is important for
Android.  This could be a “lab kit”.
* What are the price estimates for the above kits?
* Do we need to support ARM DS-5?
* Labs would be Powerpoint and code on the SD cards.  Mark Yoder
worked with “Embedded Linux Primer” in the past.  Where would we put
the code?  BeagleBoard is explicitly called out in the latest
edition--we’ll likely want to see how accurate we can keep the
information in that book.

Other concerns raised
* Can TI distribute all of this or does it need to come from
elsewhere?  Much of this is not distributable by TI due to legal
reasons (licensing of some of the open source components).

ECE requirements - lab ideas - primarily a technology overview
* Where is first Linux instruction executed?
* Display sub-system
* Basic hardware interfacing for H-bridge/robots
* Video-in via camera interface or USB webcam
* Speech recognition
* Using the DSP (C6Run, C6Accel, etc.)
* GStreamer
* OpenGLES

UT labs
* Key concept introduced is that everything is a file in Linux
* Practice is toggling LED

* End result is a collection of labs that can be added to courses.
* Just using Word is best for Mark Yoder. might be viable,
but we’d want to maintain the Word documents separately.  Powerpoint
isn’t enough--we also need student hand-outs that are a bit more

Ways of getting this done
* First to seek out volunteers
* Have a complete plan of what needs to be delivered and where the gaps are.
* Jason to deliver project plan to his management to gain buy-in and
to assign resources to fill the gaps.

Please spread the word about this project and draw people out for next
week’s meeting!

No comments: