New Online Sexual Harassment Service

We live in an era of many sexual scandals such as #metoo and Harvey Weinstein. Unfortunately, the number of cases has gone up in the past few years, 76,418 in 2018 alone (according to the EEOC). This is an issue that has not disappeared from the workplace and now, more than ever, employers need to better prepare their workforce and prevent more accidents from happening.

There are many tools that organizations can adopt such as fine-tuning the company policy, posting posters with workers rights to the kitchen wall, offer training in-person or online, etc.

While there is no one-size fit-all solution, U.S. organizations need to comply with certain regulations based on their location and size. For example, the state of California requires companies to offer a sexual harassment prevention training to all their employees every two years, or within three months after they join.

Enter EasyLlama. They offer modern online sexual harassment training for companies of all sizes. It’s very easy to get started (self-serve) and get your team trained online.

It works any any device (mobile phone [iPhone and Android], desktop, tablet, etc.) and the prices are starting at $9.95 for larger companies.

The interface is very intuitive and easy-to-use. You just have to upload the employees list (better if they have an email address), assign them a training and they will receive an invitation by email. BOOM, very easy.


Cuboyo, a new marketplace to sell 3D object files is being released soon and I’ve heard it’s going to be amazing.

Allowing designers to sell their STL files and customers to purchase them, it will change the 3D printing industry by allowing exchanges of files for lucky owners of 3D printers…


What Adsense ads have the largest area?

Here is a useful ranking of Adsense ads based on their surface. It can be useful if you want to pick up the Adsense ads that has the largest surface and increase your CPC or CPM.
All units are of course in pixels.

180000px for 300 x 600 – Large Skyscraper
96000px for 160 x 600 – Wide Skyscraper
94080px for 336 x 280 – Large Rectangle
87300px for 970 x 90 – Large Leaderboard
75000px for 300 x 250 – Medium Rectangle
72000px for 120 x 600 – Skyscraper
65520px for 728 x 90 – Leaderboard
62500px for 250 x 250 – Square
40000px for 200 x 200 – Small Square
28800px for 120 x 240 – Vertical Banner
28080px for 468 x 60 – Banner
27000px for 180 x 150 – Small Rectangle
18000px for 200 x 90 – Vertical X-Large
16200px for 180 x 90 – Vertical Large
16000px for 320 x 50 – Mobile Banner
15625px for 125 x 125 – Button
14400px for 160 x 90 – Vertical Medium
14040px for 234 x 60 – Half Banner
10920px for 728 x 15 – Horizontal Large
10800px for 120 x 90 – Vertical Small
7020px for 468 x 15 – Horizontal Medium

And as you can see, the winner is… of course the Large Skyscraper that Google only allows once (and without other ads) in the page.



Cuboyo, a new marketplace for 3D printing

Cuboyo, a new marketplace to sell or buy STL files will open soon.

Named the “eBay of 3D printing files”, aims to target all hobbyists owning a 3D printer (such as the makerbot) at home.
People who are just looking for a specific item that they would like to print can purchase it on Cuboyo for a reasonable price, usually below 10$ and print it at home.
People with more creative abilities can submit their own computer-designed files and share them with the community, whether they would like to make money or not.

What a great idea!


How to use the OV5650 camera board on the PandaBoard ES (OMAP4460)

Because the official repo for OMAP4-V4L2 does not work, I decide to use Gary Thomas instead.

My setting is the following: Ubuntu 12.04 Desktop, PandaBoard ES (OMAP4460).

The first is to get the latest files from the project:
sudo apt-get update
$ git clone -b ti-ubuntu-3.4-1485.7 . project_dir
The dot here can be replaced by the git link so the files are downloaded directly from the server (might be slow).

cd project_dir
project_dir is the name of the directory and can be replaced by anything.

Next, you have to configure the project and compile it. We are choosing AARMHF as it is the version for ARM-based processors with Graphical Card.
do_tools=false fakeroot debian/rules clean
export $(dpkg-architecture -aarmhf)
do_tools=false fakeroot debian/rules editconfigs
In the configuration menu (System Type), choose either SDP4430 or PandaBoard (and camera support of course) because both won’t work together. Also, it is not necessary to change the “armel” configuration file, just choose the second one “armhf” that is prompted during that command.

do_tools=false fakeroot debian/rules binary-arch
This step (the longest) takes up to 4 hours to complete. Sometimes the ABI check or module check was not nice with me so I added “skipabi=true” and “skipmodule=true” to this line and it still worked.

Finally, you have to patch the kernel and install all these modifications:
cd ..
sudo dpkg -i linux-headers-3.4.0-1485_3.4.0-1485.7_armhf.deb
sudo dpkg -i linux-headers-3.4.0-1485-omap4_3.4.0-1485.7_armhf.deb
sudo dpkg -i linux-image-3.4.0-1485-omap4_3.4.0-1485.7_armhf.deb
The numbers might not be exactly the same so just adjust those commands to your situation.

Finally, execute the following:
sudo flash-kernel 3.4.0-1485-omap4

You are done! Simply connect you camera adapter to the PandaBoard ES and reboot. You should be able to see the OV5650 being detected when you type “dmesg” in the Terminal after reboot.

I could make it work using yavta but not gstreamer or vlc. See the OMAP4-V4l2 project for examples.

PS: For cross-compiling on an other computer, use CROSS_COMPILE=arm-linux-gnueabihf- dpkg-buildpackage -b -aarmhf


How to run the OV5650 image sensor on the OMAP4/PandaBoard

We are going to show how to implement the V4L2 drivers on the PandaBoard (version A based on OMAP4430).

1) Download the project:
2) Unpack it on your board
3) Run “sudo apt-get update”
4) Go into the OMAP4-V4L2 directory and run “fakeroot debian/rules clean”
5) Run “fakeroot debian/rules editconfigs”
The resolution of the screen might be too small so you should zoom out in the Terminal window (View menu)
5) Make sure all the options are correct. In the “System type” menu, choose the right image sensor (OV5640 or OV5650). Same in the drivers > multimedia.
6) Run “export $(dpkg-architecture -aarmel)”
7) Run “do_tools=false fakeroot debian/rules binary-arch”. It tooke me 5.5 hours to run on the LeopardBoard A3 (based on OMAP4430).
8) Run the following commands in the upper directory (where those files are located). Change the numbers accordingly:
9) Run “sudo flash-kernel 2.6.35-1209-omap4” to flash the kernel
10) Reboot

After that, the video camera device should be detected! You should see it in the dmesg log.

I used Yavta to captures some image because GStreamer did not work.


How to (cross-)compile Ubuntu for ARM-OMAP4 with V4L2 support

Here is a little tutorial on how I did to compile Ubuntu for my ARM OMAP4 based LeopardBoard with V4L2.

I did those commands on my Ubuntu 12.04 laptop x86 computer.

1) Download from
I chose the Ubuntu Natty TI OMAP4 version:
2) Untar in the directory of your choice
3) Launch the config: ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- CONFIG_DEBUG_SECTION_MISMATCH=y make menuconfig
4) Choose the options you want
5) Launch ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- CONFIG_DEBUG_SECTION_MISMATCH=y make uImage modules
6) Enjoy

I was stumbling upon a problems when compiling, among them:
– I had to disable SPI as it did not compile correctly (and I did not need it)
– cpu_hz was undeclared and I had to modify the arch/arm/clock.c file like explained here:;a=commitdiff;h=44cf7b973e14909ea30a762e0c25381606a9cfdf


How to compile MJPEG Streamer on recent Ubuntu versions (10.04, 10.10, 10.04, 11.10)

I wanted to use that great application called MJPEG streamer and therefore, I

1) Download the source code from the project
2) Uncompress it
3) Change line 59 in the Makefile
4) Change all the include that mention videodev.h to videodev2.h
5) Comment all references to “input_gspcav1” in Makefile because they are using videolan1 and we don’t have it
6) Type “make clean all”
7) You are now ready!


Link exchange

If anybody is interested in exchanging links with me on this blog, don’t hesitate to contact me!


Import AutoCAD (DXF) files into Altium Designer

It’s very simple: simply create a new PCB document and then, go to File > Import and open your file.
Simply follow the instructions and you are there!

Unfortunately, I was not able to get the net names with it. It only brings the primitives…