Roberto De Almeida

Today’s startups are tomorrow’s enterprises. Many of the other startup folks I know share the same expectations about how software should be sold. Basically, if a given software package or service isn’t free/open, it should be as easy as humanly possible to try it, pay for it, and start using it in production. If it isn’t easy to get started with your product, I’m going to find another vendor.

The article has a nice list of don'ts when selling software. Really nice checklist for when I start my startup. :)

This is an experiment to show how some realtime audio processing can be done with the Arduino. The first set of examples alter an incoming audio signal and put it back to an audio output. We achieve effects like Reverb, Phasor, Flanger or Ringmodulator. The second set of examples are outputting computed waveforms like Sinewave, Bell and Xylophone sounds.

This is on my list of things to do. I've been working more and more with electronics lately. It's something that I learned from my father when I was a kid, and I feel closer to him working on circuits. A few days before he died I remember calling him and asking a question about potentiometers, and talking about the PCB I had just made.

Any advert in a public space that gives you no choice whether you see it or not is yours. It’s yours to take, re-arrange and re-use. You can do whatever you like with it. Asking for permission is like asking to keep a rock someone just threw at your head.

You owe the companies nothing. Less than nothing, you especially don’t owe them any courtesy. They owe you. They have re-arranged the world to put themselves in front of you. They never asked for your permission, don’t even start asking for theirs.

- Banksy

A consistent reduction in cloud height would allow Earth to cool to space more efficiently, reducing the surface temperature of the planet and potentially slowing the effects of global warming. This may represent a "negative feedback" mechanism -- a change caused by global warming that works to counteract it. "We don't know exactly what causes the cloud heights to lower," says Davies. "But it must be due to a change in the circulation patterns that give rise to cloud formation at high altitude.

Cool that this could work as a negative feedback stabilizing climate. Gaia Theory, anyone?

One thing I've noticed about electronics is the lack of simple, newbie-friendly information on the internet about specific topics. Then again, maybe I just wasn't looking properly. To help fill this possibly existent information gap, I'm going to post a nice, simple schematic and an explanation on what it is and how it works.

There are both linear and non-linear forms of the Kalman filter, with the non-linear forms being the Extended Kalman Filter (EKF), the invariant Extended Kalman Filter, and the Unscented Kalman Filter (UKF). The non-linear versions of this algorithm represent the current "gold standard" in many application areas, including GPS and navigation.

Before jumping in the deep end of the pool, I decided to implement a simple example that shows the ideas and implementation of Kalman filtering, using a recursive Bayesian approach.

I need to understand this better and see if this could be used in climate simulations.

A couple of weekends ago, after two years of this test-page nonsense, I thought: surely this must be fixable? Poor dad. I must try harder to fix it. So, up I went, turned on the printer, and it printed the test page. Which I then, for the first time in two years, actually looked at.

Turns out it's not a test page at all. When you put new ink cartridges in the printer, it prints one of these pages and then you're supposed to scan that same page; the printer uses it to re-align the printer heads, because it can see from the printout whether the lines are straight.

Nice post by Sil.

A convolution is a way to combine two sequences, x and w, to get a third sequence, y, that is a filtered version of x. The convolution of the sample x_t is computed as follows:

It is the mean of the weighted summation over a window of length k and w_t are the weights. Usually, the sequence w is generated using a window function. Numpy has a number of window functions already implemented: bartlett, blackman, hamming, hanning and kaiser.