programmingstudies

I could only find Trackpad sensitivity in the GUI settings. I followed this and this to set the mouse sensitivity for a cheap wireless mouse I picked up.

I quickly noticed that "Vector" in programming languages ≠ "vector" in mathematics" and that "you should rather learn how to translate math to the code."

Particularly, NumPy's notion of vectors and and their transposition differs from that in linear algebra.

I'll be referring to this post (partially quoted below) in the future.

For example, transposing a NumPy array of shape (a,b,c,d) returns an array of shape (d,c,b,a) -- the axes are reversed. In two dimensions, this means an array of shape (a,b) (i.e. a rows, b columns) becomes an array of shape (b,a) (i.e, b rows, a columns). So NumPy's notion of transposition matches up nicely with the linear algebra notion for 2-dimensional arrays.

I've been studying machine learning. On my iPhone 5s. Watching various lectures on YouTube, including those from Andrew Ng's ML Course, and reading from Steven Marsland's Machine Learning - An Algorithmic Perspective text. Filling in gaps in my linear algebra and calculus knowledge as necessary. I've been here for just over a week. Today I asked the hostel owner if I could install Python on the computer available for guest use. He said "don't install 2.7."

I installed Anaconda v5 and am up and running with Jupyter notebooks and Python 3.6! I went briefly through a NumPy tutorial to get the feel of typing back in my fingers. It felt good to be writing code again, however basic!

I'm excited to implement some of the basic ML algorithms I've been learning about, including multivariate linear regression, logistic regression, gradient descent, and the multi-layer perceptron. This will bring me up to the state of the art as of the early 1970s.

In generalizing a solution to the problem of “given a free throw percentage F, find the probability P of making exactly K out of N free throw attempts,” I am reminded that programming follows after knowing what to program — in this case after having figured out the mathematical model. With programming it is possible to run simulations to quickly gain intuition by exploring the possible outcomes for different values and/or visualizing these data.

P(K out of N free throws) = F^K * (1 - F)^(N - K) * N!/(N-K)!K!

where the first and second terms are the probabilities of making and missing the requisite number of free throws and the third is the number of such unordered combinations of makes/misses.

To remove all byte order marks in a directory I used this solution: http://www.yiiframework.com/wiki/570/remove-byte-order-mark-bom-from-files-recursively/

Just spent a bunch of time trying to find a terminal for ubuntu where I could select to copy and click to paste. I found that Terminator allegedly offered this, so I got it, but then in my installed version it didn't work. What gives?

It turns out my OS had been setup to not register the middle mouse button "as the middle mouse button" -- and this is the button both my existing terminal (ROXTerm) and Terminator both use to paste.

I determined this by trying to get Terminator to copy + paste as desired and bumping into this post about a hack that referenced button numbers. Then I searched for "ubuntu mouse test" and discovered the utility program xev that shows that button numbers for the mouse buttons as you click them. My left button was 1, right was 3, and middle was... 10? How'd that get there? According to this it is possible to remap the physical button output using the somewhat cryptic xmodmap.

Running xmodmap -pp shows the current mapping of Physical Button Numbers to Button Codes. Running xmodmap -e "pointer = 1 2 3 4 5 6 7 8 9 10" remapped all of the physical buttons up to 10 (for me the only ones that were mismapped were 2 and 10) to their respective Button Codes.

Bonus

Jenkinsfile -> Makefile -> Dockerfile -> package.json -> app_start.js

http://curtclifton.net/papers/MoseleyMarks06a.pdf

We can learn about life generally from such a refactoring discipline.

http://ronjeffries.com/xprog/articles/refactoring-not-on-the-backlog/

Sometimes people call this "technical debt". It really isn't: it's really just not very good code. But it doesn't look too bad -- yet.

We are tempted to ask for time from our product owner to refactor. Often, that time is not granted: we're asking for time to fix what we screwed up in the past. Not likely anyone is going to cut us any slack on that.

A modular program consists of components that can be understood and reused independently of the whole, such that the meaning of the whole depends only on the meaning of the components and the rules governing their composition; that is, they are composable.

I was just working through this Clojure tutorial and came upon the "Watch Out!" section where the authors point out that the plain old if statement in Java/C# doesn't evaluate to anything -- and hence, it must have a side-effect in order to do something! Whereas in Clojure, if is an expression that evaluates to a value -- the functional if functions identically to the ternary operator in Ruby/JavaScript/C# and other languages.

I also found interesting that declarative vs imperative is orthogonal to functional.

I've tried getting up an running with machine learning so many freaking times now I think I've completely screwed up my home ubuntu machine, as far as like where all the different python stuff is installed. Well today I made a giant leap in organizativity!

Use virtualenv! It's freaking easy!

sudo pip install virtualenv

Create an env in a dir

virtualenv ENV_NAME

Load the env

source ENV_NAME/bin/activate

Check it

which python which python3

Install the things for the tensorflow tutorial

pip3 install --upgrade tensorflow pip3 install --upgrade matplotlib

Run the tutorial!

Deactivate the virtualenv

deactivate

Get this app: https://github.com/mkropat/jumpapp Use Keyboard to create a custom shortcut to launch a terminal. For example I use roxterm so I created

Name: launch terminal Command: jumpman roxterm