Yay!
I just did the last changes necessary for my newly factored ontology bridger to actually give results!
The code is clean, good unit test coverage. The last things that needed to be changed were whitespace stripping and trimming issues.
The “walking skeleton” is assembled! All I need to do know is attach a command line interface and increase the number of features and other supporting stuff.
What do I loathe doing and what do I love? This is what I need to keep tracking of to focus on my strengths. I need to write it in a notepad after every time it happens at work or school or in my life.
Things that are work related that are extremely pleasurable:
Things I loathe
Things I might like
Friday the 13th began the time of conversations. REMEMBER THIS CHANGE
The two most important parts of a home are flow and the size of the living spaces — bedrooms and living room.
Reddit clone with smart text tags of articles — spider pages and tag it with the most interesting words based on document/corpus word count. Or entropy — something like that. Use something similar to the “semantic zooming” algorithm to have the most interesting words pop out.
Real Estate listings aggregation — alerts on new real estate listings appearing in neighborhoods
Mechanical Turk based proof reading/critiquing etc.
So, even though summer is coming to a close, I have to write down what I need to do for power kiting.
Reading online it appears that a “trainer” kite is anything that’s considered 3.5 meters square or less. I really enjoy the 4 line kites (kites with break lines and control lines), but want safety. The little wrist attachments that allow for emergency landing are called “kite killers” and cost about 15 bucks.
So, i need to purchase a 3.5 m or smaller four line kite and some kite killers.
Went there last Friday. Excellent atmosphere, nice place.
This is the first stop on my coffee shop odyssey. There isn’t any additional sign on the store, just a large one over the entrance. McDonald’s has a road facing sign, so it’s probably not a legislative requirement not to have road signs. The plaza has a couple Indian grocery stores and a Baja Salsa “food factory”.
Not much permanent traffic, just a few men 30 to 40 buying coffee and leaving. The drive in seems busy. Four people are working, one man and three women. The man is young, in his early twenties, two of the women are in their late twenties with wedding rings. The remaining woman is older.
Nothing especially interesting geographically or culturally here so far. I’ll publish if I find anything.
Coffee coffee everywhere.
According to google maps, there are over 100 coffee shops in Colorado Springs. This blog marks my start of a cultural geography survey of every one, taking note of what things (gyms, salons, bookstores, etc.) are near them, what type of people are there, etc.
I hope to go to at least once per week, maybe more.
Walking Skeleton
The first step in an agile development project is to create a “walking skeleton”. An automatic build/test/deploy cycle to support further development. This skeleton is then built on through the development to create a working piece of software.
My favored python tool chain is: Hudson, Paver, Fabric, VirtualEnv, Distribute/Setuptools, Pip
My favored Java tool chain is: Hudson, maven, tomcat, (others to be determined)
Creating the “Nebulae”
This is a simple, undifferentiated substrate that the project will develop from. Like an astronomical nebula, it will collapse over time into a final product.
This substrate only really consists of assumption of the final environment the software have to live. This means that some of the major components such as hardware architecture, operating system, development language, testing framework, etc. have been chosen.
Literate Programming
Literate programming can be integrated into this process after the first “walking skeleton” is created. Perhaps after the first acceptance test is written.
Pedagogy
One approach to teaching programming could be to start with a “walking skeleton” and have novices work on new functionality in a TDD manner. Having a mature, well documented architecture would act as a “scaffold” in the problem domain, allowing the novices to focus on the details of programming.
I hope that I’ll be able to use GarlicSim for this scaffold by next summer. I don’t know how to set up the build infrastructure yet, maybe using EC2 or something else. I’ll also have to make sure that GarlicSim has a good test infrastructure.
I still have to decide whether to use Python 3. Python 3 would be generally easier to teach to, but has fewer libraries supporting it. py.test and GarlicSim already support are ported, and those are the most important dependencies.
My current idea is to guide novices through building a game of life simulation — probably test first, then having them develop a simulation of their favored domain afterwards.
Visualization may be an issue, since the GarlicSim GUI may not be open source. If it is, I can easily create a simple GUI in pygame or pyglet…
