Pew. Pew.

Going in for a little bit of pew pew! I smell trouble ahead.

Book: Model-Based Development: Applications

I’ve been waiting for the book Model-Based Development: Applications for what seems forever, the publication date having been postponed time and time again. It arrived while I was not looking, but today I finally got my hands on it.

It’s the first volume of a series and is written by H. S. Lahman of comp.object fame. Lahman’s posts in that newsgroup reflect the power of his conviction in object technology – and particularly model-driven architecture – in the way that, in over 10 years of comp.object presence, he uses almost the exact same words in every post to drive his arguments, without so much as a hint of doubt. This makes Lahman’s posts quite tedious and repetitive, requiring several reading to extract his wisdom, and his blog is but a compilation of all his comp.object posts. So, having his ideas published is something that has been a long time coming, and finally we have this book to enjoy.

Photography Assistant for Android phones

While looking for the Android equivalent of iPhone exposure metering applications, I chanced upon Photography Assistant, a free Android application that uses the built-in light sensor to measure the lighting condition. For phones that don’t have a light sensor, it allows manual selection from a list of common EV values. In either case, the shutter speed, aperture value, and ISO setting are adjusted according to the detected exposure. The application is not as accurate as a dedicated light-meter but is nonetheless a decent replacement.

What I like about Photography Assistant is the ability to calibrate it against a light-meter or a camera. It is not very obvious how to do this, so here are some simple instructions.

The first step is to take an exposure measurement of a well-lit area with either a light-meter or the camera. Record the shutter speed, aperture value, and ISO setting.

Then, under the same lighting condition, set Mode and Exposure Value to Manual (M) and Man EV, respectively, and set ISO, Aperture, and Shutter Speed to the values obtained from the previous step.

Finally, press Menu, then select Calibrate. A message will be shown; at this point, the calibration can be confirmed by pressing the Calibrate button.

It is important that 1) the light condition remains the same throughout the process and 2) the light sensor is not obstructed.

Having calibrated Photography Assistant in this way, I have found that it gives readings that are close to what I would obtain from my Sekonic L-308S in incident mode.

Undesirable patterns

I hate software design patterns. Not because I am ignorant of them – to be clear, I am not – but because they discourage creative programming.

Programming is an art and, as such, is inspired instead of being systematic. Whenever a developer applies a design pattern as a first resort, he misses the opportunity of finding a better way to achieve his goal and may even be doing something with negative repercussions.

The use of the Singleton Pattern demonstrates this. At the surface, the pattern seems appropriate for solving the problem of how to restrict a class to only one instance; however, because a singleton behaves like a global, it is inevitable that it will be used as such, which causes issues such as instance coupling that inhibits unit-testing or lack of thread-safety.

That is not to say that all patterns are bad, but presented as “solutions for known problems”, they discourage problem-solving and cause issues that outweigh the benefits they are supposed to bring.

Draft. Revise. Edit. Publish.

So, while waiting for my blog to come back to life and allow me to publish the last few posts, I carry on reading Jeff Goins and am reminded that writing is a process that starts with a draft.