Practical Software Project Management

I'm often asked by clients, partners, and friends about the basics of software development and software project management.

There's much to write about practical software project management. I have on archive a compilation of the very interesting discussions shared with respected colleagues about the subject, and I intend to compile that into a series of articles, or a book, soon.

For now, I'll just provide the shortest possible reading list about the economics and logistics of software development.

• Extreme Programming Explained: Embrace Change, by Kent Beck is the nominal book about the modern economics of software development.
• Waltzing With Bears: Managing Risk on Software Projects, by Tom DeMarco and Timothy Lister, explains that software development is an endeavour characterized by uncertainty, and provides techniques to accept and manage uncertainty in pro of successful software projects.
• Writing Effective Use Cases, by Alistair Cockburn, is about capturing user requirements in the simplest yet most effective way.

Of course, there is much more to read and much more to say about effective software project management. But if you think you like the time, start with those six hundred pages of easy to read and well illustrated text.

Tequeños en Wok

La necesidad es la madre de todas las ideas

Muerto de hambre cerca de la media noche, y estimulado por los restos de la salsa de las albóndigas del mediodía, decidí prepararme los tequeños que no se hicieron más temprano por falta de serotonina.

Apresurado y flojo como soy, decidí que usar el caldero con el montón de aceite no era lo más conveniente para alguien queriendo comer algo en medio de la noche sin despegarse mucho tiempo de la compu portatil (alias laptop), así que decidí hacer los tequeños en el Wok, a los trancazos, y ¡Ho, sorpresa!, quedaron ¡buenísimos!, con una consistencia entre salteado/frito y horneado, al buen estilo chino.

Va la receta…

1. Con una servilleta, riega una capa fina de aceite de cocina en el wok, pon el wok a fuego máximo..
2. Mientras se calienta el wok, saca los tequeños del congelador y sepáralos para que queden bien sueltos.
3. Cuando el wok comience a hacer humito, baja el fuego a la mitad, y agrega los tequeños, removiéndolos vigorósamente.
4. Cuando los tequeños comiencen a dorar, baja el fuego a un cuarto y ponle la tapa a el wok.
5. Destapa, revuelve, y tapa, cada 30 segundos, evitando que los tequeños se doren demasiado, y procurando que se horneen completos con el wok tapado. Baja el fuego al mínimo si es preciso, y aumenta el tiempo de tapado al ojo.

El resultado de esta rápida cocción, hecha con paciencia, será unos tequeños crujientes y gorditos, con el queso bien derretido por dentro, pero sin el grasero ni las roturas y los salpiques típicos de los tequeños fritos.

¡Buen provecho!

Scanning 35mm film

About why to design before implementing

Yesterday I gave a shot at trying to scan some of the 35mm negatives of my old Black&White photos using a 1200 DPI flat-bed scanner. It failed misserably:

• The scan showed severe Moiré and other kinds of noise in the pictures.
• The sizes of the TIFF files were enormous.
• At 1200 DPI the flat-bed scanner picked up dust and stains from the glass.

Although it was just an experiment, the terrible results led me to do some research. Unless otherwise stated, the following numbers come from the article about Digital Photography at Wikipedia.

Resoluiton of 35mm film

The Wikipedia article says that a DSLR resoluiton of 6 to 14 megapixels aproximates that of 35mm film, and that one of 16 megapixels exceeds it, so lets asume that the resolution of 35mm film is 14 megapixels.

To compare with the scanner resolution we need to translate that to DPI. Megapixels are calculated as the product of the vertical and horizontal resolutions:

M=x*y

The Aspect Ratio of 35mm and DSLRs is known to be (3:2), or 1.5. Thus:

x=1.5*y
y=x/1.5

and

M=x*(x/1.5)
M=x^2/1.5
x=sqrt(1.5*M)

So, at 14 (actually 14.7) megapixels, the number of horizontal pixels in 35mm film is about 4693. If we divide that by 1.38" (35mm translated to inches), we obtain 3405 DPI for 35mm film.

Required resolution

With that result it is not surprising why scanning 35mm with a 1200 DPI scanner produces terrible results: 65% of the information in the negative is not captured. Even more, according to Information Theory, the sampling density must be the double of the information density, which means that the scanner should have a resolution of at least 6810 DPI to sample a 35mm negative without bias.

Guess what? The resoluiton of specialized 35mm film scanners is of above 7000 DPI…