Have you ever wondered why some projects develop functional prototypes almost overnight while other projects take forever to produce a working prototype? I think one of the major deciding factors in whether a project team will rapidly assemble a working product or not is if they are aiming for something that is perfect or ‘merely’ good enough.
“The perfect is the enemy of the good.” – Voltaire.
The pursuit of perfection is counter to the pursuit of a working product. To build the perfect product takes time: lots of time, more time than most companies can afford. To produce a good enough product takes significantly less time which increases the chances of actually getting it to market, making a profit and surviving long enough as a business to make a second improved iteration of the product.
“For many events, roughly 80% of the effects come from 20% of the causes.” – Pareto Principle.
According to the Pareto Principle (also known as the 80-20 rule) the first eighty percent of the output (effects) comes from only twenty percent of the work (causes). This eighty percent output is the ‘good enough’ product, therefore to produce the perfect product takes five times as long as the merely good enough.
The question you should ask is ‘Do we need perfection?’. I suspect unless you are programming medical, industrial or military equipment or software the answer is likely to be ‘No’.
Another video by the same guys that presented ‘The myth of the genius programmer’, this time from Google I/O 2008: talking about how to protect your open source project from poisonous or negative people. I think their advice is equally applicable to non-open source projects as it is to open source projects.
Hat tip to Cliff for pointing this video out to me.
I am going to be moving this sites RSS feeds away from FeedBurner this weekend. Mostly because I want per topic feeds and I have not yet found a FeedBurner plug-in for WordPress that supports that.
Consider yourself warned if you are using an RSS reader!
From the Google I/O 2009 conference:
“A pervasive elitism hovers in the background of collaborative software development: everyone secretly wants to be seen as a genius. In this talk, we discuss how to avoid this trap and gracefully exchange personal ego for personal growth and super-charged collaboration. We’ll also examine how software tools affect social behaviors, and how to successfully manage the growth of new ideas.”
Despite being almost an hour long this is a very insightful video that I’d recommend any Software Engineer watches. I find it fascinating that so many programmers want to erase their perceived (or actual) mistakes in source control systems. I guess everyone secretly wants to be the perfect super programmer. However I typically learn more from my failures than my successes: perhaps it is natural to be more introspective about failure than success?
If you find the title or video too pretentious then the question and answers session (around 42:40) is still quite interesting as the presenters get grilled by the audience.
I have been trying to install the excellent MatPlotLib graphing module for the Python programming language on my iMac for a while now. Unlike most python module installations I’ve done the excellent python SetupTools (a.k.a easy_install) has not been up to the task. So I ended up using MacPorts (version 1.8.1) to get MatPlotLib (version 0.99.1.1) installed on Snow Leopard (OS X 10.6.2) with X-Code (3.2.1 – 1613).
So here is a brief description of how to do it:
Install the X-Code developer tools (for GCC, make and other build tools) from your Snow Leopard installation DvD.
Install the X11 Window System from your Snow Leopard installation DvD.
Run ‘Software Update‘ from the apple menu on your Mac to get the latest X-Code updates.
Mount the installer image file (.dmg) and run the contained MacPorts installer (.pkg).
Once installation is complete open a terminal window from Applications->Utilities->Terminal.
Type ‘port‘ at the terminal then press enter to run MacPorts. You should see output like the following if it installed correctly: MacPorts 1.8.1
Entering interactive mode… (“help” for help, “quit” to quit)
Then type ‘quit‘ then press enter to exit port’s interactive mode.
Type ‘sudo port selfupdate‘ and press enter to update MacPorts to the latest version. You will be asked to enter the administrators password before continuing. Depending on how new the version you downloaded is, MacPorts may do some upgrading.
Once the update is finished type ‘sudo port install py26-matplotlib‘ and press enter. This will attempt to install the latest version of matplotlib for Python version 2.6.*. You may be asked to enter the administrators password before continuing. MacPorts will now download, configure, build and stage the dependencies needed to build the latest matplotlib for Python 2.6. This took at least thirty minutes on my iMac and involved lots and lots of scrolling text output from the build process.
Next we need to switch our environment to use version 2.6.* of Python that MacPorts just built and installed with matplotlib. To do this run the following two commands, note you may be asked to enter the administrators password before continuing:
‘sudo port install python_select‘ and hit enter.
‘sudo python_select python26‘ and hit enter.
To test this all worked type the following: ‘python -V’ and hit enter. You should see output like ‘Python 2.6.4‘ which should match the version of python MacPorts built and installed.
Finally to test if matplotlib was installed correctly do the following:
Type ‘python‘ and hit enter to enter the python interactive shell.
Type ‘import matplotlib‘ and hit enter, this will import the matplotlib module. There should be no output if this works.
Type ‘print matplotlib.__version__‘ and hit enter. This will print the version of matplotlib that is installed, you should see output like ‘0.99.1.1‘.
Type ‘exit()‘ to quit the python interactive shell.