The ElementTree library is an excellent way to allow your Python apps to read and manipulate XML files in a Pythonic manner.  But what do you do if you want to create XML from a Python object with arbitrary dictionaries and lists?

To tackle this task, I have created the recursive xmlObjectCreator.  Its code is noted below.  When you pass it an arbitrary Python object and a few other parameters, it will give you back an ElementTree representing the Python object.  You can then dump or write the tree to your heart's content.

There is one caveat, however.  To avoid unneeded complexity, I did not provide any method by which you can describe metadata.  This basically means that for lists, the actual items will have element names set to the list name plus a static string.  So, you might have myList['One', 'Two'], and it would generate XML like this:

<myList><myList_item>One</myList_item><myList_item>Two</myList_item></myList>

To run the method, you do this:

xmlObject = xmlObjectCreator
    (inputObject, topLevelTag, etreeRootObj, structType, etreeRootObj, True)

The input parameters are as follows:

1. inputObject.  This your arbitrary Python structure.
2. topLevelTag.  This is the tag that you want your root level object to have.
3. etreeRootObj (parameters 3 and 5).  This is a valid ElementTree tree object.  It should not have any elements.
4. structType: an empty string.  This is used internally during recursion.
5. True.  This is always set to True.

The extra string parameter, root object paremeter and the tailing True are needed because the method is recursive, and it will pass different values for these paremeters depending on the depth of recursion and the current object it's working on.  Some folks have suggested alternate, cleaner ways to handle this - and those ways will be implemented in future revisions of the library.

xmlObjectCreator.py

UPDATE:  I've updated the code to use introspection instead of try/catch exception blocks.

xoc.py

Energy prices.  The disjuxtaposition of the mortgage system in the U.S.  The decline of fish stocks near the coasts.

So what's with the future, anyway?  Some things, such as the mortgage mess and resulting economic turmoil, are a result of the need for instant gratification of a large number of our brothers and sisters.  These things are but a blip in the geohistorical radar, much like the Great Depression.  They may be a large disrupting effect on human systems of exchange, but can be recovered from.

However, as citizens of the Earth, we seem to have reached a few geological  "hard limits".  Indeed, we seem to have created some of them ourselves.  The two biggest hard limits I am speaking of are climate change and the end of cheap oil.  Regardless of the root causes of these problems - whether human events are the cause of global warming, or it is a geological cycle; and whether we have reached "peak oil" or still have a few years to go - the symptoms created by these problems present some unique challenges for our species in the coming decades.  Will we run out of oil and reach a point where we don't have enough energy or raw materials to continue even the most basic scientific research?  Or will we 'power through' the next few decades and become an interstellar species?

"Take hold of the future or the future will take hold of you - be futurewise." -- Patrick Dixon - Futurist and author Futurewise 1998/2005

There are a number of interesting resources to start a brainstorm about the future.  Some of these have a basis in reality, and some are intended to be pure science fiction.

http://www.btplc.com/Innovation/News/timeline/TechnologyTimeline.pdf

http://www.btplc.com/innovation/news/timeline/

http://www.trekmania.net/conference/chronology.htm 

I have an article up on Packt Publishing's Article Network here:

Click Here

This article shows how to secure your web-based Python applications with mod_wsgi and SELinux on RHEL5 (or CentOS 5 for that matter).

More blog posts to come shortly!

Cheers,

-J

I had intended to create a blog section on the math aspects of the show, but the folks at Wolfram publishing beat me to it.  Here's the link: http://numb3rs.wolfram.com/

The show used to do a very good job of avoding "T.V. Science", that bad habit of show producers to invent whatever science happens to be handy for solving their plot.  However, this last episode (aired on November 16) delivered a couple of doozies.

For the record, here is what a real DTMF decoder looks like:

It's not the size of a PC's motherboard, with heatsinks and all that.  Decoding the DTMF tones won't tell you anything except what numbers were dialed - it surely won't pull up the location of the phon on Google Maps.  And who solders circuits without cutting the power, anyway?

A while back, my employer sent a few of us to .NET training.  This was optional, as our particular development group doesn't use .NET.  However, just for kicks, I decided to take the class.  This was a mental challenge for me, as I'm well aware of the problems that are caused by the vendor lock-in precipitated by, and supporting the monopoly of, Microsoft.  But I'm open to most things that will make me more valuable to my employer, so I bit the bullet.

We had the pleasure of taking the class from a Microsoft MVP, Steve Smith. Overall, the class was very productive, and Steve was informative.  And, as I learned,  Microsoft's developer tools make developers very productive.  It's very apparent why smaller software development firms use .NET; it keeps their costs lower.

However, there are still problems with the Microsoft way of doing things, which I will describe in a subsequent post.  Where the costs of developing software using Microsoft tools may seem to be lower, the cost of running a Microsoft platform is higher (than running Linux or Mac OS X), in some cases much higher.  So, while I was taking this .NET class, I was wondering in my mind how we could get the same excellent productivity while developing cross-platform apps.

In the past, I had heard about Python and how easy it is to pick up, so I thought I'd check it out.  A number of companies have found that their use of Python enables them to save a lot of time while increasing code quality, even when non-developers are trained to use it:

• "We achieve immediate functioning code so much faster in Python than in any other language that it's staggering," says Friedrich.  United Space Alliance
• "Nexedi is sometimes asked: Why not Java and J2EE? While it would be possible to create a similar system with Java and J2EE, development costs would be much higher."  NexEDI

Python offers drop-in simplicity for most of its functionality.  For example, I do most of my hobby development on CentOS, a clone of RedHat Enterprise Linux.  CentOS comes with a very mature, but relatively old, version of Python.  I wanted to use a newer version with a few nifty addons such as wxWidgets and Zope.  It was very easy to drop python into its own directory, add the external packages, and then tell my scripts to use the new python instead of the system-wide, original version.  It ws also very easy to deploy these new components onto a Windows-based system that had Python installed.

So will Python live up to its promises?  Watch this space to find out. Being able to quickly develop a GUI app that needs no recompiling for Windows or Linux is a definite plus...

Cheers,

-J 

The First (and Last) Run

• Availability. (Is this a mission-critical network that must be available 24/7?)
• Capacity.  (How many nodes on the network?  At what speed?  Are there any special circumstances?)
  
• Contingency Plans.  (What happens if the network goes bump in the night?)
• Human Resource Issues.  (Do users require retraining?  Will they need incentives?)
• Responsibility.  (Who is responsible if the network goes bump in the night?)
• Psychological or Political dimensions.  (While this is generally not a part of the written agreement, it is nevertheless an important part of any implementation.  What rules or responsibilities must we define to ensure that management and the userbase is comfortable with our decision?  What is the value of "go with BigBrandX because then there's somebody to sue if it goes bump" (even though licensing may preclude this), compared to a more financially cost-effective approach?

Freedom 4 You

Corporatism and the Constitution

Capstone Handout