James Tauber's Blog 2006/02

blog > 2006 >

Two Years Ago

I just missed my anniversary.

Two years (and two days) ago, I started this blog!

Thanks to all of you who read it regularly.

by : Created on Feb. 27, 2006 : Last modified Feb. 27, 2006 : (permalink)

Three Key Camera Subsystems

When comparing different video cameras, I like to think in terms of three subsystems that each massively impact the quality of the image that results.

Note that there are lots of other features that will distinguish one camera from another, but the three above are the key ones that directly affect the image.

The JVC GY-HD101E I just bought has the following:

I'll say more about what each of these means in subsequent posts. It's a pretty fascinating area of technology.

by : Created on Feb. 27, 2006 : Last modified Feb. 27, 2006 : (permalink)

HD101 Arrives

My JVC GY-HD101E arrived today.

I'll report more as I get the chance to try it out. I also want to start getting down some notes on digital cinematography.

by : Created on Feb. 27, 2006 : Last modified Feb. 27, 2006 : (permalink)

Congratulations Nelson

Nelson Clemente, my singer/songwriter/model/actor friend (who is also the better half of Nelson James) just won Most Promising New Talent 2005 at the Limelight Theatre Awards. Way to go Nelson!

by : Created on Feb. 27, 2006 : Last modified Feb. 27, 2006 : (permalink)

Starting to Blog About My First Feature Film

I've mentioned before that Tom Bennett has started blogging about Heist, the feature film we're working on together.

Well, it's about time I start too.

Tom recently finished the first draft of the script and is currently reworking the script based on my initial feedback.

In that last post, he mentions that I've just bought the camera I plan to shoot Heist on.

It's a JVC GY-HD101E and while Tom will continue to blog about the screenwriting process, I plan to start blogging about digital cinematography, getting up to speed with the camera and planning for Heist.

by : Created on Feb. 21, 2006 : Last modified Feb. 21, 2006 : (permalink)

Thank You Daily Python-URL

It warms the cockles of my heart that, despite no longer being on Planet Python (a state I still hope is temporary), Daily Python-URL still posts (and therefore has editors that read) my Python-related posts.

Thank you PythonWare!

If you read this blog and have any interest in Python, do yourself a favour (to quote Ian "Molly" Meldrum) and add Daily Python-URL to your feed reader.

Of course, that doesn't mean I don't want you sticking around here as well :-)

by : Created on Feb. 16, 2006 : Last modified Feb. 16, 2006 : (permalink)

Are Absolute Imports Available in Python 2.4 or Not?

PEP 328 looks like it will solve my recent import issues.

However, the PEP mentions that

In Python 2.4, you must enable the new absolute import behavior with from __future__ import absolute_import

which got me excited because it means I don't need to wait for Python 2.5 to take advantage of it.

However 

from __future__ import absolute_import

gives 

SyntaxError: future feature absolute_import is not defined

So that part of PEP 328 was just teasing me. I think it's an error in the PEP.

by : Created on Feb. 15, 2006 : Last modified Feb. 15, 2006 : (permalink)

First Pass of Demokritos Persistence Done

@84 on the Demokritos trunk now has full persistence of members and collections into a subversion repository.

At the moment there's no caching (so subversion gets hit every request) and transactions are overly granular. As a result, performance isn't great. But it seems to work!

by : Created on Feb. 15, 2006 : Last modified Feb. 15, 2006 : (permalink)

Bug Fix to Python Unicode Collation Algorithm

See Python Unicode Collation Algorithm for background. This version fixes a major bug that prevented the collation algorithm from working properly with any expansions:

http://jtauber.com/2006/02/13/pyuca.py

UPDATE (2012-06-21): Now see https://github.com/jtauber/pyuca

by : Created on Feb. 12, 2006 : Last modified June 21, 2012 : (permalink)

The Naming of Musical Notes, Part IV

In Part III, I introduced what I called absolute note names and relative note names.

I'd like to introduce a third possibility now which is the approach underlying much of Western music note naming. I'll call it fixed note naming.

The fixed note name of a note is what the relative note name would be if we were in the key of <1>-major.

So with fixed note names, we are still naming based on a diatonic scale (and so will have to use + and - to get the chromatic notes) but the diatonic scale we are using might not be the diatonic scale of the key we are in.

Let's use parentheses to indicate fixed note names. And so a <3>-major scale in each of our note naming systems would look as follows: 

<3> <5> <7>  <8> <10> <12> <2'>
{1} {2} {3}  {4} {5}  {6}  {7}
(2) (3) (4+) (5) (6)  (7)  (1'+)

The third note is (4+) because <7> is not part of the <1>-major scale. Why didn't we use (5-) which also maps to <7>?

Well notice that 

(2) (3) (4+) (5) (6) (7) (1'+)

has the nice property that each number appears exactly once. This property is a convention we adopt. This way, the scale can be thought of in terms of how each note in the <1>-major scale gets modified.

If all this seems too abstract, you can think about the example above in terms of the D-major scale: 

D E F# G A B C#

Note that this Western choice of note naming is what I've called fixed note naming in that the notes of the scale have been named based on modifications to the C-major scale under the convention that each letter appears exactly once.

This is one way of thinking about why the third degree of the D-major scale is F# and not Gb. There are other approaches that come to the same conclusion which we'll explore later on.

I'll finish this part, though, with a final thought. It is possible to have a Gb in a piece in D-major. But it is a flattened 4th and is not the same as F# even though in 12-ET the pitch is the same. In absolute note naming, they are both <7> (if <1> = C). In relative note naming (specific to D-major), Gb would be {4-} whereas F# would be {3}.

by : Created on Feb. 12, 2006 : Last modified Feb. 12, 2006 : (permalink)

Pyworks Common Library and Import Issues

Demokritos contains some modules that might be useful in other software, so I decided to create a little package for these common modules called 'pyworks' (seeing as that's the umbrella I'm starting to use for my various Python-related projects).

So I was thinking I could rename webbase.py to pyworks.web and call my Subversion library pyworks.svn. I also have a library for helping build domain objects from xml parsing events. I was going to call it pyworks.xml

But that means a 'from svn import repos' in pyworks/svn.py won't work and nor will 'from xml.parsers import expat' in pyworks/xml.py

This problem doesn't exist outside of the pyworks package itself. 'import xml' only means the standard library package and 'from pyworks import xml' gives you the pyworks xml module. Similary with 'import svn' versus 'import pyworks.svn'.

For this reason, I don't mind a little hack internal to pyworks.

I seem to recall the py library (that py.test is in) does something similar.

Unfortunately, calling the module 'pyworks_xml.py' and putting 'import pyworks_xml as xml' in 'pyworks/__init__.py' doesn't work.

UPDATE: The following seems to work in pyworks/__init__.py: 

import sys
import pyworks_xml
sys.modules["pyworks.xml"] = pyworks_xml

UPDATE (2006-02-15): Looks like PEP-328 might be exactly what I need. Another reason I'm looking forward to Python 2.5.

by : Created on Feb. 9, 2006 : Last modified Feb. 15, 2006 : (permalink)

January Stats

Looking at my web stats, it appears that last month I had almost 100,000 visits (97,571) from almost 20,000 distinct IPs (19,484) averaging almost 10,000 hits/day (9,801).

That's double January last year. Almost.

by : Created on Feb. 8, 2006 : Last modified Feb. 8, 2006 : (permalink)

Using the Python Subversion Binding

I've previously talked about my desire to use subversion as a persistence layer.

I've now started to work out how to use the Python bindings that Subversion comes with—mostly through studying the source code of ViewVC (formerly ViewCVS)—and it looks like it definitely meets my needs.

I've written up my notes so far as a page for others wanting to use the Python binding. Hopefully it's useful to other people:

http://jtauber.com/python_subversion_binding

It's a work in progress and I'll add more as I discover and use it.

by : Created on Feb. 8, 2006 : Last modified Feb. 8, 2006 : (permalink)

Back Home

No blog posts for a over a week because of a busy work week and travelling back to Australia.

But in that time, the number of bloglines subscribers to this blog increased 5% to 181 (after being flat for months)

Does that mean I'll get better readership if I don't post at all? :-)

But seriously, I'm not sure why the sudden increase. It could be because I'm no longer on Planet Python.

If you're a new reader via bloglines and you don't mind telling me how you came to subscribe, email me.

by : Created on Feb. 7, 2006 : Last modified Feb. 7, 2006 : (permalink)

JAMES TAUBER
Technologist • Entrepreneur • Scholar • Artist
↳ Open Source • Python • Pinax