Ward Cunningham and Wikis

Floss Weekly this week was a stormer! Ward Cunningham invented Wikis back in the early nineties whilst working on SmallTalk workstations at Tektronics (not their video test and measurement division!). The original C2 Wiki is still up and running and discussing agile development and extreme programming techniques. Ward makes the point that if you can't explain some point of technical development to your grandmother/bartender then you've over-complicated it. He is very humble and I enjoy listening to him speak.
In 2001 my chum Richard Drake was developing a commercial application for Wikis using his Clublets software. In fact I used clublets to run engineering at Resolution (that Wiki is still being served - but for how long, and how much is still there?). That particular engineering log has outlasted the company! Rik got me into the whole business and he knows Ward.

So, now we live in a Wikipedia world and everyone knows about Wikis - I'm still a fan and am trying to move this venerable blog over to one. I did mention it a couple of weeks ago and I've made some progress. Although the wiki isn't running on my Linux box yet I am staging it on Apache on that machine but the data lives on my wife's G4!
See the new colour scheme here.

Whole drive encryption and disk performance

I've been interested in volume encryption for a while. TrueCrypt ticks all the boxes. Being a piece of security software it should be open source (you don't want any back-doors after all). One thing peaked my interest on a recent edition of Security Now! - Steve Gibson discovered that booting Windows off a system partition that has the TrueCrypt driver installed gives a system that has a significant improvement in disk performance;

...so I wrote a little batch file using that EndTimer tool and the Windows defrag and Vopt and Windows defrag. I ran those three in sequence. With no encryption, Windows defrag took 8 minutes and 35.765 seconds. Vopt took 4 minutes and 31.046 seconds. And then a final Windows defrag took 1 minute, 54.765 seconds. Okay, so just look at the first number, 8 minutes and 35 seconds. I did it; I did it again. That is, I restored the image, ran the script again, and it was 9 minutes and 1 second. So, you know, about 8 minutes and 45 seconds on average. And the difference are just we're doing a lot of head-seeking. And so where the disk's rotation happens to be is going to affect timing a little bit.
They say on their web page that they've got 100 percent pipelining of some sort. Apparently once upon a time it was too slow, and boy did they fix it.

I intend to start using TrueCrypt - so I'll blog about it when I've got it figured.

The Schubin Report with Mark Schubin

Mark Schubin is one of those well known engineers who has columns in various technology publications and for the last year has been producing a superb podcast. The most recent has a good round-up of test equipment that was shown at NAB this year. The Harris OFI-20B Optical Fibre Identifier seems nothing short of miraculous. It flies in the face of a few of the obervations I made about optical link hacking - I really wish I had a chance to see some of these developments.

MP3 for podcasts

Rupert asked me about hosting a podcast for a chum of his who is involved in the Pandora music project. In fact he made me a Vigilantes of Love station on there - very cool!
Anyhow with all this in mind I thought I'd summerise some of the steps I use for the perfect sounding speech podcast!

• Record straight to hard disk - don't go via analogue tape or any compressed system. Pay attention to levels - you can't recover clipped levels after the event so it's better to record low rather than high - try and aim for peaks at -10dBfs.
  
• Once in Audition or Audacity you need to remove any DC bias and normalise the signal.
  
• Save it back to disk (keeping it uncompressed) but convert it to mono - stereo does chew up twice as much space/bandwidth - and you don't need it.
  
• The Levelator is a superb post-process compressor for speech.
  
• Re-import the Levelat'ed(!) file back into your DAW software and save the file out as an MP3 with the following specs; 48kBits, CBR, mono. Why CBR Phil? Well, this MP3 might be played on one of hundreds of ancient and new devices. My first MP3 player was a Diamond Rio500 which definately didn't support variable bit rate! For mono speech the difference in quality is not great.
  
• ID3 tags - get them right - for both ID3 v.1 and v.2 flavours. I prefer Tag & Rename for getting that stuff done - it supports all the standard as well as custom frames. It's very cool to imbed an image with your podcast as well as notes, URLs etc. That MP3 file will have a life of it's own once it's running out via an RSS feed and you want to have the best chance of folks being able to find you. You'll be suprised how many aggregator sites scrap your file.
  
• FeedBurner is the best way to make compliant XML for the RSS feed with a multi-media enclosure.
  
• Finally, make sure you subscribe to your own feed (in Juice or iTunes etc.) - you should be the first to know if your feed gets broken.

That's it - go forth and podcast!

A tale of two podcasts

I did my degree in the eighties and after an abortive start on a Physics Bsc. I switched to a maths and programming course (ironically the maths was easier!). I came across an old paper I'd written on the two branches of artificial intelligence - commonly refered to as the strong AI model and the weak AI model. I don't know if I'd read it somehwere (this was more than twenty years ago!) but I'd made a rather insightful comment;

....It shows maturity within a discipline when valid sub-sections become vibrant and recognisable to external observers.....

I think that can now be said of podcasting - gone are the days when every podcast you listen to (and The Daily Source Code is probably the worst offender!) was about everything and nothing (and actually mostly about podcasting). Two shows that I've recently discovered that relate to this industry are;

• VFX: The Visual Effects Show - Ron Brinkmann, Alex Lindsay and friends review visual effects of the latest movies while discussing the challenges and technologies of today's visual effects pipeline. In a recent episode they talked about the effects work on Flyboys which caught my ear because I knew it had been shot on the Genesis digital film camera. We'd had some rushes at work and I was amazed at the clarity of the images. They had only bad things to say about the grade of the finished film, which is a shame because digital cinema cameras are really starting to shine. They also talked at length about the green-screen/gimbal rig work and how all of the part-models were static and the tracking shots had to work as if they were shot from one moving plane looking at another moving plane. I'd have loved to see the automated cameras movements swinging around the model to achieve that look.
  All in all a really interesting podcast if you have any involvement with digital effects.


• The Schubin Report - Technologist and engineer Mark Schubin looks at the past month's digital television news and events - He's an old-school engineer who has very much kept on top of current developments. This month's podcast has a great section on why every HD television currently available is a bad buy! Rupert put me onto this one and initially I though he was a bit of a schill for JVC - but if you ignore the adverts he really is quite balanced.

So - my recommendations for new listening.