So 104 people were killed by police in the USA during August, 2014. To my eyes, that's an absolutely enormous figure. As a Brit, I compare it to the 1 person killed over 3 years by the UK police. Yes, they're two different countries, yes there's a lot more people in the US, yes they have different cultures, yadda yadda yadda; people are dying here.

Let's do some maths:

• Population of the USA: 319 million (source: http://tinyurl.com/bpotuf9)
• Percentage chance for a person to be shot in August is then: (104 x 100%) / 319,000,000 = 0.000033%

That's a scarily huge percentage, given that it's normalised by population. Bear in mind that police in the USA are not ... shy ... at shooting at suspects, and neither are they 100% accurate. Some of the casualties are in fact bystanders.

Now let's consider extrapolating for the period of time that most shootings occur (i.e.: suspect between the ages of 15 and 40), and see how that changes things:

• Chance to be shot over 25 year period = (104 x 12 x 25 x 100%) / 319,000,000 = 0.0097%
• Rounding that, since this is an extrapolation, we get 0.01%

Now that's an amazingly large percentage chance of being shot dead by a policeman. Let's do the same thing for the UK:

• Population of the UK: 65 million (source: http://tinyurl.com/kzsalbe)
• Percentage chance for a person to be shot over last 3 years is then: (1 x 100%) / 65,000,000 = 0.0000015%
• Therefore percentage chance for a person to be shot in August 2014 is 0.0000015 / 12 / 3 = 0.0000000427%
• Therefore percentage chance to be shot over 25 year period is 0.0000000427 x 12 x 25 = 0.0000128%

Compare 0.01% and 0.00001% and remember these are normalised by population. Yeah.

In the standard distributed monitoring described in the Nagios docs, check results flow one way - from leaf to root. I needed something a little different - peer to peer distributed monitoring. There are several problems that drive this need. One is checking public services. Our nagios server runs in the same tiny backroom "data center" as the public web server. It can check things internally, but can't check that the general public can actually get to hosts and services that are up internally. Another is that we have larger customers who manage their own network. Our nagios server rightly does not have direct access to all the internal services that need to be checked.

One way to handle both problems is to use nagios remote plugin execution (nrpe) to run the problematic tests on another server. However, our larger customer also wants his own nagios server, with only his network. Having our server monitor the same network via nrpe would be redundant, so I decided to try the peer to peer distributed approach. Each nagios server has a mix of active and passive services. The active services come in transmitting and non-transmitting forms. Both systems are first configured as "regional servers" according to the standard model. Then, I add "obsess_over_host 0" and "obsess_over_service 0" to the root host and service templates on the central server. Specific hosts and service to be sent to the customer's nagios are marked with "obsess_over_host 1" (similarly for service). In nagios, "obsessing" over a host or service means to run a script with each check. For a distributed setup, that script sends the check results to another nagios server (usually via send_nsca).

I added passive-host and passive-service templates:

define host {
                name passive-host
                use linux-server
                active_checks_enabled 0
                notifications_enabled 1
                freshness_threshold 3600
                check_freshness 1
                obsess_over_host 0
                register 0
}
define service {
                name passive-service
                use generic-service
                obsess_over_service 0
                active_checks_enabled 0
                notifications_enabled 1
                check_freshness 1
                freshness_threshold 93600
                check_command check_dummy!3 "No passive update yet"
                register 0
}

and used these for the hosts and services to be check by the other nagios server.

Actually, originally they did not have the "obsess_over_host 0" (and for service) entries, and this led to my passive obsessive checking disorder problem. The symptom was that the log showed passive check results coming in continuously, with checks for the same host or service a second apart, not every 5 minutes as configured. I got frustrated and stayed up late, and finally after sleeping on it realized the problem. A passive check triggers the nagios "obsessive compulsive" behaviour the same as an active check. And this is actually a feature, because you might want to relay the passive checks on to yet another nagios server. I just need to turn off obsessing for the passive hosts and services to prevent a feedback loop between the systems.

Laptop hard drives have long come with power saving features. This makes sense for laptops, which are generally single user systems. I just had the misfortune of installing a pair of "green" WD5000AADS-00M2B0 drives in a server. I soon noticed the problem of rapidly rising Load_cycle_count acknowledged at the WDC Faq.

The fundamental problem with these "green" drives is that they assume a single user system. This was an OK assumption for laptops, but it is rather annoying for a desktop drive. I suppose a desktop can be single user, but I guess we have carefully buy "server" drives instead of "desktop" drives now. Just like you have to buy a "server" desktop to get ECC. While the WDC suggestions to tune logging and setting laptop_mode for linux (which they don't mention) can produce periods of inactivity for a single user long enough to be compatible with "IntelliPower", they are ineffective on a server with multiple virtual machines, or on a SAN server, with many clients, or even on a busy email server.

For laptop drives, power saving could be disabled on linux via "hdparm -B 255". This doesn't work for the new "green" desktop drives. The inactivity timer of this model seems to be set at 8 seconds, so I wrote a simple C program to read a sector from each drive every 8 seconds in O_DIRECT mode (to bypass caching). WDC provides a DOS utility to adjust the inactivity timer - setting a very high value effectively disables it. Unfortunately, these drives were in the field before I noticed the problem.

Because I have a blog at http://fyngyrz.com/.

It kind of makes the whole journal thing redundant. If you really want to see what I have to say about random things, by all means, you're invited to the blog. If not, well, it seems you're in substantial company, if nothing else. :)

OK, you've done it again. When I follow a link from the front page to a story, I get the nice usable old style interface. When I comment on an article, then click on the link to the article at the top of the page, I get the screwed up are-you-mad-who-likes-this-manure beta interface.

Hmmm...

The link from the front page is: http://tech.slashdot.org/article.pl?sid=09/07/16/1335256
The link on the page itself is: http://tech.slashdot.org/story/09/07/16/1335256/Typography-On-the-Web-Gets-Different

Hey, Slashdot, when I'm looking at the page I can see the bleeding title on the page itself, AND in the link, I don't need it in the URL. And why the heck does the URL make a difference?

So, with some trepidation given the media focus surrounding the new 3GS and pointing out how expensive it is over time, I decided to "donate" my iphone 2G to my fiancée and go for a 3GS. Since it's another 2 year contract, I figured I'd go for the top-of-the-range and wait it out again. To my (pleasant) surprise, my needs are relatively cheap...

Initial costs are a bit steep at $415 including tax, shipping. But the monthly charges are $56 (including the data-plan) for my particular needs. I don't use the phone much for talking (450 minutes a month is overkill for me) and I rarely text people (an average of 25/month is (again) overkill, and this corresponds with the '200' dollar amount I'd otherwise have to pay for in bulk). What I *do* use on the phone is the data service. A *lot*. And that's built in as unlimited - it breaks down as $32 for the Nation 450 w/rollover, and $24 for the unlimited data plan.

That comes to a total of $1759 over two years. ($415 + 24 * $56), and I can comfortably afford that. That's also a *lot* less expensive than the $3000+ (over 2 years) that people have been bandying around. It's worth looking at the options, and seeing what suits you before coming to a decision...

Simon

Rewind back to 2000. While everyone was taking a breather after Y2K turned out to be a relative non-event (thanks to hard work from the technical community everywhere), I was coming up with ideas. Ideas for things. Things that would do stuff.

Some of these things caught the attention of my then-employer (a company often associated with the words "big" and "blue"), and the slow wheels started grinding them towards some patents. Two of them in particular made their way through the internal grinder, and became actual applications: "Executing Native Code in Place of Non-Native Code", and "Dynamic Generation of Program Execution Trace Files in a Standard Markup Language".

Then that company gave me the boot.

Over the years since, I've kept an eye on my ideas through online databases. Both were filed in both Canada and the US, with the US applications appearing to be "links" to the Canadian patents. I'd look in on the CIPO database here in Canada every few months, generally to see the only "progress" being that my former employer had paid some yearly renewal fee.

This changed briefly back in 2006, when ""Dynamic Generation of Program Execution Trace Files..." was listed in CIPO's database as "dead". You win some, you lose some.

Ever since, nothing has changed...until I decided on a lark to take a peek today, to find:

• EXECUTING NATIVE CODE IN PLACE OF NON-NATIVE CODE - Issued 2009-04-28

I AM AN INVENTOR!

So I decided to do a quick search of Google's Patent Database to see if it shows up there too, only to find an unexpected entry instead:

• Dynamic generation of program execution trace files in a standard markup ... - Issued 2007-07-31

...so I have been an inventor on a patent since 2007, and didn't know it. The one that was marked as dead in Canada turned out to have been issued in the US. So not only was I surprised today to find out that one of my inventions was just issued a Canadian patent, but that another one was granted a US patent nearly two years ago.

Regardless of what I might think about software patents, this is still a pretty happy day. Both of the ideas patented in these two patents are in use in the wild (and presumably without a license from IBM), and I personally hope it stays that way. I have no say over how my old employer uses these patents (I technically didn't have any say in them applying for these patents either), but it feels pretty good to have these two added feathers in my cap today. It's been a very long wait, and I had long ago given up on anything ever being granted, so this has been a rather pleasant surprise for me.

Yaz.

Now I can't even see the home page, it jumps right into some kind of RSS feed.

Quit the fuck fucking around with Slashdot, let us read the fucking thing without ten kinds of broken fucking AJAX crap.

Somehow Slashdot turned on the Beta index back on for me, and since I had no idea that it had been turned on I assumed this was more leakage from Beta back to the regular Index.

Good god.

All I can say is, they must be subcontracting to Sirius Cybernetics Corporation. It's appalling. It must never be released.

Last month, I finally decided to upgrade my old Celeron 550-based home server to a modern system, and found a pretty sweet Core 2 Duo-based system from Compaq available for just under $300 CDN as a factory refurbished unit. I quickly unpacked it, put Debian Lenny on it (Sarge was a bit too old to recognise some of the hardware, particularly the built-in network adaptor), and copied all the data from the old file server. It's most compute-intensive task is to run pyTivo, and moving from a 550Mhz system with 256MB RAM up to a C2D at 2.xGhz with 2GB RAM was a huge improvement -- we went from about 4fps up to about 60fps (at which point it would pretty much saturate the 100Mbps network connection). It was beautiful to behold.

Last week, after three weeks of flawless operation, I started getting a huge pile of network timeouts, and ultra-slow transfers from the new server. Slower than the old Celeron 550Mhz box it replaced. An order of magnitude slower. So I decided to do some simple diagnostics. In the end, it turned out there were two independent failures in two different pieces of hardware: firstly, the hard drive in the new server was crapping out already. More insidious however, and the actual cause of the slowdown I was trying to diagnose, is that my old 100Mbps switch appears to be having significant problems.

This switch was what most of our wired network devices are plugged into. We have a Tivo, and Playstation 2, the file server, and two Vonage phone adaptors wired into the network, and usually try to have one spare cable for times when the wireless network in our building goes crazy, and we need to plug one of the laptops in. This, of course, is more plugs than the 5-port switch could handle, so we also have an old LinkSys BEFSR41 router, set to switch mode plugged into the switch to offer yet more ports.

The new server is in getting its drive replaced (I wanted an empty drive, but they absolutely insist on putting Vista on it, even though the first thing I'm going to do is reformat/repartition it). As for the networking problem, I decided to leapfrog the issue altogether. The whole reason for having the 5-port switch was because our main routing device is an 802.11g version of the Apple Airport Extreme, which only has one ethernet LAN port on it. So, in an attempt to be a bit more forward-thinking (not to mention allowing me to use 802.11a or 802.11n to hopefully bypass the problem we have with too many wireless devices in other units in our part of the building -- I can usually see at least 12 other SSIDs from our unit), I bought a new Apple Airport Extreme, 802.11n edition, which comes with four built-in gigabit ethernet ports.

Now unfortunately, between our old 802.11g-based Airport Express and my wifes 802.11g-based PowerBook, I probably won't get to take advantage of the 11n speeds all that often. I'm willing to live with that for now. However, the possibility of some really fast transfers on the wired portion of the network by adding a gigabit ethernet adaptor to the server once it's back from being repaired would be fantastic.

My current plan is to continue to use the BEFSR41 as a switch/hub for those devices which are only ever going to be 100Base-TX, such as the PS2 the Tivo, and the Vonage boxes (which will fill it up right there), and other than the BEFSR41 itself, use the Airport Extreme's ports for Gigabit enabled devices (the fileserver, once so equipped, and the spare cable for my MacBook and work MacBook Pro when I bring it home). All of which hinges on finding a good Gigabit adaptor for the fileserver.

Any recommendations on a good Gigabit card for running with Debian Lenny? The board has some free PCI Express x1 slots, and thus a card supporting this would be preferable. Any ideas?

Yaz.

Well...this is it. After three years of work, I'm defending my Masters thesis in Computer Science tomorrow. Entitled "Optimizing Synchronization Cost for Mobile Devices: The Expedient Trickle Sync Algorithm", my research revolved around coming up with a set of heuristic algorithms which could intelligently manage the synchronization of mobile devices in order to reduce the overall cost (where cost involves two opposing factors: the cost of data transfer across the network, and the more conceptual cost of potentially basing decisions based on out-of-date information; thus the trick is to synchronize at times and frequencies to try to guarantee that the users data is up-to-date when they need it, while minimizing the frequency of synchronization, the amount of data transferred, and the use of expensive networks).

The research turned out to be significantly bigger than it should be, and I'm very proud of it, however some of my committee members have been a bit of a PITA. Regardless, I'm going to survive tomorrow, come out the other side, and kick some ass and take some names along the way ;).

Yaz.

This weekend I spent both days riding around the city on my new e-Bike. The sun was out, the weather was fine, and it felt god to be out and one the road and bike trails here in Victoria. Saturday I happened upon a Ska festival down at the inner harbour. Today I was one of the few to join in on an electric bike brigade ride along the coast.

Both days, unfortunately, I wound up having to do some serious pedalling as my batteries started running dry. Today was worse, in part because I had a passenger with me during the brigade ride (the young son of the woman who started the brigade ride really wanted to ride with me, so I let him jump on the back) -- by the time I got home, the battery was pretty much dry, and I had to pedal up a moderate grade on the way home, as the motor didn't have sufficient power to move the bike uphill on its own (and it wasn't giving me much help, either).

Fortunately, this bike has an optional second battery which goes under the seat, however I haven't spent the $90 for this option -- yet. But this weekends rides have got me thinking about power issues to extend the range of my bike.

Purchasing the secondary battery seems like a no-brainer, but like virtually all rechargeable batteries, it has a maximum lifetime, and it's probable that my usual home-to-University (and son home-to-work) trips won't really need it. Charging it as part of the system without ever discharging it will reduce its lifespan (the secondary battery doesn't run in a parallel circuit with the primary battery -- you actually have to switch the key to a different position to use it). I suppose I can simply switch which battery I use on a day-to-day basis, to "level" them equally...

Other thoughts go toward perhaps exploring some sort of portable solar charging system -- in the summer when it's hot and sunny at least, when the bike is sitting parked somewhere for several hours, I should be able to take free advantage of the sun's rays to give the battery some sort of boost. The trick here is that I'd want something ultra-portable (Canadian Tire has some neat foldable panels, but they don't generate much in the way of watts, don't output the correct voltage, and don't have the correct sort of interface) that can simply be plugged into the existing charging socket. Something that can go into the box at the back (or perhaps under the seat, which currently has lots of room as I don't have the secondary battery, although if I do get the secondary battery this space will probably disappear) would be ideal.

Or, perhaps I should just be a bit more sinister and take advantage of "public" power outlets. Today I spent several hours at a BBQ at a city park, near the outdoor stage. The stage has power outlets for use in staging public music shows, and I could have taken my bike up there after the show this afternoon, plugged it in in a corner, and locked it up (setting the alarm) and just leech my 15 worth of power from the City. I didn't do this, naturally, but the thought did cross my mind (especially as the thought of having to pedal home wandered through my brain. While you can pedal it, in reality the pedals aren't positioned to emphasize optimal body mechanics, and the bike is heavy, and it only has one gear, so pedalling it without the electric motor over long distances isn't particularly enjoyable. Some of the people I met at the e-Bike brigade today have actually removed their pedals altogether, as they simply never use them, but I guess they aren't doing 3 - 4 hour tours like I have been this weekend :P).

I need to take the bike in this week for some minor warrantee maintenance (they told me to come back to get the brake cables tightened, and there seems to be a minor switch malfunction where the emergency off switch won't actually disengage the power when switched off every so often...), so I think I'll pick up the secondary battery then. Anyone have any ideas on the use of solar panels to charge a 48V battery?

Yaz.

Gas prices here on the island have finally surpassed $1.50/L, thanks in part to the new BC Carbon Tax coming into effect on July 1st (which, FYI, I fully support). As I'm driving a '97 Chevy Lumina V6 Sedan with a 60L gas tank, if I were to drive everywhere our gas budget would be huge. Never mind the fact that parking is pretty expensive at the University. As such, for the past 6 months or so, Gigi and I have been taking the bus whenever we don't need to transport a sufficient quantity of goods (such as doing a big grocery trip). This works for us right now as we're still registered as grad students at the University, and have a bus pass included in our tuition that is significantly cheaper than the usual pass.

The bus, however, isn't exactly a speedy way to get around from our place. You waste a lot of time walking to the nearest bus stop, then waiting for the bus, riding on the bus (as it stops at nearly every stop to let people on or off), transferring to one or more other buses if we're trying to get anywhere more interesting than the University...etc.

I'm finishing up my thesis this month (defending in late August is everything goes well), and have accepted a full-time development job here on the island starting August 5th at a location only about 6km from our place, and so I decided to take a look at alternate forms of transportation. I settled upon and purchased a GWEV Super 8 electric scooter. And let me tell you -- this thing is just so much fun to drive around the city I keep looking for excuses to get out on it. I've been using it between home and the University for the past week, and its significantly faster than the bus, with less expense (or hassles) that driving the car (and nearly as quick so long as I don't have to get on a highway). As the Province of BC classifies it as a Motor Assisted Cycle, it can be driven anywhere you can ride a bicycle, and can be parked anywhere you can park a bicycle. It only costs about 15 to fully charge the battery from empty, and in our case we're not even paying that -- our building management has given us a special underground, secure parking spot next to a concrete support pillar with an electrical outlet for free -- so they're paying for the electricity.

Some people do look at me a bit oddly now and then -- mostly people who mistake it for a gas powered scooter and think I can't park it on sidewalks and such. It's also less than whisper-quiet -- even under power, about all you can hear is the sound of the rubber meeting the pavement.

When I bought it, as the dealership is downtown, Gigi and I went in the Lumina, but as the bike is too big to put into the car, I had to ride it home. The dealership had it fully charged and ready for me, so Gigi and I left at the same time (me on the bike, her in our car). We had a ~6km trip home each -- she taking the roads, and I taking an old railbed which has been converted into a cycling trail through the city. We got home at exactly the same time.

I can't recommend this gem of a vehicle highly enough. Obviously, it's mostly useful in an urban area with good cycling infrastructure (and legislation which permits you to ride it anywhere a bicycle can). I keep looking for excuses to get out and go for a ride, and now that I don't have to worry about paying for parking or the cost of use, I find myself wanting to go downtown more frequently.

Our only issue now is getting one for Gigi so she can join me [0] :).

Yaz.

---
[0] - Technically, the vehicle does have room on the seat and the necessary foot rests for a second passenger, but apparently only children under 12 can ride as a passenger in this configuration. Still, we have tested it and it will physically work, but we're just not up to testing law enforcement on this one. Besides which, we don't have a second helmet for her at this time.

This summer, I'm expecting to finish up my Masters thesis, and wind up my role as a student in a formal educational environment for many years. Gigi is also finishing off her Masters project, and should be finished with her studies at the same time.

Education has been expensive. I've been fortunate that I've been able to teach some undergraduate courses (I'm teaching a 4th year software engineering course starting May 5th) for income, but with rising food and fuel costs, we barely scrape by. As such, I'm interested in moving straight from school back to the work environment with minimal delay; putting food on the table pretty much requires such. So I've been applying for jobs.

Flashback nine years ago when I finished my undergraduate degree. My initial desire was to do my Masters degree right away. I had just finished releasing the very first feature complete version of the jSyncManager (which is celebrating it's tenth year of development this year, FWIW). However, I also had no less then seven job offers from companies in both Canada and the United States (and took the one offered by IBM Canada, where I worked for nearly 3 years) six months before graduating.

Fast-forward back to today and I've got nothing. I have significantly more experience under my belt (my work at IBM, my time in the Canadian Forces, a consulting gig with the Faculty of Medicine at UBC, three courses worth of University level teaching experience, conferences, publications, four patent applications (with IBM), not to mention nearly a dozen Open Source projects (some obscure, some successful)), but getting my foot in the door anywhere just doesn't seem to be happening.

As such, I'm trying to re-brand myself. My first step is to re-design my personal homepage to be more of an interactive Curriculum Vitae -- a one-stop shop listing my research, publications, patents, teaching experience, work experience, and OSS projects. I'm going to hit my university's Career Services office to see what they can provide.

Has anyone else here gone through the process of self-rebranding? If anyone has and useful insights, please share them below.

Yaz.