It’s the 21st Century! – Paper

It’s no secret that I’m rather disappointed with humanity’s overall lack of technological progress over the last decade, and some things in particular stand as an embarrassment to people everywhere.  Every once in a while, when one of these makes me particularly frustrated, I’ll post about it, and hopefully someone out there with the right connections will be inspired to do something about it.

Today’s chosen humiliation of humanity is paper.  It is 2009.  There is no reason for medium-large companies or governments to use paper for forms.  This sort of ridiculous, unnecessary waste goes way beyond being more aggravating and more time consuming for all parties.  I would go so far as to say that wasting paper and physically mailing it defiles modern society.

Don’t get me wrong; I still often use paper for jotting down notes and diagrams, since touchscreen technology is still insanely overpriced, but that’s a matter for another time.  What I’m talking about is the kind of duplicity exhibited by the system to apply for Ontario Graduate Scholarships, OGS.  It purports to have an online application system.  After “completing” the “online application”, it tells you that this is, in fact, just one small part of eight (nine if you’ve attended more than one post-secondary institution).   More insultingly, all parts, including the supposed online application, must be printed and signed, mailed to your current university department, processed physically, then mailed to the OGS organizing group, who then mail some parts to other universities for professors to review, who then mail reviews back.  At no stage will OGS accept even a form that has been faxed, let alone an emailed PDF.

I can see no possible reasons for this other than that the people managing OGS are incompetent or they enjoy knowing that they are wasting people’s time, money, and planet.  Why do I think that there’s no excuse?  Many thousands of the students applying for OGS could, in a week or two, single-handedly create a web application that manages the entire process with no paper, no waiting for the postal system to finish wasting so much fuel on your behalf, and no hassle of dealing with physically processing forms.  There’d be none of the pointless duplication of data on the forms either.  How many times do I need to tell you that my name is Neil Dickson and I’ll be studying in the field of Computer Science?

Everybody would win.  Fix it.  Now.  No paper.  No excuses.

iTunes+QuickTime = Malicious Software

I’m sure I’m not the only one who’s realized this, but iTunes and QuickTime are about as close to malware as you can get without being automatically removed by anti-virus software.

The malware-like behaviour started long before iTunes, with QuickTime and its blatant lying that you have the option of not having it start on boot.  I rarely used QuickTime, but it’d be sitting in the background taking up a non-negligible amount of memory, so I’d close it, but whenever I’d restart it’d be there again.  Checking in the Task Manager, sometimes when I’d supposedly closed it, it was still running, but it no longer had a tray icon or any other UI components.  I eventually found the checkbox hidden amongst other options to unclick to have it not start on boot.  However, on the occasions I’d happen to use QuickTime, it’d CHANGE THE SETTING BACK without telling me, putting itself back in the list of programs to run on boot.  This, of course, is exactly what a malicious program would do to maximize the complete destruction of your computer’s usability.  Even Spybot Search & Destroy’s features to disable programs from running on boot didn’t work; QuickTime would add duplicate entries in the startup list to force the system to run it on startup.  That’s about when I uninstalled QuickTime and refused to reinstall it, regardless of what it was that anyone had posted online in .mov format.

I was free of QuickTime for several years, and fully intended never to experience it again.  However, it got reinstalled a couple of years ago, without my knowledge, by another program on my system.  Yes, I am of course talking of iTunes.  The only things I use iTunes for are watching TED Talks, (which I’ll probably now do on the TED website), and listening to internet radio stations, (which can also be done from their particular sites.)  Apple started admitting that iTunes had installed QuickTime, (who knows how long it was there before), when they started calling the program “iTunes+QuickTime” in its updates.  This is around the time when they started the fiasco of listing Safari as an “update” to iTunes.  That’s a sentence that’s so dumb-founding that I even typed it with the rhythm in which William Shatner would say it.

So, I then knew I had QuickTime on my system once again, without my permission.  However, that wasn’t all that iTunes had installed without my permission.  I got an error message one day saying something vaguely along the lines of “Bonjour was unable to access the internet.”  My immediate reaction was “WTF is Bonjour, and why is it on my computer, trying to access the internet?”  Sure enough, Bonjour is a piece of software to find other computers and devices on the local network, so my first thought was “Oh crap, I really do have some sort of malware on my system.”  As it turned out, this piece of software was installed by iTunes without asking, and it has no relevance to how I use iTunes; quite frankly it shouldn’t have any relevance to any feature of iTunes.  It was also running on startup without asking.

However, I chose to overlook this malicious behaviour, as they hadn’t caused any data loss, only negligibly slower responsiveness… and possibly sending my personal info to other computers on the network.  This was not to be the end, though.  I disliked that Apple wanted me to download from scratch and install a new version of iTunes+QuickTime every few days, so I had put it off for a while, and it seems I had good reason to.  The next time I installed an update, completely inexplicably, it one-by-one closed each of my Firefox tabs, then closed Firefox.  I have not known any other program to do this, and I can’t possibly imagine why iTunes+QuickTime would need to do such a stupid thing.  This DID cause me data loss, as a tab or two had unsaved text, and I use Firefox tabs like a todo list, so it had wiped out that text and my todo list, for no logical reason, without ever asking.

Months passed and I eventually forgave again, and today I noticed that iTunes+QuickTime 9.0.0 is the version it’s telling me to update to now.  I thought “Surely they’ve fixed the massive UI design flaws and stopped writing malicious update software by now.”  I was wrong.  The updater still closes Firefox without asking, it still has Safari listed as an update, and updates for Bonjour too.  That’s not to mention the two hours for which it made my computer completely unusable by thrashing the harddrive non-stop.  Mind you, it has improved slightly: instead of closing the Firefox tabs one-by-one, it simply closes Firefox normally, so the tabs were still there when it restarted (though that doesn’t always work), and Safari is now listed as recommended software instead of strictly an update to iTunes+QuickTime.  It also now admits the existence of Bonjour instead of completely hiding the fact that it was secretly probing the network for other computers and devices.

For the record, a few of the massive UI design flaws are still there.  If you click on a video in a podcast and press play, the user expects the video to play.  Instead, iTunes+QuickTime plays the PREVIOUS video I watched, which is one of the dumbest design decisions ever, if anyone even bothered to think about what the play button is supposed to do.  To get a different video to play, you have to double-click it.  At least they fixed that their UI was completely unresponsive to clicks whenever it wasn’t the last window selected.  You had to click it somewhere first, then click the pause button, for example.  Talk about UI design failure.

I’m going to try to eliminate iTunes+QuickTime+Bonjour+whatever from my system before the growing parasite installs anything else to destroy my computer.  As much as I admire the idea of the iTunes store, I suggest that others do the same if they can reasonably do so.

More Tutorial Videos!

Now that my honours project is over, I’ve been busy catching up on other things, including the Assembly Language Video Tutorial and a new Performance Optimization Tutorial.  The assembly tutorial now has two more videos, one showing how to do a 3×3 Gaussian blur in a very simple way, and one showing how to identify red and orange in the image.

Input Image: Map of the Carleton University Tunnels

Output of Episode 4: Map Photo Blurred to Reduce Noise

Output of Episode 5: Red and Orange Thresholds to Identify Tunnels and "You Are Here" Sign

In the next few videos, we’ll be finding which red blob is the “You are here” sign, finding where it’s pointing, finding the farthest emergency telephone symbol from it, and finding the shortest path there.  The premise is that you’re running from zombies and want to warn campus security from as safe a position as possible, but you’re out of minutes on your cell phone.

The Performance Optimization Tutorial is something that I’ve been meaning to get underway for some time, and my co-workers encouraged me to start it after getting an unexpected 10x-20x speedup on the AQUA@Home simulation code.  It’s going to be a much larger tutorial, so it’ll take a while to get into stuff like vectorization, but it’s very important to get a solid grounding in the fundamentals, so after the first two videos, there’ll be 4-6 videos on how common operations on different data types work at the bit level.  There’s a ton of stuff that can be exploited about the nature of these operations, some of which can make the difference between something being vectorizable and not vectorizable.  The hope is that almost any dedicated programmer should be able to get big speedups where it counts.

Edit: There’s a second video of the optimization tutorial out now.  I’m hoping to make this a weekly thing, or at least releasing a video in one of the tutorials each week.

Roots of Two and the Median of the Minimum

Yet another weird math observation.  As k becomes large (like k=1000 has less than 0.034% relative error):

I haven’t actually looked into why this is, but it means that:

So why am I looking at this?  Well, I’m really looking at the finding the median of the minimum of k independent, identically-distributed random varaibles.  With k such variables:

The question is, what value a is the median of the minimum?  Let’s define and solve for it:

If it’s the minimum of many independent variables, it’s then .  So, as long as we can find the value a at percentile p of the original distribution, we’ve found the median of this distribution of the minimum.  Weird stuff.

Inventor IDE Alpha 5b Released, but…

So technically Inventor IDE Alpha 5b is now released on codecortex.com; the problem is that nobody can access codecortex.com right now, even though the bill got paid in time thanks to my father.  Well, I’ll just hope that it gets resolved soon and say a bit about this release.

This release finally lets users create 64-bit applications and operating system images by selecting them in the new project options dialog:

Project Options Dialog Box

There are also many user interface improvements that probably nobody else will notice, but that will come in handy in the next videos of the tutorial I’ll be continuing again soon.  One that people may notice is one that I’d wanted to avoid, but there is now a progress bar for loading.  Personally, I wanted the loading performance to be improved to the point there that wasn’t needed, but that’ll have to wait for the Beta version.

There are some encoding bugs fixed, mostly for the support for 64-bit code, inline functions, and not-yet-released performance test functions.  As an example it turns out that general registers r12 and r13 need some undocumented encodings that I hadn’t supported until today.

The next things potentially on the agenda for Alpha 6 releases are:

• Find/replace bar, a twist on the classic find bar
• Find all references
• Unit/performance testing system
• Exports & relocations, to really support making libraries
• Listing files (a request from Prof. Reichstein for COMP 2003, and something that’d help debugging operating system code)
• Navigation side strip, for quick navigation within a file, especially for errors
• Output formats: OBJ, LIB, ELF, and Mach-O, to support Windows, Linux, and Mac with the same source code
• Input OBJs or LIBs

Things higher up are currently more likely to come sooner, but if people have any requests, please let me know.  The most likely is that only a few will get implemented for Alpha 6, and the rest will be put off for when Inventor IDE has been ported to C/C++.

Possible Downtime and More Random Math

www.neildickson.com and www.codecortex.com may or may not be down on Sunday or Monday.  If they are, I blame Scotiabank for:

• needing 6-8 business days to make a simple transaction that can be made in less than half a second because of two little things I like to call: the internet and the 21st century
• sending two versions of a credit card within a week or two where activating the wrong one makes them both not work
• needing over a week to make and send a new credit card instead of just creating one on the spot in about 30 seconds as they easily could (just need a punch for the numbers and a magnetic strip writer)
• charging me an outrageous $54 annual fee for the privilege of carrying a piece of plastic around that didn’t work anyway
• charing a record high 30% interest rate on the credit card even though I’ve never missed a payment
• paying a record low 0.25% interest rate on my savings account, and even then only on amounts over $5,000; less than 1/3 of what RBC pays, and less than 1/4 of what HSBC pays
• working only 9:30am-4:00pm to avoid dealing with customers who have jobs

Anyway, hopefully the sites don’t go down, and here’s some more random math-ish stuff.  Suppose you have a function in terms of its Taylor series about zero (it’s Maclauren series):

How can you find the square of the function?  In hind sight, it should have been obvious, but it’s:

In fact, more generally, if we also have :

You can also calculate fairly simply:

Simply use forward substitution to solve for the values of .

It’s much harder to work out what non-integer powers of the Taylor series are, though.  The pattern for seems too complex to figure out from 5 terms.  What’d be really cool is to get the series for , since then most anything else would be pretty easy to figure out.

The Harmonic Series and Beyond

I was going to write a post about AQUA@Home, the cool quantum simulation system you can help with that I’ve been working on at D-Wave, but I’ll leave that for a couple of days and talk about one of the miscellaneous things that I’ve thought about at random.

I’ve known for quite some time that (the harmonic series) divergesas n approaches infinity, and more recently, that it scales as .  However, converges, along with any other positive function being summed.  For a while I had assumed that any function would converge, but I learned in my second calculus course at Carleton that also diverges.

Several weeks ago, I was curious where the true bound between converging and diverging is, and I seem to have found it.

I started by finding a simple proof of the scaling of the harmonic series, and it goes something like the following.  Let’s assume that n is a power of two:

and:

Therefore, .  Using a similar approach, one can show that .  Doing this again can show that , and a pattern emerges allowing each one to be expressed in terms of the previous.

The final conclusion:

where the number of logs in the last factor and in the scaling is k.  Note that the is just to prevent any of the logarithms from going to zero, so assume .  If any one of the logarithmic factors in the sum has a power greater than one, the series converges, but with them all equal to or less than one, the series still diverges.  In effect, the scaling can get arbitrarily close to , so this appears to be the very brink of divergence.  Some may argue that having a finite k makes it insufficient, since there are theoretically slower-growing functions, but that’s really splitting hairs.

Biking Directions

As much as my land lady is quite kind, she’s unintentionally given me humourously bad biking directions two out of two times now.  As an example, last night, she suggested I take a shorter route to work today.  The route I normally take to work is:

1. Bike west along Kingsway to Willingdon Avenue
2. Bike north down the huge hill on Willingdon Avenue to Still Creek Drive

It’s about 6.8km total, and it takes around 30 minutes total (at least in this heat wave).  Her suggested route:

1. Bike west along Imperial Street to Gilley Avenue
2. Bike north down Gilley Avenue to Deer Lake Parkway
3. Bike west along Deer Lake Parkway to Willingdon Avenue
4. Bike north down Willingdon Avenue to Still Creek Drive

Unfortunately, the valid set of directions closest to that is:

1. Bike west along Imperial Street to Gilley Avenue
2. Bike north down Gilley Avenue to Deer Lake
3. Swim across Deer Lake to Deer Lake Parkway (actually illegal, but that’s another matter)
4. Bike west along Deer Lake Parkway to Willingdon Avenue
5. Bike north down Willingdon Avenue to Still Creek Drive

What I actually had to do was:

1. Bike west along Imperial Street to Gilley Avenue
2. Bike north down the big hill on Gilley Avenue to Oakland Street, where Gilley Avenue ends.
3. Bike back up the big hill on Oakland until Royal Oak Avenue
4. Bike down the huge hill on Royal Oak Avenue to Deer Lake Parkway
5. Bike west along Deer Lake Parkway to Willingdon Avenue
6. Bike north down Willingdon Avenue to Still Creek Drive

Even if I cut out going back up the hill by going to Royal Oak first, it’s the same distance as my original route, and contains more uphill parts on Deer Lake Parkway because the hill on Royal Oak is even larger than the hill on Willingdon.  (I wish I could find topographic maps for the area to give an idea of the scale of these hills, but all the programs online I’ve tried failed or didn’t have data in sufficient detail.)  I’ll probably just stick to Kingsway and Willingdon, even though they’re very busy roads.

Striking During a Recession

Let me start by saying that I’ve got absolutely nothing against unions as a whole, and I think that they’re necessary for modern society.  That said, I can think of few things that are more cruel and immoral than flaunting your job security and financial stability while so many people are losing their jobs and going bankrupt.

The Canadian unemployment rate is currently 8.6% and the US unemployment rate is 9.5%, but with the way that so many unions are going on strike now, you’d think that nobody was desperate for work.  It’s no wonder that people have been protesting against the striking workers in Toronto’s garbage strike.  It’s so absurd that I can’t even fit all of the reasons in one breath, so here they are as a bulleted list:

• I’d conjecture that most people can’t afford to simply choose not to work for weeks on end.
• Most people need to do their job much better in order to get paid much better, instead of refusing to do their job until they get paid much better.
• Demanding huge pay raises when your employer has much less money than before is not exactly the greatest strategy.  Many people are having to take pay cuts to keep their jobs, or risk having their entire company go under.
• The purpose of a union is to make it such that if one worker is fired for striking, they’d all quit, which normally would be a big bargaining chip against unreasonable companies.  With unemployment this high, many employers, e.g. the city of Toronto, could fire all striking workers and hire new workers who evidently want to work more than the current union does.  In that case, all of the striking workers would be out of work, which is probably not what they want, but considering the contempt they have for the rest of us, more and more people are beginning to think that it’s what they deserve.

I’m all for strikes when we were and when we will be back in economic stability, but striking now is like punching the general public in the gut, laughing at us, and then asking us for our lunch money in exchange for them no longer punching us in the gut.

The Toronto garbage strike is getting heated, but it’s at least not as dire as the Ottawa transit strike this past winter.  Many people lost their jobs because of the transit workers on strike demanding an insane 9% pay raise.  That raise was ripped right from the flesh of those other workers.  The fact that they want a raise in the first place when they do their job so poorly seems ludicrous to me (e.g. it should not be common for a bus to be 30 minutes late when it’s supposed to come every 20 minutes).  The Vancouver transit system is an incredible amount better, partly because of the automated SkyTrain, but the bus drivers also do a much better job, and they haven’t been on strike since 2001!

I realize that I’m fortunate to have a job in this rough economic period, and I think it’s about time that these few unions that want to screw the rest of us for their own benefit start realizing it too.  I recommend, to any employer that can manage it, that any union currently striking without legitimate cause should be fired outright, as long as there are enough unemployed people to fill the positions, as there probably are in Toronto for the sanitation workers.  These unions should learn that they’re not invincible and that they can’t hold everybody hostage whenever they become more greedy.  They make a mockery of the union system and of modern society, and we shouldn’t stand by and let them.

Inventor Performance Viewer

I recently put off the release of Inventor IDE Alpha 5b (probably coming in August, since my honours project is highest priority now) to get something ready for DemoCamp Vancouver that I’ve been wanting for a very long time.  It’s a feature that probably won’t be released until Alpha 6, but you can be certain that I’ll try to make good use of it at D-Wave as soon as I can.  I consider it the first truly awesome feature of Inventor IDE.  It’s easiest to introduce with a picture:

Rock-solid performance data literally in the blink of an eye

(In case you’re wondering, the times are in clock cycles on a 1.66GHz laptop, so 100,000 clocks = 60 microseconds.)

A big problem in doing performance analysis is that getting reliable results and useful scaling info usually means tests that take minutes, so that the impact of the OS or other applications levels off to a constant factor.  If you try to do short tests, you get issues like these:

Bad Performance Data for 2 Mersenne Twister Implementations

Bad Performance Data for 3 Sorting Implementations

These aren’t exactly reliable results; running a short test multiple times produces wildly different results because of just a few bad data.  On Tuesday around 4am (Pacific Daylight  Time), I finally managed to find a decent approach to eliminate almost all bad data without eliminating much good data.  It’s certainly not perfect, but here are the results of running the two Mersenne Twister tests thrice each while AQUA@Home is trying to max out the CPU and the harddrive is thrashing like mad (for unknown reasons):

Much Better Mersenne Twister Performance Data

It’s worth pointing out that these times are in fact noticeably worse than the results when the CPU and harddrive aren’t under heavy load elsewhere (about 1.5 times as long).  The important thing is that the results are still remarkably consistent, and Mersenne128 is still 3.8 times faster than Mersenne32.  You can even still roughly see that Mersenne32 regenerates its data array every 624 elements and Mersenne128 regenerates every 2496 elements.  In case you’re wondering why CPU times would be so consistently longer instead of scattered, it has to do with process switches flushing the cache and TLBs, only causing cache misses in the good data, whereas the bad data still contain any process switches and other interrupts (like the 1kHz tick and page faults).

I’m still shocked that I can effectively triple-click the test run button on the performance test functions and have the results show up as fast as Windows can show the windows.  It’s even more responsive than the regular run button in the ImageProgram sample app, and that had shocked me.  That said, I found the ironic-but-not-so-important problem that it names the output files with the time only down to the second, so one that ran in the same second as another wrote over the first’s data file.

This kind of capability finally allows developers to rapidly iterate through possible performance improvements, not only making it faster to optimize, but much easier, because there’s time to try out many more optimizations in the same amount of time.  For example the faster mergesort in the top graph is from adding a special case for n=2.  In that way, it can also be used as a learning tool for developers to determine what does and doesn’t improve performance.  Best of all, because of the performance viewer’s language-independence, its platform-independence, and the many useful analyses it could do on the data, it’s applicable to many different situations; probably many I won’t think of myself, so I leave that up to you.  I’ve just shown you the first glimpse into what I’ve got planned for it.