|
|
Monday, October 5th, 2009
| |
11:45 pm - VIFF coincidence
|
|
| Thursday, October 1st, 2009
| |
9:37 am - Vancouver International Film Festival
|
Nice rainy Vancouver morning today for the beginning of VIFF.
In alphabetical order, the films I am most eager to see:
|
|
(comment on this)
|
| Wednesday, September 16th, 2009
| |
1:10 am - TIFF
|
|
| Thursday, September 10th, 2009
| |
2:35 pm - TIFF programme guide
|
My biggest laugh reading through the TIFF official programme guide this year was the opening of the description of Last Ride:Every few years a film comes along wherein every element works, leading to a completely realized piece of cinema – Last Ride is one of those films. Okay, so every other film at the festival this year has some element that does not work, and is not a “completely realized piece of cinema”? And presumably no film at the festival last year or the year before was “completely realized” either, since it’s just every few years that such a film comes along.
|
|
(comment on this)
|
| Wednesday, September 9th, 2009
| |
11:01 pm - TIFF
|
|
| Thursday, August 6th, 2009
| |
4:51 pm - Pour elle
|
Currently playing at the Ridge, Pour elle is the best thriller I’ve seen in ages. It begins with a young couple eating breakfast with their five-year-old son and getting ready for work, when the police bust in, arrest the woman for murder, and search the place while the five-year-old cries. You could find out more about the plot, but it’s better to go into the movie not knowing. Vincent Lindon and Diane Kruger are fantastic, and the pacing is perfect.
Some other movies I saw recently:
La Fille de Monaco: Very disappointing. Well, technically I shouldn’t even call it disappointing because I thought from the trailer that it would be bad, but then I read the description of the movie that there was a mystery in it, and “things are not as they seem” or some such thing, but that turned out to be a lie, things were exactly as the seemed, and the film was a bit preposterous and thoroughly dull.
500 Days of Summer: Good film, definite props for effective use of a drastically nonlinear chronology. The problem was that during the film I kept thinking of Eternal Sunshine of the Spotless Mind, which of course also told the story of a failed relationship by use of a nonlinear chronology, and of course was much better than 500 Days of Summer.
Parlez-moi de la pluie (Let It Rain): Reasonably interesting French film about two guys incompetently making a documentary about a politician.
Tokyo Sonata: The best Japanese film I’ve seen in a few years, although the plot is not entirely original; there was a French movie a few years ago with the same premise, L’Emploi du temps or Time Out.
Sin Nombre: Set in Central America, in the world of crime, gangs, and illegal migration, the film has an interesting narrative structure but was not entirely successful; I found myself confused until half-way through.
Crossing Over: For some reason I hadn’t heard of this apparently big-budget movie that stars Harrison Ford, but it turned out to be surprisingly good.
The Pool: Set in Goa, light but clever film about a teenage boy obsessed with a swimming pool at the neighbouring property that nobody ever swims in.
Un baiser s'il vous plaît (Shall We Kiss?): Engaging French film, structured as a story within a story, and sometimes a story within a story within a story, about kissing, intimacy, and infidelity.
|
|
(comment on this)
|
| Saturday, July 18th, 2009
| |
4:27 am - Privacy Commissioner’s report on Facebook
|
I am impressed by the detailed report about Facebook’s policies under PIPEDA, issued Thursday by Assistant Privacy Commissioner Elizabeth Denham. Arising out of a complaint filed for CIPPIC, Denham’s report cogently addresses each issue, arriving at conclusions that (for the most part) accord with both the letter and spirit of PIPEDA. This is a welcome change from the nature of reports issued from the OPCC during the first few years of PIPEDA.
Denham’s report is online here. (Although labelled a “case summary”, this appears to be the full text of the report.) The OPCC’s press release is here, which also links to a backgrounder and some prepared comments by both Denham and Commissioner Jennifer Stoddart.
CIPPIC made 24 allegations against Facebook of noncompliance with PIPEDA; the report divides these into 12 subjects. The conclusions divide neatly into three parts: in 4 subjects the complaint was “not well-founded” (i.e. Facebook was found to be in compliance with PIPEDA), in 4 subjects the complaint was “well-founded and resolved” (i.e. Denham concluded that Facebook had contravened PIPEDA, but Facebook has agreed to change its policies to bring it into compliance), and in 4 subjects the complaint was “well-founded” (i.e. Denham concluded Facebook had contravened PIPEDA but Facebook is unwilling to alter these policies).
Although 385 paragraphs long, Denham’s report is quite readable. I would advise anyone with a Facebook account to read the report or at least the point-form summary (Appendix A).
I have a few miscellaneous comments on the report.
On the subject of third-party applications, Denham has identified a serious breach of PIPEDA which Facebook is unwilling to resolve (and would find it very difficult to resolve). Briefly, a third-party app is a piece of software developed by a person other than Facebook that runs within the Facebook platform. Examples are games like Scramble (which I spend lots of my time playing, incidentally), or daily horoscopes, or matchmaking/dating services. To use a third-party app, you have to expressly enable it; this will let the app interact with your Facebook experience. For each user who has enabled the app, the developer of the app receives a “key” that gives the developer access to the user’s profile information (other than contact information). The key will also give the developer access to information of the users’ friends which is visible to the user—even when these friends have not themselves enabled the app.
For example, suppose you enable a horoscope app. The app’s developer will get access to your DOB, which is reasonable for the purpose of providing a horoscope service. However, any other kind of app, such as a game, would also be able to get, with its key, the same access to your DOB. Facebook argued that every developer agrees (contractually) not to collect any information except what is necessary for the particular app. Moreover, the developer is supposed to destroy its copy of the information within 24 hours. Denham found, however, that these contractual terms were not good enough. PIPEDA requires safeguards to prevent improper access to personal information. As Denham put it at paragraph 200:When I speak of limits to access, and especially when I consider the vast amounts of Facebook users’ personal information potentially available to large numbers of application developers, I believe something much more substantial in the way of safeguards is required. Specifically, I mean technological safeguards that will not simply forbid, but effectively prevent, developers’ unauthorized access to personal information that they do not need. And at paragraph 201:Since developers can in effect copy users’ personal information from the Facebook site to their own servers, there would appear to be no way for Facebook to effectively monitor the developers’ subsequent use and disposal of the information.... Moreover ... Facebook in effect puts the onus on the users to detect and report problems. The inability to monitor developers’ usage after the fact is all the more reason for Facebook to take effective preventative measures. It will be interesting to see if in future, the OPCC applies the same reasoning to issues involving credit reporting agencies like Equifax Canada. In the past, when a credit reporting agency has released information about an individual to someone who did not have proper consent to obtain it, the OPCC has sided with the credit reporting agency, finding that it was not their fault since they had been relying on their contractual terms (e.g. PIPEDA Case Summary #2003-182).
In my opinion, Denham’s reasoning on this point in the Facebook case makes much more sense; I hope that in future, the OPCC will hold credit reporting agencies to the same standard.
The most insidious aspect of the third-party apps privacy issue is that personal information about a user can be released to a developer even if the user has not enabled the app, but a friend of the user has. According to Facebook’s representations to the OPCC, “the application provider is effectively authorized to stand in the shoes of the individual user on behalf of whom the data is requested” (paragraph 157). For example, suppose you have a Facebook account and in your privacy settings you specify that certain information will be visible to your friends, but not to the general public. This information might include DOB, relationship status, sexual orientation, religious views, et cetera. Then if one of your friends enables a third-party app, the developer of the app (who might be anyone, maybe a 17-year-old halfway around the world) will have access to all that information by “stand[ing] in the shoes” of your friend.
In my view this is the most serious privacy breach identified in Denham’s report, and it is disturbing that Facebook is not inclined to remedy this breach by altering its policies.
Another topic covered in Denham’s report is the advertising on Facebook. She notes that in many earlier PIPEDA cases involving advertising, it was considered to be a secondary purpose, and thus in those cases, consumers had a right to opt out of their personal information being used for ads (pursuant to clause 4.3.3 of PIPEDA’s Schedule). Denham continues (at paragraph 131): Facebook has a different business model from organizations we have looked at to date. The site is free to users but not to Facebook, which needs the revenues from advertising in order to provide the service. From that perspective, advertising is essential to the provision of the service, and persons who wish to use the service must be willing to receive a certain amount of advertising. In my opinion that is an apt and workable interpretation of clause 4.3.3.
The report goes on to discuss the way Facebook uses targeted advertising. As a user, the personal information you have entered in your profile is used in determining what ads you see. For instance, if you have listed a certain television program as one of your favourites, you might be shown ads for DVDs of that program. In particular, ads are targeted to you based on your age (which Facebook calculates from your DOB, which you are required to supply when you sign up). This is true even if you have the maximum privacy settings, not showing your age on your profile at all.
As explained in the report (at paragraph 115):Facebook explained that advertisers who purchase Facebook Ads specify the characteristics of the users to whom they want their ads served. Facebook guarantees that the ads will run to people with those characteristics, and provides the advertisers with statistics such as numbers of ads served and numbers of people who clicked on the ad. Thus, even though you might receive an ad specifically targeted to someone of your age (for example), the advertiser has not been told your age (or anything else about you).
On that basis, Denham considered the practice of these targeted Facebook Ads to be acceptable (although she thought that Facebook should expand its Privacy Policy to explain more fully the role of advertising and inform users of the use of their profile information for targeted advertising, a recommendation to which Facebook has agreed).
I like this conclusion although I think the issue is very subtle. For example, ads for the web site travelalerts.ca appear on Facebook targeted very specifically to a person’s age: if you are 45, for example, you get an ad headlined “Travel on your 46th b-day”. I think that some people might be creeped out by this, if they had set their Facebook privacy settings to keep their age and DOB confidential. Even though it is not a privacy violation (since the advertiser has not been told your name, or your age specifically, but just that the ad was shown to a certain number of 45-year-olds), it may feel like a privacy violation when you view that ad. (I should also point out that at a news conference, whose audio is online here, Denham misstated Facebook’s policy; she said in reponse to a question from Reuters that if a user’s date of birth is not in the profile, then ads would not be targeted to that user based on age.)
Incidentally, one of Facebook’s terms of service is that users must provide their true date of birth. But during news coverage of the OPPC report, a reporter for BNN admitted that for privacy reasons she had put a fake DOB in her Facebook profile: video from BNN via The Globe and Mail.
On a different topic, one of the subjects on which Denham disagreed with CIPPIC’s allegations was the collection of personal information from sources other than Facebook. The allegation was based on this sentence in Facebook’s Privacy Policy (paragraph 224):Facebook may also collect information about you from other sources, such as newspapers, blogs, instant messaging services, and other users of the Facebook service through the operation of the service (e.g., photo tags) in order to provide you with more useful information and a more personalized experience. Denham noted that according to Facebook, it does not currently collect personal information from outside sources, “but may do so in future and therefore included the above-cited passage in its Privacy Policy” (paragraph 225). Thus she concluded that this part of CIPPIC’s complaint was not well-founded.
I find this troubling as a general principle. I believe that an organization’s Privacy Policy should give a clear sense of how it currently uses personal information in fact, as well as how it might use personal information in future. The Privacy Policy ought to include both these categories, and distinguish between them, and in addition provide some notice to its customers when its practices change such that it begins to actually use personal information in a way that it previously had said it might use it.
If an organization says (in its Privacy Policy) that it may use my personal information in a certain way, then ideally I should have the right to complain to the OPCC, and have it investigate whether such a use complies with PIPEDA, even if the organization has never actually used the information that way. However, I do understand that the OPCC has a heavy workload, and therefore might not want to devote resources to investigating theoretical uses of personal information that are contemplated by a privacy policy but not implemented yet in practice.
Incidentally, one matter that apparently was not raised in CIPPIC’s complaint was the security of the Facebook login process. When you log in normally to Facebook, your password is sent unencrypted over the Internet, which makes it possible for someone to intercept it. This is especially true if you are using someone else’s WiFi network (at a coffee shop, for example). The operator of the WiFi router can get access to all data that goes through the router over an http connection. For that reason, some services (like yahoo.com) default to an https connection (which is encrypted) for the login screen. But with Facebook, the default login page is the unsecure http://www.facebook.com/login.php which does not even have a link to the secure login page https://www.facebook.com/login.php.
Now, the OPCC has given Facebook 30 days to respond to the report. Many people are asking the question, what if Facebook refuses to comply? Since Facebook is an American company, whose servers are presumably all in the US (or in any event, outside Canada), how can it be forced to comply with Canadian law?
In general, the scheme of PIPEDA provides that recommendations made in a report from the OPCC are not legally binding. Thus Facebook certainly would have the right to say, “we disagree with the OPCC’s interpretation of PIPEDA and therefore we are not going to comply with the OPCC’s recommendations; we believe that our current practices fully comply with PIPEDA already”. The next step would be the bringing of an application in the Federal Court pursuant to section 14 of PIPEDA. If the Court determines that the organization has contravened PIPEDA, then the Court might order the organization to correct its practices.
If the Court makes such an order against Facebook, then Facebook might choose to comply because it wants to present the image of being a “good corporate citizen”. But for the sake of argument, what would happen if Facebook still refused to comply? I don’t know the answer to that. Presumably there would be proceedings against Facebook for contempt of court, but then what? Would a finding of contempt give the Federal Court any real power to prevent Facebook from doing business in Canada? Perhaps Facebook might lose the domain facebook.ca but other than that, I’m not sure what the court could do. Would it be practical to prevent Facebook from doing business with Canadian advertisers? This matter might blossom into some very interesting litigation.
|
|
(1 comment | comment on this)
|
| Friday, July 17th, 2009
| |
1:58 am - recent movies
|
The Hurt Locker is opening this weekend; I strongly recommend it. See my comments here.
I liked Moon but it’s too bad that the trailer and most reviews spoiled one of the plot twists. Nevertheless a provocative science fiction movie; Sam Rockwell is great.
Whatever Works was pretty disappointing after a good start. Usually I like Patricia Clarkson but this movie went way downhill as soon as she appeared. Actually it was heading downhill before that. I did like the first ten minutes or so.
Summer Hours (L’Heure d’été): a great film by Olivier Assayas.
Séraphine: I liked this but I got the impression that it was following the documented facts too slavishly. Maybe I’m wrong but it just felt that the events at the end happened without a reason. The film is about real people and in real life it may have been that there is no known explanation for what happened, but if that’s the case, it would have made a better film to depart from that and give some kind of foreshadowing.
|
|
(comment on this)
|
| Sunday, July 12th, 2009
| |
11:56 pm - Burrard Bridge bicycle trial
|
The official start date was supposed to be tomorrow (Monday, July 13) but the barriers were in place today for the Burrard Bridge trial. Previously the bridge had six lanes of car traffic (three each direction) while each sidewalk was split (with a painted line) between pedestrians and cyclists.
Now, one of the southbound road lanes has been converted to a bicycle lane, and the east sidewalk has also been converted to an exclusive bicycle lane northbound (no pedestrians). The west sidewalk is now for pedestrians.
The new barriers were put up yesterday I guess (or perhaps today). Here’s a photo showing a cyclist going the wrong way:
Today, there was plenty of room in that southbound bike lane. There were quite a few cyclists incorrectly going northbound in that lane, and it did not cause any problems that I saw. Will cyclist traffic on weekdays really be heavier than on a nice Sunday afternoon?
As a pedestrian, it was great having exclusive use of the sidewalk. It feels much safer than before, when cyclists would zip by on the same pavement, inches away. But I also am a little upset that the east sidewalk, with its nice views of Granville Island, will now be closed to pedestrians:
|
|
(1 comment | comment on this)
|
| Tuesday, July 7th, 2009
| |
7:17 pm
|
I hate it when the news media report only the relative change in risk, rather than the absolute change. For example:
The Vancouver Province places this text on its cover: “Cancer Risk: Children who use cellphones are five times more likely to develop malignant brain tumours, say researchers”.
First, no one should ever say “five times more likely” because it is ambiguous. They mean “five times as likely”. Literally, “five times more likely” means “six times as likely”.
But that’s incidental. My main objection is that it’s useless to convey the amount of the increased risk as a multiple of the original risk when they don’t tell you the original risk.
For example, suppose the research had found that among children who don’t use cell phones, the risk of developing a malignant brain tumour was 1 in 200, but among children who use cell phones the risk was 1 in 40. If that’s what the research had found, it would be extremely worrying and a very persuasive case to cease using cell phones, at least for children.
But on the other hand, suppose the research had found that among children who don't use cell phones, the risk of developing a malignant brain tumour was 1 in 2,000,000, and among children who do, the risk was 1 in 400,000. That result would not be a strong argument at all against cell-phone use: it would represent a very small increase in the risk of a malignant brain tumour (comparable to the risk of being injured in a short car trip, for example).
In both these examples, it would be accurate to say that the risk among children who use cell phones was five times as high as among children who don’t. But that’s not enough information to make any kind of rational decision. You need to know the absolute difference in the risk, not the relative difference. In these examples, the absolute increase in risk would be 0.02 in the first example and 0.000002 in the second example.
There was a great article on this topic in Scientific American: Knowing Your Chances: What Health Stats Really Mean. As the authors of that article put it: Absolute risks are more informative because they take into account information about background rates. Given the absolute risks, a person can derive the relative risks—but not vice versa. After all, a relative risk reduction of 50 percent could describe either a substantial mortality reduction from 200 to 100 in 10,000 patients or a much smaller one from two to one in 10,000 patients. Randomized trials provide some of the best information in medicine, but unless the results are reported adequately, people will not be able to assess them. Moreover, when reporting on studies like this, without a randomized control group, the news media ought to caution against inferring a causative link.
Thus it appears that the studies linking cell phone use to brain tumours were conducted by interviewing people already diagnosed with brain tumours, asking them whether or not they had used cell phones as a child, and then interviewing a control group of people without a brain tumour, and asking the same question. Regardless of the result of this type of study, you could not conclude that cell phone use had caused an increase in the tumour risk. To prove causation, you would have to start out with a large number of children and then randomly choose some to use cell phones while the rest did not. Only then could you be reasonably sure that the cell phone use caused the increased tumour risk (rather than, for example, some other factor being responsible for both the cell phone use and the increased tumour risk).
Neither the Canwest News Service article nor an article from CTV BC mention the absolute risk. These articles refer to a paper in the journal Pathophysiology. The abstract of the paper is available here but I didn’t feel like paying $31.50 for the full text.
|
|
(2 comments | comment on this)
|
| Sunday, May 17th, 2009
| |
7:58 pm - Star Trek (the 2009 movie)
|
While watching the new Star Trek movie I was enjoying it quite a lot, but the more I think about it, the more problems I see. First of all, could they not have given this movie its own name? Giving it the same title as the franchise, with no subtitle, just causes confusion (especially since there was already a film, in 1979, titled Star Trek: The Motion Picture).
Stop reading here if you have not seen the movie and don’t want to see major spoilers.
( Read more... )
|
|
(2 comments | comment on this)
|
| |
5:42 pm - B.C. Referendum — Why was STV defeated?
|
Last week, British Columbia reran the referendum on electoral reform that was held four years ago, and it is remarkable how different the results were.
In 2005 (link), 57.7% voted in favour of switching to STV, and in 77 of 79 districts a majority voted in favour of switching to STV.
In 2009 (link), 38.7% voted in favour of switching to STV, and in 7 of 85 districts a majority voted in favour of switching to STV.
What accounts for the results? First, in 2005 British Columbia experienced a hugely disproportionate legislative assembly; in the 2001 election, the Liberal Party won 77 of 79 districts. The NDP, as Official Opposition, won only two seats in the assembly despite receiving 21.6% of the vote (link).
Second, I do not recall any television advertising campaigning on either side of the 2005 referendum. There might have been some, but certainly not as much as this year. I would conclude that the ubiquitous TV and print advertising opposed to STV, much of which was deceptive, was a major factor in the referendum results.
|
|
(comment on this)
|
| Sunday, May 10th, 2009
| |
7:57 pm - B.C. Referendum — Is STV too complicated?
|
Clearly the vote-counting process of STV is more complicated than that of FPTP. Is STV therefore too complicated? My answer is no.
First, vote-counting in STV is less complicated than you might think; you can figure it out within ten minutes by reading my earlier posts (here and here).
Second, for voters, deciding how to vote in an STV election would often be simpler than in an FPTP election. The reason is this: FPTP can place voters in a dilemma about whether to vote their true first choice or whether to vote strategically. For example, suppose that of three major parties, there is one you strongly like, one to which you are indifferent, and one you despise. You may believe, either from opinion polls or from a general impression, that the party you strongly like has little chance of winning in your riding, but it will be a close race between the other two. The rational decision, therefore, is to hold your nose and vote for the candidate of the party to which you are indifferent, in order to prevent a victory by the candidate of the party you despise.
But then, what if your assessment turns out to be wrong, and the candidate you like actually has more support than you thought? Then your choice to vote for another candidate might prevent your first choice from being elected!
Under STV, you are never presented with this dilemma. You simply vote for your favourite candidate first, your second-favourite candidate second, and so forth, as many as you wish. The STV system will tabulate your ballot in an intelligent way, so that your ranking of your less preferred candidates will not hurt the chances of your most preferred candidate. You do not have to spend time thinking about how everyone else is going to vote (as you do under FPTP); instead you can spend that time learning about the substantive issues of the election.
That is the strength of STV, and why it is a far more democratic electoral system than the one we use now. That also is why I say that STV is simpler for voters (although it is more complicated for the vote-counters).
|
|
(comment on this)
|
| Tuesday, May 5th, 2009
| |
3:30 pm - B.C. Referendum — STV election results in a Northern Ireland election
|
I had a look at the results for the 2007 election to the Northern Ireland Assembly. There were 18 electoral districts, each electing six candidates. The list of those elected as well as full results of the count (in both spreadsheet or presentation format) and various election statistics are all on the web site of the Electoral Office for Northern Ireland.
I discovered that in 13 of 18 districts, the six candidates elected were the ones with the six highest numbers of first-preference votes.
In the other 5 districts, here’s how the six elected candidates ranked in terms of their first-preference votes:- 1, 2, 3, 4, 5, 7 (Belfast West)
- 1, 2, 3, 4, 6, 7 (Fermanagh and South Tyrone)
- 1, 2, 3, 4, 5, 7 (Foyle)
- 1, 2, 3, 4, 5, 7 (South Antrim)
- 1, 2, 3, 4, 5, 8 (Upper Bann)
Thus, in this real-world example, STV produced results that were very close to what you would get by looking only at the first preferences of voters.
In all 18 districts, the top four vote-getters (on first preference) obtained seats, and in 17 out of 18, the fifth vote-getter (on first preference) did as well.
Now, what about those districts where #7 or #8 got elected but #5 or #6 did not? What is the reason for that, and is it justified?
The reason is this: although the candidate might have been #7 in first-preference votes, there were sufficient numbers of voters who listed that candidate as their second or third preference, thus allowing the #7 candidate to overcome the gap and beat #6. This might well happen when #7 and #8 belong to the same party. You would expect that most voters whose first choice was #7 would have put #8 as their second choice, and vice-versa. The first-preference vote for that party was split between #7 and #8, which is what allowed #6 to finish ahead of both of them. The STV counting system would then aggregate those votes, allowing a candidate from that party to get elected ahead of #6.
|
|
(comment on this)
|
| Sunday, May 3rd, 2009
| |
1:42 pm - B.C. Referendum — web sites dealing with the referendum
|
I am disappointed with the web sites of the official campaigns on both sides of the referendum:
http://stv.ca/ (supporting a switch to the STV voting system)
http://nostv.ca/ (supporting retention of the current first-past-the-post voting system)
Both the proponent campaign and the opponent campaign are funded with public money, each side having received $500,000 from the government.
The stv.ca site is very slick, but I found it difficult to navigate and locate solid information. As a friend of mine said: “Wow, that stv.ca site is political malpractice. There should be a clear, one-paragraph description of what the hell they're talking about, right up top. (Instead, there's a baffling TV spot that explains nothing.)”
The nostv.ca site is less slick, and easier to navigate, but I find many of the arguments there to be slightly misleading. For example, it says: “In the proposed seven-MLA Capital Region, a candidate would be declared elected with just 12.5% of the vote, while in the Northeast, a two-MLA region, a candidate would be elected with 33.3% of the vote. That would give BC what amounts to two different voting systems, and that is not equal effective representation.” That is misleading because it doesn’t mention that the riding with 7 MLAs has a much larger population than the riding with 2 MLAs. It makes total sense that a candidate might get elected by obtaining either a bigger percentage of a smaller number of votes or a smaller percentage of a bigger number of votes. There’s nothing unfair or unreasonable about that.
The web site of the Referendum Information Office, BCreferendum2009.ca, is quite good. It is a neutral service run by B.C.’s Ministry of Attorney General, and it’s easy to find good, clear information there via the navigation bar.
If you want a formal description covering all the details of the proposed BC-STV system, you have to go to the site of the Citizens’ Assembly on Electoral Reform. The Citizens’ Assembly was an interesting experiment sponsored by the government; briefly, two citizens were selected from each riding by random draw from the voters’ list. The assembly convened over a period of months, heard from experts and from the general public, and deliberated on whether it would recommend a change to the electoral system and if so, what system should be recommended.
The formal description of the BC-STV system is in the Citizens’ Assembly’s technical report, beginning at page 31 of the PDF (which is page 17 of the Report).
If you want to watch a video clip explaining STV, you could go to the stv.ca site, but I found their videos a bit tedious. There’s a better one at a web site of the New Zealand government: www.stv.govt.nz/STV/how.htm. It should be noted that the version of STV used in New Zealand (for some local elections) is a little different from BC-STV. The New Zealand version is more complicated, but they are still able to explain it clearly.
I also found a good brochure from New Zealand concisely setting out the pros and cons of both STV and FPTP.
The Library of Parliament research service has put out what looks like a good report: Canada’s Electoral Process: Frequently Asked Questions.
As for blogs covering the B.C. referendum, Bill Tieleman’s blog is covering the most cogent arguments to vote against adopting STV, such as in this post. For a pro-STV viewpoint, I would suggest visiting ereform.blogspot.com. Another pro-STV blog is FairVoteUBC.wordpress.com, but I don’t find the content there particularly informative.
In my next blog entry, I am going to comment on some recent elections where STV was used as the voting system, from Northern Ireland and Scotland.
|
|
(1 comment | comment on this)
|
| Saturday, May 2nd, 2009
| |
12:53 pm - B.C. Referendum — STV in detail
|
The most common criticism of the STV voting system is how complicated it is. I have two responses to that.
First, even if you consider the vote-counting process of STV to be complicated, the voting process is not. As a voter, all you have to do is indicate your first choice with a “1”, then if you have a second choice, indicate that with a “2”, and so forth, ranking as many candidates as you want. That is simple to understand. I concede that ideally, all voters would understand the vote-counting process, but as long as they understand how they should mark their own ballot, it isn’t really necessary that they understand the intricacies of how the ballots will be tallied.
Second, the vote-counting process is not really all that compicated. Yes, certainly it is more intricate than our current voting system, “first past the post”, where in each riding you just tally up each candidate’s votes, and whoever has the most votes is declared elected for that riding. Vote-counting in STV takes a few minutes to comprehend, but it isn’t rocket science. All you need is to understand fractions, which I believe is covered is grade six math.
The three basic points of STV were set out in my post yesterday. Once you’ve looked at that, I can explain the vote-counting process in detail (along with a detailed example), within ten minutes, as long as you understand grade six math. In fact, I’d appreciate it if anyone reading this would time how long it takes to read the following, and then respond in the comments, so I can know if my “within ten minutes” claim is correct.
BC-STV vote-counting process in detail
Each riding will be represented by a number of candidates, from 2 to 7, depending on the population of the riding.
As a voter, you rank the candidates by preference, starting at 1 for your first choice, then 2 for your second choice (if you have one), and then continuing like that. You don't have to rank all the candidates; you could choose just your top 5, or top 2, or top 1, or however many you want.
Each party would (likely) run several candidates. For example, if there are five candidates to be elected, the Liberal Party might field four candidates (or more if they want). The idea would be that the Liberal campaign would encourage voters to rank the four Liberal candidates as their top four choices.
In counting the votes, the first step is to calculate the "Quota" based on the number of valid votes and the number of MLAs to be elected. For example, if there are 100,000 valid votes and 3 candidates to be elected, the Quota would be 25,001. The logic behind that is, with only 100,000 total votes, you could not have four candidates each with 25,001 votes. If there were 100,000 valid votes and 5 candidates to be elected, the Quota would be 16,667.
Then you tally the votes according to the first preferences only. Any candidate whose votes meet or exceed the Quota is declared elected.
If your vote helped get that candidate declared elected, then there will be a fractional transfer to your next expressed preference. The fraction is based on how much the candidate exceeded the Quota. For example, suppose the Quota is 16,667 and a candidate "Amy" receives 28,806 votes. Then, by dividing, we can say that 57.86% of Amy's votes were "necessary" to get her elected while the remaining 42.14% were "surplus". Therefore, if your first preference was for Amy, then 57.86% of your vote will "stay" with Amy, while 42.14% of your vote will be transferred to your second preference.
For example, suppose that of the 28,806 first-preference votes for Amy, exactly 10,000 of those ballots had "Bob" marked as a second choice. Then, after the first count, Bob would get 4,214 additional votes, added to however many first-preference votes Bob received.
So after each iteration you transfer the "surplus" of any elected candidate to the voters' next choice. Then you check whether any other candidate has a vote total meeting or exceeding the Quota. If so you declare that candidate elected and transfer that candidate's surplus votes.
If no candidate is elected, then you eliminate the candidate who currently has the least number of votes. All votes of the eliminated candidate then, obviously, are transferred to those voters' next choice.
This is repeated until you have as many elected candidates as you are supposed to have.
So continuing the example from above, you're a voter, and your first choice was Amy, and your second choice was Bob. Recall that after Amy was declared elected, 4,214 votes were transferred from Amy to Bob (or to put it another way, 42.14% of each of 10,000 votes were transferred to Bob). Now suppose that the transfer of Amy's surplus does not push anyone else up to the Quota of 16,667. Then we eliminate the candidate with the least number of votes. Let's suppose it's Bob (your second choice). Okay, so Bob is eliminated and all Bob's votes go to the next choice on the various ballots. As for your particular vote, let's say your third choice was "Charles". So 42.14% of your vote is transferred from Bob to Charles. (Recall that the other 57.86% of your vote stayed with Amy.) And let's say that now, Charles's total is 18,700. This is over the Quota, so Charles is declared elected, and then Charles's "surplus" 2,033 votes are transferred to the voters' next preference. Since 2,033 (the surplus) is 10.87% of 18,700 (Charles's current total), 10.87% of the value that each vote had (with Charles) will be transferred. In other words, if I voted for Charles as a first choice, then 10.87% of my vote would be transferred to my second choice (the other 89.13% of my vote stays with Charles). But for someone who voted for Amy ahead of Charles, only 42.14% of their vote was transferred to Charles, so to find out the amount of that vote which is now transferred to the next preference, you just take 10.87% of 42.14%, which is 4.58%.
As for your own ballot, let's say your fourth choice was Donald. Then your contribution to Donald's total will be 4.58% of a vote.
Then if you want, you can analyse exactly where your vote has gone, at this stage of the count:
57.86% stayed with Amy, helping her get elected
37.56% stayed with Charles, helping him get elected
4.58% currently with Donald at this stage of the count
(The total of course is 100%.)
So I think the above is a pretty good explanation of the counting process, with a detailed example. It doesn't cover every technicality such as how ties are dealt with, and a few other things, but overall I think I'm right in saying that anyone who understands grade six math can get a good understanding of STV in ten minutes.
|
|
(comment on this)
|
| Friday, May 1st, 2009
| |
11:13 pm - B.C. Referendum — STV in a nutshell
|
In 2005, British Columbia held a referendum on replacing the electoral system.
A majority, 58%, of voters supported changing the electoral system to a version of “single transferable vote” (STV). However, the electoral system did not change, because it was felt that a threshold of 60% should be reached in a referendum before changing the system.
As a result, the government chose to repeat the referendum this year. It will be held on May 12, in conjunction with the provincial general election.
I am going to post a series of blog entries relating to the referendum. To begin with, what is STV? Here are the three salient features of this electoral system:
(1) on your ballot, you rank the candidates according to your preferences;
(2) during counting, if one of your candidates is eliminated (having the least number of votes at that stage), then your vote will go to your next expressed preference;
(3) if your vote helps one of your candidates get elected, then a portion of your vote will go to your next expressed preference.
That’s STV in a nutshell. If STV is adopted, then the above three points are all you would really need to understand, in order to participate in the next election.
In my next blog entry, I will elaborate on the vote-counting process under STV and give a detailed example.
For further information, I would recommend visiting the Referendum Information Office (a neutral service run by the B.C. government) at BCreferendum2009.ca.
|
|
(comment on this)
|
| Friday, March 20th, 2009
| |
8:52 pm - Time zone abbreviations
|
One of my pet peeves is erroneous or ambiguous time zone abbreviations. A sidebar on the front page of today’s Globe and Mail provides a good example:
(The story is online here.) The timeline says that 10:40 a.m. GMT is 6:40 a.m. EST, but it should have said 6:40 a.m. EDT.
Eastern Standard Time (EST) is used in Montreal (and elsewhere in the Eastern Time Zone) from the first Sunday in November until the second Sunday in March.
Eastern Daylight Time (EDT) is used from the second Sunday in March until the first Sunday in November (except in places like Coral Harbour, Nunavut, which stays on EST year-round).
This is not complicated. A reputable newspaper should be able to get this correct, particularly on its front page.
A useful alternative to EST/EDT is the abbreviation ET (for Eastern Time), meaning whichever of EST or EDT is in force on the relevant day. For example, Nokia Canada’s contact-us page gives its customer-support-line hours as “9:00 AM - 7:00 PM EST” which creates ambiguity about what the hours are when Daylight Saving Time is in force. Nokia ought to have used the abbreviation “ET” instead of “EST”.
While I’m on this topic, let me argue against the term “GMT” (or Greenwich Mean Time). The problem is that GMT traditionally refers to a time scale determined by the rotation of the Earth (now referred to as UT1). This time scale differs by up to 9⁄10 of a second from UTC, which is based on a more accurate atomic time scale. Civil time as distributed by official sources is invariably based on UTC, not UT1. Therefore the terms “Greenwich Mean Time”, “Greenwich Time”, and “GMT” should be avoided in favour of UTC (or, if appropriate, WET/WEST for Western European Time/Western European Summer Time).
Relevant links:
|
|
(2 comments | comment on this)
|
| Friday, October 3rd, 2008
| |
1:30 pm - VIFF Day 2
|
As I Lay Dying: Sort of interesting, though I didn’t quite get it; were those supposed to be stigmata? 6
Escape: I found this boring. 3
Summer Afternoon: The basic concept of this short is unoriginal; the execution of the concept does occasionally show flair but the film is just too muddled. 5
The End of the Tunnel: The best of this group of shorts, but still not that great. 6
Corridor #8: I like maps. In my opinion, movies in general do not make enough use of maps. Even in narrative film, if a movie is set in a small town in France or wherever, why not show us a map of where the town is? But particularly in documentaries, I always find it frustrating when a place is being discussed and they don’t show you where the place is, what it is near to, and so forth. Film is a visual medium, so showing maps is easy. Corridor #8 is a prime example of this common failing. The film is about a proposed highway from the Black Sea to the Adriatic Sea, traversing Bulgaria, Macedonia, and Albania. But the film never shows any sort of map or diagram of where these countries are in relation to each other! The film does use various graphical effects like a particular way of indicating the number of miles along the proposed route, and I appreciate that the point of the film was not so much to discuss the road itself but rather the differences between these Balkan cultures, but even there maps would have been useful. I just find it hard to believe that it never occurred to the filmmakers to show such maps. 4
Orz Boyz: Kind of cute, but overlong. 6
|
|
(comment on this)
|
| |
12:03 pm - VIFF Day 8
|
When It Was Blue: A thoroughly engrossing film made by overlapping two 16mm films. The images are often recognizable, sometimes abstract, but never boring. I strongly recommend this; it plays again tonight (Friday) at 9:30. 9
Sheltered Life: I have seen a lot of boring films at VIFF this year, where all I could think of during the film was “when is this going to end”. Sheltered Life was not like that; it was completely watchable and interesting, but nevertheless not quite successful. Something seemed off about the geography of the movie; the script and actors suggested an Ontario location, but some scenes had obviously been shot in Vancouver. It gave the movie an artificial feel, reinforced by the style of acting (which you often see in Canadian movies and TV) that is more theatrical than cinematic. Questions kept occurring to me about how the characters were moving around. 6
The Romance of Astrée and Céladon: This movie was just so stupid, I kept think that Rohmer must be playing an elaborate in-joke, but I couldn’t figure out whether he was mocking the audience, movies in general, or himself. By the end I was wondering if he was intending to be serious; if so he failed completely. 3
Heaven’s Heart: A clever little Swedish drama, about two married couples. You find out at the beginning of the movie that one of the couples is getting divorced, then the action moves back several months to reveal why. The style is to use many close-ups, and it works because of what it catches in the faces of the actors. 8
|
|
(1 comment | comment on this)
|
|
|
|
|
![[info]](http://l-stat.livejournal.com/img/userinfo.gif){kind=small}
mathew5000's journal
