Sunt Lacrimae Rerum

A little bit of checking showed that the algorithm was essentially as it had been four years ago. Several lines were marked as mine because I had converted the macro TRUE to the boolean value true on a couple of lines, and one because I had taken a freestanding C function and turned it into a member of the class in which it operated.  For practical purposes, the code was the same as it had always been - but my name was all over it. In addition, the problem would be expected to take the form of a line having dropped out, and there is no blame tracking attached to deletions.

In actual fact, the conclusion to be drawn was that the code was legacy code. Minor tweaks obscured that fact.

Git blame operates on a per-line basis. But any change to the line - tweaking parentheses, for example, or converting a C-style cast to a C++ cast - makes you the owner of the line.

On a greenfield project where responsibility is doled out in blocks it might be useful, but on a legacy projects it's worse than useless.

By coincidence, I had been looking at the blame record for a makefile the day before.  The makefile had an if-else block where both branches had the same statements. (In other words, there should have been no if-else block, but just the list of statements.) Blame shows five different names associated with the block of the code (all of whom, except (I think) the oldest one have some passive responsibility for the poor structure) but not in any coherent manner. 

When I look at a block of code and want to see its history, I want to see how the algorithm evolved, not how different lines were tweaked while retaining the same algorithm.

You can't avoid this problem as long as your algorithms are line-based. It's a whole different level of difficult to provide a program which divides a program into logical chunks and applies that analysis to the raw record; or (worse) to determine when apparently minor changes create new logic but skip over better implementations using the same logic. (A for loop and a find_if statement may be exactly formally equivalent, but don't expect any automated help to know that.)

So I will continue to avoid git blame. If I really need to look at the history of code .... I'll look up diffs from the history of the codebase and look at them as integral wholes.

class Foo;

bool matches(const Foo& inVal);

std::vector<Foo> x;

std::vector<Foo>::iterator i = x.begin();

while (i != x.end())

{

    if (matches(*i))

        i = x.erase(i);

    else

        ++I;

}

I raised two issues with it: one, that there was a bug in it (it left out the else, meaning that where consecutive members of the vector matched the condition the second would be skipped).  The second was that the cost was fairly high; at the time I raised this by asking whether it made sense to replace the vector by a deque (reduces the cost for large sets of data) or a list (reduces the cost for all sets of data, but can raise the cost of other operations significantly), and was told that it would be a short vector and that the overall expense would therefore be low. (The cost is high because every time an element is deleted all the following events have to be shifted left once.)

About 24 hours later I asked about the use of remove_if and erase, shown below:

class Matcher {

public:

     bool operator()(const Foo& inVal) const {

         return matches(inVal);

     }

 };

x.erase(std::remove_if(x.begin(), x.end(), Matcher()), x.end());

I received the answer that the use of remove_if required extra work for a minor bit if code and it was left there, as far as that issue went.

But I remained curious.  How much extra work was it? So I actually drew up the code bits above and counted lines. If we leave out the setup (first three lines) the STL code snippet is actually one line shorter.

(This had to be workable in a C++03 compiler. A C++11 implementation with the STL algorithm using lambdas would be shorter still.)

What are the benefits of the STL approach? Well, first, it's not fragile; the bug caught on the code review is impossible in the equivalent STL code. The bug is directly related to another minor weakness of the iterator-based code: because of the variable incrementing logic it can't use an idiomatic for loop, but has to use while().

Secondly, it's faster; remove_if is an O(N) operation and the final erase, which only resets the end of the vector, is very cheap indeed.

Third, it is immediately clear what the code is doing. Because of some additional verbiage in the actual code (plus the missing else, which obscured the firm of the loop) I had to look twice to be sure of what was going on.

This is, in a nutshell, an illustration of why, in general, STL algorithms are preferable to hand-rolled loops: more reliable and generally faster. There are some subtle ways of using find_if and transform in unusual ways[1] but remove_if is clear about what it does.

But it's also an illustration of the implicit prejudice I find in developers against using the STL algorithms. The combined use of remove_if and erase is a very well-known idiom; there is nothing difficult about it. But the perceived - but not actually significant - overhead of having to provide a functor to feed to the algorithm [2] seems to constitute a mental barrier.

[1] For example, using find_if for "process every item in this ordered vector below a given value", for which it is admirably suited. Using for_each is less efficient, as find_if will cut off processing once the relevant value is found or surpassed if your test is written correctly. But you throw away the result of the search, which is not normally expected.

[2]Of course, even that is not real. If your test is in a function, as above, you can just pass a function pointer, although it is likely to be slightly less efficient.  If it's not in a function, it's a logical unit and either should be in a function or in a functor in any case.[3] (In the actual code the test was actually not a function, but a comparison expression. This actually did mean that the number of lines needed for the STL approach would be a little greater, as the test needed to be encapsulated for the STL where it had been merely dropped into the loop.)

[3] The marginal case being where this test is used only here. If it expresses a meaningful concept in the problem domain, though, it's a good bet that it is not.

I had just finished West's book on Indo-European myth and in particular noted at the end of it his identification of the deep roots of the comitatus in Indo-European culture - the king as the leader of a band of warriors whom he attached via, essentially, handing out booty or providing the opportunity to pillage.

The next day I was reading a book dealing with (inter alia) the conversion of Anglo-Saxon England, and read: "Seventh-century English kings did not 'govern' in any sense that we should recognize today. Their primary business was predatory warfare and the exaction of tribute from those they defeated. The spoils of successful war - treasure, weapons, horses, slaves, cattle - were distributed to their retainers as payment for past and lien upon future loyalty."

So the pattern described above has deep roots in Indo-European culture. Traditional poetry, whether about the Trojan War or a successful cattle-raid, reflects this.

In Europe generally, it is the Twelfth and Thirteenth Centuries which see the movement away from this pattern, at least at the local leader level. Prior to that point it was common for leaders well below the level of monarch to carry out raids and low-level warfare against their neighbours; it is in this period that the state begins to exert its centralizing powers to curtail this activity.  In areas where national borders were involved it persisted for rather longer (such as the Scots Borders). The general basis for lordship becomes, not the distribution of booty, but the conferring and defending of rights to land (or patents, or other privileges) which can generate a continuing stream of revenue for the holders.

(Booty doesn't entirely go away. Soldiers continued to be given the implicit, and sometimes explicit, permission to pillage on campaign, and even after it became entirely frowned upon (consider Wellington's army in the Peninsular War) remained (and remains) a problem. (How many Allied homes have "souvenirs", like the Roman bust discovered recently in Texas, picked up by soldiers in the Second World War?) But it was no longer the systematic basis of lordship.)

Put more bluntly, the government doesn't care about people getting sick, they care about hospital overcrowding. (This has been visible and explicit since the start of the pandemic.) They also badly want the whole thing to be "over" by the time of the June election.

It is also important to recognize that the primary benefit of masks - except at the high end, which involves respirators proper - is (1) at a population level and (2) that they protect other people from the mask-wearer more than vice-versa. So saying that people can still choose to wear masks misses the point; the people who choose not to wear masks are likely to skew less cautious in other ways and therefore as higher risk.

Masking is a low-cost high-benefit practice if generally adopted in public places. Dropping a general mask mandate does not just verge on the irresponsible but goes well into that territory.

At an individual level the obvious strategy to take is to limit going to places where there is a significant number of Individuals one does not know and to wear a respirator, not just a cloth mask, when inside public places. Shopping online is still a better option in most cases; if one chooses to shop in brick and mortar locations, relatively smaller locations with better ventilation are better choices. Voting against the government in June is also a good idea - assuming that the opposition parties are willing to back continued restrictions. (I hold no real hope of this.)

(A classic example in software development is to measure productivity in lines of code, and make compensation dependent on productivity. At best, this provides no incentive for clean, concise code; frequently, it actively encourages coders to write deliberately verbose code to boost their lines of code.)

I have recently observed a specific example of this in another domain, namely, the TTC.

I regularly use a bus route which is relatively short and which is essentially a direct line from the station for most of the route, but splits into a loop at the end. I catch it on the leg going back to the station.

During rush hour, the predictions provided by the feed based on real-time data from the TTC are fairly accurate. Mid-day, they regularly would result in missing the bus.

The route is scheduled with busses N minutes apart, based on the number of busses on the route at the time; every 15, 20, 30 minutes (10 at rush hour).

To maintain the schedule, the drivers are supposed to get to the far point of the loop where there is a stop they can wait at and then come back at a scheduled time.

Because the apps available for TTC predictions track busses in near-real time, I can see what happens: busses wait at the wait point until they are three minutes early, and then leave. This means that what was a ten-minute prediction when I looked at it, and a six-minute prediction at the point they start to move, suddenly collapses by a significant amount.

What is going on? The problem lies with the TTC's metrics. They count a bus as being on time if it is within a three-minute window on either side. This is meant to allow for problems with heavy traffic or construction which drivers can do little about, or for random times when fewer people are waiting for stops than the statistical average, speeding up the bus by reducing the number of stops. It is not meant to encourage drivers to start moving as soon as they are technically "on time", but that is what systematically happens.

They start moving early because they will get to the station "on time", but really three minutes early, and then have a three-minute longer break. (The same drivers tend to arrive on one side of the station, let off their passengers, and then sit there until just before they have to leave, proceeding to their loading bay only at the last moment, even in very unpleasant weather. The TTC consistently conveys a sense of being run for the benefit of its employees rather than its customers.)

At rush hour there are no delays built into the schedule; the only divergences from longer-term live predictions will be as a result of heavy traffic, which is rare on this route.

This leaves users with no recourse. The drivers have committed no formal infraction. The TTC could analyze the collected route data and penalize this behaviour if a driver engages in it systematically, but I'm willing to bet that the likelihood of the Union to grieve such a change in the rules is a barrier to such a step.

Surpris’d I was with sudden heat which made my heart to glow;

And lifting up a fearful eye to view what fire was near,

A pretty Babe all burning bright did in the air appear;

Who, scorched with excessive heat, such floods of tears did shed

As though his floods should quench his flames which with his tears were fed.

“Alas!” quoth he, “but newly born, in fiery heats I fry,

Yet none approach to warm their hearts or feel my fire but I!

My faultless breast the furnace is, the fuel wounding thorns,

Love is the fire, and sighs the smoke, the ashes shame and scorns;

The fuel Justice layeth on, and Mercy blows the coals,

The metal in this furnace wrought are men’s defiled souls,

For which, as now on fire I am to work them to their good,

      So will I melt into a bath to wash them in my blood.”

      With this he vanish’d out of sight and swiftly shrunk away,

      And straight I called unto mind that it was Christmas day.

- Robert Southwell

In what fit of absence of mind they concluded that Covid would be so completely gone in another six months that they can plan on pulling back vaccine mandates, it is hard to imagine. The opening stages - allowing more density in places with vaccine mandates - is not prudent, but it is somewhat understandible. To assume that magically Covid will go away by March is pure wishful thinking. (Even if by some miracle Ontario could wrestle it to the ground, there would still be extensive reservoirs elsewhere.)

This is the extreme form of the government's complete failure to take the one critical step that is necessary: to say, clearly, that there will be no "return", no ability to resume the life of 2019. Even if the government were to drop its vaccine mandates entirely, all the employers and other fora will continue to worry about insurance and liability and are unlikely to drop their restrictions. Many people will, rationally, continue to avoid places where they are crowded together; although evidence from other jurisdictions suggests that fewer people are rational than one might hope. Masks will continue to be an important public health tool. Many people will continue to work from home, affecting the life in urban centres.

The one genuinely bright spot on the horizon is the (likely) very near approval of Covid vaccines for children and the (equally likely) of such vaccines being made mandatory (as many other vaccines are) for attendance at school.

It is, frankly, the task of the government, the parens patriae, to tell everyone the truth: life has changed, irreversibly (like one of those catastrophe theory transitions on a folded manifold).

Last election I was already noting that the Liberals would do well to get rid of Trudeau, though that was before Covid, which pushed up his stock somewhat. I would reiterate the comment; he is clearly a drag on his party, but his retention of office without a significant loss of seats probably leaves him with enough power to avoid being pushed out.

What the Conservatives need to do, to deflect their usual problems next time round, is take the relatively moderate platform O'Toole ran on and maintain it up to the next election. As I gather that there are members who already want to depose O'Toole, I doubt they will manage it.

(The best grounds for attack on O'Toole the Liberals had, they wasted: pointing out the drift between his run for the leadership and his run for PM and suggesting that he is a weathervane seems to me like the obvious weak point. By next election he will be gone or his platform will have some stability. I doubt he can reinvent himself yet again.)

But the map shows the kind of current electoral trap Canada is in. There are two broad areas of solid blue: the west (for all practical purposes: the NDP has a few new seats there), and rural Ontario. (That these are the areas with the biggest Covid problems is not coincidental: in addition to Alberta and Saskatchewan, about which little needs to be said, it is the rural caucus in Ontario which by all accounts has been most opposed to reasonable anti-Covid measures.) In Ontario the whole GTA with a couple of exceptions went Liberal, and of those exceptions only one - Thornhill - went Conservative.Other urban areas voted Liberal, except where they were weighed down by large rural components of a single riding (such as Peterborough).

There are two big blockers to any majority government. First, the Bloc essentially prevents the Liberals from gaining a majority; their only other normal way there is a thoroughgoing collapse in the Conservative vote. (Both Bloc and Conservatives have been weak before; the Reform/PC split gave the Liberals under Chretien a long time in majority office, taking all, or nearly all the ridings in Ontario to make up for the Bloc in Quebec, and the Bloc was weak in 2015.)

The Conservatives, on the other hand, have tended to win only when dislike or distrust of the Liberals - Chretien's cronyism, Trudeau's / Turner's patronage appointments, distrust of Dion and Ignatieff - drives the electorate in the swing ridings of Ontario to abandon the Liberals for the Conservatives. Although the Tory base certainly feels that way about Trudeau, his weaknesses among the general population - seen as naive, self-centred, not terribly bright - have not, to date produced that result. So the Conservative road to a majority runs through the suburbs of Ontario. The past several elections have been marked by a significant number of voters voting, not for the Liberals as such, but against the Conservatives when there is a risk of them winning, and younger voters in central Canada are even less likely to vote Conservative. To make things worse, the Covid crisis has made government with an activist and intrusive bent more generally attractive, which means that the Harperite/libertarian block making up the core of the parliamentary party is even more out of step than they might otherwise be.

The geographical split encourages the parties to vanish into the event horizons of their own bases, which decreases the likelihood of a breakthrough.

The NDP lacks a base which could realistically allow it to come within shouting distance of power, except of the sort they have right now as the key to the continued governance of the Liberals. The Greens essentially disintegrated.

All that being said, it would be unwise to claim, as I've seen some commentators do, that Canada is locked into a permanent minority situation, or - a closely related claim - that the Canadian people chose a minority. Over half of active voters almost certainly belong to the bases of the two major parties. Presumably they wanted a majority. (The NDP base might welcome a minority, but I doubt they voted for one.) An average of people's choices, filtered through FPTP, cannot be given a notional will or mind, any more than the players of Prisoners Dilemma want the probable outcome.

More critically, a lot turns on immediate personal reactions. The voters of Ontario elected Doug Ford because they were fed up with the Liberals, despite the chaos of the Conservatives and the very real doubts about Ford's suitability to govern. If the Liberals do something stupid, not necessarily more stupid than the various kerfuffles which marked, but did not end, Trudeau, O'Toole or an equivalent could walk into government. If the PPC or an equivalent splits the CPC vote more heavily and along structural fault lines, we could be in for another King-St.Laurent reign.

Or at another level, if post-Covid remote work sends more urbanites out into semi-rural Ontario, that block of rural Ontario CPC ridings could become far less certain. If the Alberta NDP becomes the government again and the post-Kenney PCs became a hissing in the dark, that solid blue block could splinter. Or if the dislocations of climate change throw most people's lives into uncertainty, the resulting resentment may drive some voters towards the right while driving others towards "a safe pair of hands". The current social and economic context is the opposite of stable.