30 March 2009

The beauty of the rain

is how it falls.

29 March 2009

Spring Cleaning

Every time I see a bird doing this, I think I should be putting in a bird bath, and then I think about how to keep it usefully clean and despair. Maybe some place where I can have a circulating pond and fountain sort of setup.

Anniversary Rhino

Somewhat too much sunlight on the hay, but I think the disturbing view of the neck wrinkles makes up for it.

It seems like a better photograph than the 2008 rhino, so maybe I am learning something.

28 March 2009

The purpose of a system...

is what that system does.

What the system is actually observed to do; it doesn't matter what people say they want it to do, or it is meant to do, or is trying to do; the purpose of the system is what the system is actually observed to do.

If you want different behaviour, you need a different system.

One of the major problems with "too large to fail" financial institutions is that the system we've actually got is one in which the people running the institution make millions of dollars irrespective of the performance of the institution. (Canadian chartered banks aren't particularly bad about this, but they're not good. Everywhere else is apparently worse, which horrifies me.)

There's absolutely no need for a dedicated financier caste; there may have been, way back when finance was about personal reputation and letters of credit and moving gold around a hostile landscape, but in a world were money is increasingly not merely an abstraction but an electronically mediated intangible abstraction, and there's any number of numerate, educated people who would be quite happy to do the job for a modest professional salary, there's absolutely no need to treat people who work in finance as special. (No one's special, at the level of public policy.)

It's very easy to define a bank as any institution that takes money from a member of the public and either stores it or invests it on that individual's behalf, and which is, as a bank, constrained to compensate its employees, officers, directors, and members of the board no more than some small integer multiple of the median family income for that country. (I would suggest that this integer multiple be selected as 3.) This is total compensation; salary, stocks (priced at larger of the market price at time of grant or exercise), mobile phone plans, provided car, everything.

If a corporation is a person, it's a member of the public.

If a corporation is not a person (and I don't agree that they are), I still can't think of a reason for the rules to be different for taking a corporation's money and a natural person's.

No single bank can have more than 3% of the market. If a bank exceeds that size, it is immediately liquidated. Not sold, split, or otherwise altered and continuing, but fed into the existing provisions for the winding up of a business under the bankruptcy laws.

I can't see how this would do any harm to the provision of financial services whatsoever, and it would provide the considerable good of removing working in financial services as a plausible means of accumulating great personal wealth.

26 March 2009

Righteous Profit

Some of the discussion of various CEO and bankster bonuses generally has come down to "people have a right to make a profit".

Rights are things that apply equally to everybody.

If making a profit is to apply equally to everybody, that's a rather strong form of socialism.

Doesn't seem likely that socialism is what they mean, does it?

If it does not apply equally to everybody, the version being put forward is a con; they're using "right" as a word that can't be argued with, instead of "privilege" or "prerogative". ("Members of my class deserve to have society arranged to guarantee we get money"; it's not really any different as an attitude than not letting the serfs move around was, in particularly oppressive aristocratic regimes.)

There's a huge difference between profit as a necessity; the thing you have to have to keep doing the otherwise interesting and worthwhile thing, and profit as a goal.

Profit as a goal, profit maximization, eventually destroys value. Value is the ratio between cost and benefit; five dollar shoes that last a month are worse value than fifty dollar shoes that last three years. (And yes, I'd like it if I could get long-lasting fifty dollar shoes, too. But the point of Vimes' Boots isn't real numbers, it's the ratio of cost and benefit.)

So if you're maximizing profit, you are setting out to reduce that cost:benefit ratio; you're reducing the value your produce or service represents to your customers.

So, sure, profit, is a necessity on some scale of time; what it isn't is a virtue. Virtues are good things in any quantity; you can't have too much generosity or courage or kindness in the world.

It's very easy to have too much profit; any time the benefit doesn't exist and the profit does, the value is undefined, and the profit is too great. For people who have crashed the banking system of most of the world in such a splendidly thorough way the crater is not all dug yet, I'd say it was very clear the benefit doesn't exist and so any profit is illegitimate.

Same thing with lots of pork barrel projects and public-private partnerships; the private half is rarely interested if they don't see it as an opportunity to either guarantee profit (irrespective of benefit, which is again damaging to value) or increase profits.

It's not even that difficult to measure value; measuring value could do with being discussed a lot more.

22 March 2009

I think Brad DeLong is wrong.

Brad's come up with a "useful incremental step" summation of the Geithner Plan, and noted, with due trepidation, that he thinks Paul Krugman is wrong to be appalled by said plan.

Well, no; Krugman is not wrong to be appalled, for a very simple reason Brad consistently leaves out of his analysis. (I cannot decide if this is naivete, the pose of naivete, or an habitual personal benevolence leading to a kind of blindness, nor would I be likely to ever get enough information to decide, but it's a very curious blind spot for an economics professor to have.)

The people who run the US financial system are effectively unanimous in believing that its core purpose is to make sure wealthy people stay wealthy. This is strongly tied to beliefs about class; confirming those beliefs about class is a very large part of Regan's success as a politician. (That they all believe that they themselves properly belong to the class of the wealthy is also a problem, but a secondary one. The core problem would exist if they were a bunch of obsequious managers for wealthy patrons and themselves content to subsist on crusts.)

While that characterization of the people running the US financial system is true, disproportionate societal assignment of risk from what might as well be the capitalist class to the general public purse will happen not by necessity but by preference. The people running the financial system believe that this is the only really legitimate function of government; to make sure the money comes to them from everyone else. Given the slightest chance, they will act to achieve another private profit by means of socialized loss. Since the management of a fund more or less requires authority over the monies, and we know for sure that these are not honest people -- because no one honest would ever have worked for decades to remove the regulations that prevented them from creating the meltdown, nor played the game once the regulations were gone -- it is not a probability but a certainty that there will be another mighty theft.

Leaving aside that this deserving-of-wealth attitude really is the attitude of slave holders -- that they have a real and active right to the benefit of the labour of others, a right that those who labour may not refuse to grant -- and that these are the same people who have run the global financial system as, at best, blind fools, and at worst as something that uses crimes between fraud and counterfeiting as a means of personal enrichment, that belief in a proper immunity from the consequences of failure guarantees that, long term, the economy won't work. No system will stay functioning when the negative feedback has been assigned to a portion of the system lacking the command authority to generate the action that produces the negative feedback. It may, if it was diverse and diffuse and tough to start with, take it a couple of generations to kill it, but that is a protracted and agonizing death, not a stay of execution.

No-one is actually special, in any statistically valid sense; special to spouse and children and parents, close colleagues, the individuals of the small-scale social network that defines human interaction, sure. Special on the scale of economies and nations, none and no-one are. Every single individual in the banking system could be replaced; it would be moderately more difficult to do it all at once, but it could be done.

In the case of the those responsible for the meltdown, it's not enough to replace them. Banks work on trust -- you have to believe your money is really in there, if you need it, and if you want to believe that, you do not give your money to clever thieves -- and until that trust is restored, no amount of policy will be efficacious.

The CEOs and CFOs, the board and the blind regulators and the auditors, the top layers down to the people who aren't called VPs any more, need to wind up pauperised and imprisoned. (Both, not one or the other.) The regulations need to be re-written to function in an utterly transparent and fully distributed way, so that there are many auditors, most unknown, too, because with a billion dollars to do it with, you can corrupt anybody. But for a beginning, the people the whole world knows are thieves have to be removed from positions of responsibility over money.

No plan that doesn't start with will have, or could have, a positive result.

21 March 2009

Minor lapses in dignity

The wrong lens, in some respects -- this is a crop out of a much larger frame -- but it's the lens that was on the camera.

19 March 2009

Alive and well and made of cuteness

I don't get away with all of these shots, but I get away with some of them:
I'm doing pretty well, too, but no one has ever suggested I'm made of cuteness.

Almost done this contract; need to confirm that everything works in all the tested permutations, but so far it's looking fairly good. And I can now say I've done DITA bookmaps to PDF processing and multi-level indexes with XSL, which isn't a bad thing, whatever my neurons have felt like at certain points this week.

15 March 2009

The Blue Jay of Happiness

It's the blue jay of happiness because while I often hear them, I don't so often see them, and still less often do I manage to get the camera out before a very suspicious corvid hastens away.

14 March 2009

Type mismatch

Not being on Live Journal has its advantages; the vast mess that is RaceFail I have encountered extremely peripherally if at all.

In that peripheral awareness, the idea has come up a couple of times that "so and so is smart, of goodwill, and conscious; of course they're not racist", and, being me and possessed of the sense of rhetorical self-preservation of a stunned duckling, THIS I want to comment on.

Racism isn't a personal moral failing; it might involve personal moral failings (maintaining public racist views as a means of maintaining profits, for example, or for purposes of avoiding effort or social difficulty), but at root racism is worse; it's a general failure of world view. This is not something intelligence or a conscious approach to one's life are directly helpful with. OF COURSE that person of good will, general smarts, and conscious approach to their life can be racist; if they're using a world view based on types, they are more or less inherently condemned to be, because it's not possible to avoid being racist if you model the world using types.

It is, unfortunately, a very easy failure of world view; thinking by types is the historical case, and is correctly the case in the Newtonian physics and inorganic chemistry that so influenced Enlightenment thought. There really is an ideal type of an electron or a type of mineral crystal, about which valid generalizations can be made. That Enlightenment kick is not something the notion of types, pervasive in theology and philosophy for millennia, needed.

It is not correct to think in terms of types in any historical, and thus irreducible, science; knowing all chemistry is not sufficient to understand biology. Biology has to obey the rules of chemistry, but there is a real other layer of interactions there, one with an historical and contingent component, that you can't reduce to the chemical reactions.

People are biology. Reductive arguments of type, such as any use of race as a concept, are wrong not just by results but for being tried; it's completely illegitimate as a mechanism for getting an accurate result. (As a mechanism for social control or maintaining privilege, now, that's a different question. Works well for that. Hopefully you don't want to argue that those are legitimate goals.)

What we do have is populations; with lots of work, observation and very careful statistics, some statistical generalization about populations is possible, but there are two things that are really important to remember.

One is that a population is not a static thing; it changes. So even the very best effort will produce a statistical abstraction bound to a single specific time and place. Two is that applying statistics on a scale of individuals is futile. (Trivial example—the average adult mass of elephant seals falls between the range of adult female and adult male masses; juvenile males pass through it briefly. If you don't know that elephant seals are really pronouncedly sexually dimorphic, you might try to use that average adult mass for something, and come to grief. Never mind that it's generally statistically invalid and you don't necessarily know what population an individual actually belongs to...)

It's not enough to make the—very considerable and awkward; those practising the science of biology, where population thinking was introduced as a thing obviously more correct starting in about 1940, still have struggles with it—change from thinking in types to thinking in terms of populations, where every new individual is necessarily an unknown; it's necessary to get the notion of types out of rhetoric and planning.

That's one of those in principle easy things; substitute objectives and statistical models and expectations, and argue on that basis. It's also one of the better-but-harder things, though; someone who is losing that argument is always free to invoke types again. (This is a common pattern; I could wish it could be treated as being every bit as complete an admission of loss as ad hominem attacks in formal debate.)


It's about different axioms. Axioms of incompleteness (I never have all the data), axioms of the inapplicability of generalization (knowing something about a person doesn't mean I know anything about other individual people), axioms of incompetence (to a first approximation, everyone is better at everything than I am[1]), and axioms of insignificance (I'm not special. (Neither are you, but it's much much more important that I believe I'm not special than that I believe you're not.)).

Most importantly, it's about an axiom of positive expectation; of course any one individual person is of peaceable intent, is worthy of polite attention, and does not have to conduct themselves in an attitude of deference. (If you're expecting someone to defer, you're making negative assumptions about their worth. If they, or you, elect to defer, that's a situational judgement, but expecting deference is, at very very best, being oblivious to hierarchy, which is bad, and at worst, an actively evil demand for surrender.)

That's hard, and is probably never going to be fully achieved by anybody; it's really, really tough to treat the fourth member of an identifiable group, the first three members of which gifted you with trauma, as a their own distinct individual self. I certainly don't claim I can always do this, or even that I always remember to try.

Trying makes the trend better. Arguing about types never can, because types are wrong to start with; it's not an issue of attaching the wrong thing to the type, as if there's some way to get it right; it's just broken in its beginnings. Spreading that idea can start to provide a quantified, factual basis for dealing with policy issues as populations and people issues as individuals, with strong arguments for why you have to do it that way to get it right. (Individuals are always individuals, never representatives of a population in this model; in a types model that includes race, individuals are always a representative of their type, which is ... unhelpful.)

So, anyway; I think those axioms are important. I think doing the wretched awful slog to stop thinking in terms of types about anything to which types do not apply—and that's at least everything alive—is important. So is listening, and remembering that to a first approximation, everybody is smarter, more experienced, more knowledgeable, and more successful than you are about something, and you have no idea what that thing (or things, or host of things) might be.

[1] Yes, really; of course you're good at some things, but there are a whole lot of things and a whole lot of people, and the odds of being the very best at even one narrow specific thing are terrible. It's like noticing that to a first approximation, every species is extinct.

12 March 2009

Feline indecision

Aoife sometimes has serious choice-management difficulties between locative territoriality (all this bed is belong to me!), disapproval (the Wrong Monkey! ON the bed! reading!), and glaring at the camera.
Every now and again I get a picture out of it.

10 March 2009

Rotational Transformation

Shadow of the Feline:

All this bed is belong to me:
I don't know why it didn't occur to me earlier that I'd get better contrast, taking pictures of my stealth cat on a dark blue and white background, if I processed them as black and white.
It didn't occur to me this time, either; ufraw remembered what the settings were, from the squirrel, and presented me with the first one in black and white, which made me go "hmm". Framing of the top one could have been a bit to the left, but oh well. Not a bad likeness and she has some intimation of contrast this way.

09 March 2009

Spring is not yet

Well above freezing with dry sidewalks and a sunny-ish sky with some scudding cloud by the time I left work today, yes, but it was also snowing when I got up this morning.
It being not a proper snowfall unless there's a manically overfocused squirrel doing its best to suck peanuts out of the—fairly heavy guage steel—peanut feeder by sheer force of will, we may here observe the use of the ploofy tail as a drag device. (Possibly to keep the feeder from rotating in the roughly 15m/s wind.)
This is black and white processing practise; I can hope it's not appalling for tonal value on another screen. (Reading the PDML has started to make me very aware of that, and to faunch after much better monitors.)

08 March 2009

The Towels Are Next Door

Not the same sparrow, but much the same idea, just a little later in the bathing process.
This one is through the glass doors, and I paid for it in blur.

In the Bath

The back yard is rather damp this time of year.
I try very hard not to think about how badly the wee sparrow must want a bath to be willing to bathe in icy icemelt.
Rokinon 800mm hand-held through the open door; I might be improving on the 1 in 15 usable, too.

07 March 2009

Work-Life imbalance

I have picked up some contract XSL programming work (well, mostly XSL programming; some ant and some perl to go with; it's the output generation stack for a DITA XML CMS). This is back where I used to work, but not for that company directly; I'm working for, and being paid by, their CMS vendor.

Two weeks into this job, I am, rather embarrassingly, at least two weeks ahead of my own schedule. I was apparently way too pessimistic when I came up with time estimates. (On the plus side, as evidence that I wasn't just whining when complaining of having my productivity hammered by thrash between the roughly three jobs I had at this place before, this might be described as being at least suggestive.)

Stupid bugs 1 and 2 are dead; a new merge process for PDF output (it's DITA and topic based authoring; if you want PDF, you need to turn everything into one big file; this process is called merging) that supports the DITA 1.1 spec's better/cooler maps (map=hierarchy of topics); the code for top level indexing is done, and I have a two column flow to stick this into, too. Second level indexing, attaching the index to the table of contents, and a bunch of testing to go, but I ought to be able to get that done in a week, even giving up a day to get the existing fixes on the production server, so it looks like I'm going to stay ahead of schedule.

So on the one hand, yay!, productivity; I like getting things done and feeling less incompetent than usual in consequence. On the other hand, observe the nigh-total collapse of blog posting. It's gone along with really minimal photography and a general collapse in the ocaml self-study.

This leads me to conclude that I pretty much suck at maintaining a balance between what I want to do, and what I'm being paid to do. With any luck I can now figure out why this might be so in a way that lets me do something useful about it, and yes, I have noticed that the sense of being less incompetent is probably a significant factor.

Stupid bug 1, as a footnote:
<template match="*[contains(@class,'+ topic/keyword')]">
is the opening tag of an XSL template element that matches all the specialized keywords in the DITA class hiearchy. One wishes to match like this because it's more efficient in a processing speed sense to match "specialized keyword" and then figure out which one and what should be done about it. ("Shorter XPath is faster XPath" being the general rule, along with "simpler XPath is faster XPath").

Once this template has matched, whatever element this is (wintitle, cmdname...) is the context node. Once it's the context node, element matches are tests for its children, not it.

So *[contains(@class,'sw-d/cmdname')] within this template, to find the command name element, doesn't work; cmdname isn't a child of any keyword-derived element, and certainly not cmdname itself!

contains(@class,'sw-d/cmdname') does work, because it's a match on the context node.

The syntactic difference can be damnably hard to see as part of code inspection.

Not much whistling

I suppose this constitutes behavioural plasticity in the probe-feeding starling. The whole local flock doesn't do this (or the contents of the peanut feeder would vanish in an afternoon) but at least three or four of them do.
Perhaps the other starlings feel they have a perverse taste for peanuts.
And yes, wretched invasive bird, but pretty, and doing a very good job of the theropod ancestry demonstration in the bargain.

03 March 2009

A pugnacious air

I am left with the impression that this fellow ought to have a comb in his back pocket, somehow.

Ground cuckoo; I rather like how the body contour feathers came out.

01 March 2009

Emerging from the darkness

Pretty close to what it actually looks like.
She does love her sunbeam. This second one is brighter than it really was in the hopes of leaving the evidence of good circulation in her ear evident.

These may constitute some slight evidence that I'm learning something; use of histograms to avoid clipping highlights and shot in TAv mode, rather than letting the camera decide absolutely everything.

For her part, Aoife has stopped paying any mind to the camera shutter sound. Makes getting snoozle-kitten pictures much simpler.