Water in Boston

The big (10 foot) pipe that brings Boston it’s water has suffered a break and we have all been instructed to boil our water.  I see on the MWRA web site that they  deliver  167 million gallons of water to 2.5 million residents a day.  That comes out to 66.8 gallons per person per day.  Of course they also deliver water to industry … and more typical leaks.

The paper says that the leak was dumping 8 million gallons an hour into the river.  167 gallons/day is equivalent to about 7 million gallons an hour.

This water is drawn from some resevours in the middle of the state.  The largest of which holds 417 billion  gallons, or 6.7 years to draw down at 167 million a day.

Problem: Immovable installed base -> Solution: closed system?

The ongoing puzzle, debate even, about the choice points between a closed and open system appears to have picked up a new aspect.  At least I had not noted this one before.  What differentiates a closed model from an open one is the extent that the business hordes the options created by it’s product offerings.  For the firm the open v.s. closed debate comes down to questions about short and long term advantages for the firm and it’s managers.  A closed approach hold out the hope of better price  discrimination.  While an open approach offers a change at wider adoption, more innovation, stronger network effects, and better protection against competitors.

So, the new  argument  I’d not previously seen goes as follows.  It is on the closed side.  A firm can adopt a closed architecture to deal with the rapid technology change.  Closed addresses the immovable installed base problem.  By making closing things down the firm captures the  ability  to migrate the installed base more quickly.  Or, so this story goes.

When dealing with the risk of technology driven displacement a large installed base is both a blessing and a curse.  It’s a curse, of course, because it is slow to move.  But it is a blessing since there is safety in numbers, e.g. an assurance that when technology changes you’ll not be left out in the cold.  The suppliers and users will have plenty of common cause to help solve to manage the switching costs.  In a closed system the vendor coordinates that.  You could argue that in a closed system the coordination problem is easier to solve.  I suspect that’s not so clear in practice.  The installed base ties it’s fate to the competence of the vendor.  The vendor is a single point of failure.

Open makes firms nervous.  Management is all about juggling risks.  Closed offers the impression your in control.  This impression is often wrong.  Closed systems lead to the systemic blindness; provincialism.  And yet I find the  argument  that a closed system is better able to handle rapid technology change sounding pretty good.  Possibly that’s because it’s about supply side risks.  I wonder why I think the closed system blindness issue is less serious on the supply side v.s. the demand side.

Having engaged with this debate for twenty years, and now finding it kind of dull, it’s a delight to see a new aspect to it.

Problem/Solution

I’ve not read Stephen Prothero’s recent book “God is Not One.”  But, listening to him interviewed last night I was much attracted to his list. What each of the eight major world religions treat as their big problem, and what their solution is.

Religion Problem Solution
Buddhism suffering awakening
Christianity sin salvation
Confucianism chaos social order
Daoism conformity naturalness, simplicity
Hinduism endless cycle of reincarnation release
Islam pride submission
Judaism exile return to God
Yoruba disconnection follow our destiny as revealed by diviners

It is likely I’ve made mistakes in the above, it’s based on some web browsing and what I recall from the interview I listened to. More rows and columns would be fun. For example principle rituals would make a great additional column.

I am a sucker for frameworks like this.  That is a variation on one of my favorites.  This one about sketching out the differences between Puritans and Quakers –  a dialectic that has much to says much about American culture.

It would be fun to have a table like that for programming languages.

Facebook PAIN

In the best scenario all Facebook is doing with their new service that allows 3rd party sites access to your Facebook identity along with a bit of what they know about your is a slightly more transparent version of what, say, Google’s Doubleclick can does. They are selling a service to their partners that identifies the visitor. It removing their anonymity. These tracking networks are troubling from a privacy point of view since they enable trafficking in a surprisingly detailed user profile. For example it enables knowing that your currently working, or shopping, or away from home. Browsing Yelp for a resturant? Working on your a Microsoft document.

We need a name for these networks that enable the trafficking in personal data. How about persona-attribute-info-network or PAIN. There are other PAINs. The credit reporting in the financial industry for example. There are ones in health insurance. There is usually one for every kind of license, i.e. drivers licenses. In the long run, i.e. after fortunes are made and I’m dead, these tend to get a complementary “privacy protection act” that serve to limit the liability of the network owners, raise barriers to entry, and add a modicum of consumer protection.

A key term or art here is “globally unique identifier,” GUI. The social security number is the poster child of a GUI. Leaking a social security number bad for two very discrete reasons. The first is it’s role as a password, but ignore that. The more serious concern is how it is a dependable key that vendors can use to unique identify you. Once a GUI tags your account data the vendor can then trade the data in that account with each other. You licensed them to trade when you assented to their “privacy policy.” I like to joke that they do not lie when they say “your privacy is important to us.” Well yeah, it’s an asset that it is important that they leverage.

GUI come in variations of quality – Social security numbers, email addresses, open ids are all pretty high quality. Cookies are actually pretty good. Google’s Doubleclick cookies can be very high quality. What your are licensing when you leave the Facebook toggle on is tagging you with a high quality GUI owned by Facebook.

A PAIN will have rules that govern the exchange of data between members. And all the usual questions arise. What are the costs, benefits, and risks of membership. Who sets the rules? I think we can assume that Facebook has not bound the members to limit data exchange laterally, i.e. Yelp and Microsoft can traffic in info about you using the facebook GUI as a key. At that point do we care what info Facebook shares with them?

Now, mind you that was all written wearing a care-about-privacy hat. There are other hats!

How about the were-things-are-going hat. It’s obvious that reach, accuracy, and tracking skill of the PAINs is only going to continue to grow. Scenarios long imagined, like enabling the car rental agency to prefill the forms based on your recent airline ticket purchase – a behavior that it pretty trivial to enable, but spooks the user if he hasn’t been carefully preped to comprehend how it happened – are inevitable.

Put on the business-strategy hat; the puzzle is who owns the PAIN that enables the scenarios like, will they make a good landlord, how many such networks will exist, should you try to establish one. The business-tactics hat depend on the answers to those questions. But moving fast maybe necessary or it could be fatal.

Having written that, I think I have a brilliant solution … but putting it here on the end … well it really doesn’t fit.

White light sighted

Many many years ago I read in some high end Optics Society journal about what I came to think of as “the white light.” The article argued that if one extrapolated the patterns in communication there would come a time when the cost of routing information around the network vs. the cost of broadcasting everything everywere would cross and at that point the end points on the net would see everything and just pluck out just the bits intended for them.

Well today I see an early example of just this idea. Here being used to provide optical interconnect for a multiprocessor.

Back in the 70s, when I last worked on multiprocessor interconnects, I sketched out something like this; but the response times on the optical  receivers  were too slow and the complexity of the rest of the supporting junk was way high.  I wonder if you could hack a modern digital camera’s control electronics along these lines in some fun way.  This kind of thing would be a good project for some MIT 6.111 students.

The Patriarchy’s Pollen Preference

Paul Krugman is much amused by the realization that the  origin of Pigouvian taxes(i.e. the idea that the temptation to dump your pollution into the community water supply  might be tempered by a tax on such behavior) never actually mentions pollution but rather only mentions the problem where in one landowner’s enthusiasm for hunting causes a surplus of rabbits to spill out damaging his neighbor’s gardens.

Here’s another fun example of an externality.  Apparently landscape designers have a preference for male plants over female.  The female plants drop more litter, i.e. seed pods and fruit.  Somebody has to sweep up the mess, the clients complain, the landscape designers respond.  While the male plants just pump out vast clouds of  pollen.  Which triggers more allergic reactions from the general public.  But the general public isn’t paying the landscape designers, now are they?

I guess there are alternative schemes to address this.

  • Ignore it.
  • Follow well worn path laid out by the tobacco industry, fast food, and anti-enviromental forces; and mount a disinformation campaign.  (Hint: Be sure to mention freedom, choice and unintended consequences.)
  • Campaign to change the behavior of landscaping  profession.
  • Regulation
  • Tax Pollen
  • Pollen cap and trade.

I’m totally in love with the idea of pollen cap and trade, and I look forward to the fraudulent pollen credit story.

hat tip: Heebie-Geebie, and see also Bee Hive.

Blowing Bubbles to Watch ’em Pop

The Jesse Eisinger and Jake Bernstein of  ProPublica’s recently released work on the economic train wreck’s roots is  marvelous.  Listen to it on This American Life!

A quick overview of the story:  First recall that a key link in the supply chain of the subprime mortgage industry was the step where the risky mortgages where bundled and sorted.  The rational was to capture the benefits of diversification, while the sorting’s purpose was to allow investors to select the risk profile they  preferred.  Like all the other stages in the industry’s supply chain this part starts out as a sensible thing and over time it transitions into insane.  The story uncovered by Eisinger and Bernstein outlines a plausible story about how a  predator appeared on the scene and figured out how to enable the insanity to both continue and redouble.  Doing this in a pattern that allowed the  predator  to make huge profits when it all blew up.  Credit default swaps, i.e. bets that paid off when a loan defaults, provided a way for the  predator  play place a bet that would pay off if the house burnt down, and place those bets on houses he had built out of tinder.

An interesting question is then why did the builders willingly create these houses of tinder?  The  explanation  suggested in the piece is that the builder’s incentives caused, or at least encouraged, the shoddy workmanship.  Each time the bankers put together a new meta-toxic CDO built out of toxic mortgage bonds they got paid a fee; a large part of which flows right into their bonus pool.

But it seems to me there is another aspect to the story.  Industries, just like people, have a strong tendency to be consistent.  Behaviors are hard to change.  You have to discard your current expertise and start over creating entirely new skills – a very very  risky  proposition.  So there is strong incentive, largely rational, to attempt to continue to with current practice.  So it became increasingly obvious that the entire subprime mortgage industry supply chain was going insane; all the players in the chain were faced with exactly this puzzle.  Into that landscape arrives the  predator.  What he offers, sells, the existing players is the perfect product.  A solution to their principle pain point.  A way to keep on doing what they are already doing.

I’ve seen that a few times over my life.  Though never seen it with the added trick of piling on an adverse selection insurance fraud angle.

Moore’s law and his friends have regularly displaced existing, but deep pocketed, industries over the last few decades.  Each time the old, deep pocketed, players grow  increasingly uncomfortable.  So they look for ways to continue their current habits and they are willing to pay large sums, acquire foolish startups, and take significant risks to do so.  Sooner or later folks appear on the scene happy to help.  In most of these cases continuing in the old behavior is hopeless; but hope springs eternal eh?  Throw another layer of sandbags on the levee.

But really two things make the Eisinger and Bernstein story orders of magnitude more potent.  The first is the insurance fraud angle.  The second is the bubbling.  If you can play into this game just as the bubble begins to peak; encouraging the bubble to expand while betting on it popping.  That’s when the – ah – fun begins.

Originally Written?

I wonder why Apple included the word “originally” in “Applications must be originally written in Objective-C, C, C++, or JavaScript”?

If you are building a platform of unhackable devices then you need to control the gateways to hacking; i.e. the tool chain and the application distribution channel.  So the clause above’s purpose is to due just that; e.g. if you want to hack code to run here you gotta pass thru our  interpreters, our compilers.    But they wrote something stronger.  They wrote is that you have to write in language A,B, C, or D.  Only.    Your not allowed to write in language X, Y, or Z – even if you cross compile X, Y, or Z into A, B, C, or D before you deliver.  I don’t get it

For clarity: this means you may not write a program in say Python, and then run that program’s source text thru a Python to C translator before compiling it before submitting for Apple’s approval.

One of the things that lead me to Lisp was noticing how most of the programs I wrote weren’t in the language the compilers offered me, but instead we would design a custom language better suited to the task at had.  For example the rule system that drive a compiler’s optimizations, get some delicate thing to work on multiple systems, or the custom notations that generate the object system.  Once I accepted how central that is to software architecture, I wanted a language that respected and encourage that.

I’m not terribly surprised that Apple has decided to run the audacious experiment of creating a platform that strives for devices are no longer computers once in the hands of the end user.  Devices which the end users can not hack upon.  I think it’s kind of vile and it worries me that they may succeed, but it doesn’t surprise me.  I should write another post about how potent such a platform will be if they can sustain it.

They aren’t the first to go down this path.  It’s what the locked phones do.  This is what some of the game consoles do.  And I guess if you fully embrace my presumption that the right metaphor for modern websites (aka applications) is that they are massively multi-player games, then it’s only  inevitable  that game  industries  habits would move in the rest of the application stack.

By way of amusement this clause would appear not prevent you from doing a less interesting but still common architectural trick.  I.e. where you write in a dialect of language A, call it A’; and then cross compile from that into A before running thru their tool chain.  So if you can cast your need for a domain specific language into a dialect then you can sort of wiggle around the restriction.  To my surprise there is an important example of this, Caja.  Caja is a dialect of Javascript with some extremely desirable security features.  It’s used by Yahoo and Orkut (for example) to assure that third party widgets are tightly constrained in the damage or snooping they can do to other portions of the web page.

And I guess there is always C++; you can do some very wonky things at compile time in C++.

Sardine

The library of congress has an wonderful collection of photographs taken at sardine packing plants.  Thus I came to learn the word cartoner.  Which was once a person, but is now a machine.      Today comes news that the last such cannery in the US is shutting down, along with a few pictures.  This all resonates against a conversation my wife and I had yesterday about how maybe none of the high tech companies that were in Boston when we moved here in the late 70s still exist.