A portion of City: Urbanism and Its End by Douglas Rae bears retelling.

Rae argues that cities solved a coordination problem, but only for about a century bringing together energy, industry, transportation and labor.  That century started around 1850 and ended around 1950.  Everything changed.  The car and truck displaced the railroad, so factories could be sited outside of the urban core.  Electricity displaced coal so that you didn’t need to be sited near flat water to get your energy supply.  Telecommunications displaced physical presence so the need to be collocated declined.  Television and the melting pot displaced the motivations that held sustained a rich fauna of civic organizations.  Industry consolidation – into national rather than local firms – broke the common cause between civic and commercial interests.  Hobbled city government, an American tradition that cities are particularly crippled by since they exist only at the pleasure of their state governments, meant cities could not effectively react to those changes.  He makes a very convincing case for all of that.

But the story that bears repeating is about the black migration out of the South and into the Northern cities.  It’s timing could not have been worse!  He calls it horrible.  Just after the cities had begun their decline the migration picked up.  They moved toward the economically vitality centers, but it was too late.  Those centers were already imploding.  What that means, among other things, is the usual story of white flight is misleading.  Urbanism was evaporating and the black migration moved into those urban centers just after the trend had become unstoppable.  They moved into what was soon to be an increasingly empty carcass, abandoned.

Sort of like: buying real estate at the peak of the recent bubble, moving to Boston just as the mini-computer industry lost it’s way, or moving to Montreal just after the 2nd world war.  Timing is really hard.

Apparently since last summer Apple has changed how my olde MacBook Air deals with the heat.  Last summer the machine would get hot and then it would slow down, disabling one of it’s processor cores.  I’d observer that in my Menu Meters where on processor would go idle.  This summer they are doing something different and the Menu Meters report that both processors are running at 50%; which they aren’t.  Rather the operating system has decided to slow them down.  If you look at the process tables it shows a kernel_task eating all those cycles.  So, it’s out with the cold packs again.

MasterCard announced it’s 2nd quarter results.  Year over year revenue is up 26.4% in spite of purchase volume being down, at least in the US, 8.8%.  In other news one of their many presidents editorializes about why  interchange fees are good.

I have always liked that one of the oldest creatures on the earth, the blue-green  algae, is easy to find.  It is the planet’s bathtub ring, appearing in an infinite band along the rocks of the ocean’s tidal zones.  I like that combination of esoteric knowledge, and things turned up to their max: maximally old, and maximum circumference.  In my head the blue green algae bathtub ring is filed somewhere near the San Francisco cable car system – which gets high marks for largest flat continuously running mechanical device of some kind.

But I’m thinking that this article in today’s New York Times by the ever entertaining Carl Zimmer may trump those.  Apparently bodies of water have, living on their surface, a thin film of stuff.  Some sort of vast membrane.

…the top hundredth-inch of the ocean is somewhat like a sheet of jelly. And this odd habitat, thinner than a human hair, is home to an unusual menagerie of microbes. “It’s really a distinct ecosystem of its own,” …

And obviously for some definition of big this is the biggest ecosystem on the planet.

The discovery of this new world leads to lots of questions.  How does it effect the waves?  How does it effect the movement of gas, carbon dioxide for example, across the boundary.  How does effect the movement of heat?  What role did this play history of life.

It with some pride and concern that my son’s first bemused question upon having this revealed to him was – “Have we figured out how to kill it?”

Adverse Selection is the name for a common syndrome in markets where “market participation is a negative signal.”  For example: life insurance office, enters the husband and he announces: “Quick, I need a million dollar life insurance policy on my wife; by 6:45 this evening!”  The agent thinks: “Yeah, quick commission!  Ka-Ching!”  Off stage the insurance company notes both agent and husband’s particpation in this insurance market are signalling a negative.  Adverse selection.

Markets can be structured to encourage adverse selection.  When the guy in the plaid suit swoops down on you in the parking lot most of us think “oh dear, here comes the salesman” but the sophisticated observer of markets thinks: “adverse selection.”  When a car company announces that their sales people aren’t paid a commission they are trying to signal the absence of this problem.

When the mortgage industry rejiggered their risk management architecture they created a market with an abundance of adverse selection.  The mortgage brokers were encouraged to ask few questions while gathering their commissions.  Mortgage buyers where invited to murder their financial lives.  Apparently the entire hierarchy of this financially innovative market encouraged adverse selection and everybody’s participation was a negative signal.

Insurance contract often have clauses to temper the problem of adverse of selection.  If you die shortly after buying a life insurance contract they will take a close look at that clause.  The preexisting condition clauses in health insurance contracts are mutant versions of these.  Horrible horrible stories are common, insurance companies abusing these clauses to claw their way out of their contracts.  But notice how any market with even a wiff of the taint of adverse selection problem suffers another problem.

If you go to bar, or sign up for a dating site, or even sign up for a course at night school your entering a market for new relationships.  Such markets are riff with adverse selection problems.  So by walking thru the door you immediately become suspect.  Your participation in the market is a negative signal?  Who are these losers?  In the healthcare debate the reluctance of optimistic healthy people to participate in the market would seem to have a bit of that.  I.e. they aren’t just suffering from a naive misunderstanding about time, they are also reluctant to hang out with all those sick losers.  It’s not just “I’m healthy I don’t need insurance, it’s I don’t want to be associated with that kind of people.”

All the markets for creating new relationship have serious adverse selection problems; e.g. consulting, sales, hiring.  The problems are greater the longer term the relationship is going to be.

Hiring is a great example.  If you list an opening anybody who applies is immediately suspect; particularly if they are unemployed.  Their participation in your hiring process is a negative signal.  This signal is probably more accurate when the economy is strong and applicants are few, but ironically when the economy is weak and the applicants are numerous then the need for a cheap rule of thumb for sorting the applicants increases.  Here’s a nice article (thanks Luda)  about this syndrome.  The HR or head hunter jargon for this problem – two words: actives and passives.  Anybody who is actively looking is immediately suspect.

This is the Groucho Marx story:  “I sent the club a wire stating, Please accept my resignation.  I don’t want to belong to any club that will accept me as a member.”

Since adverse selection taints relationship building markets you get a plethora of work arounds.  There is demand for faux passivity.  Eight percent of the folks at the evening drawing class maybe looking for a relationship, but the twenty percent who are there to learn a new skill provide a plausible cover story.  Most of the folks in the bar maybe looking for new relationships, but if they can get a gaggle of their existing friends to head out to the bar with them then they have a good cover story.  No doubt with a little effort you can think of lots of activities that include in their value proposition a dose of faux intended to treat the problem of adverse selection in relationship building.

The problem is perfectly symmetric.  These days you see lots of job listings that read along these lines: “paid opportunity!” or “grow into in-house and salaried positions” or “great resume experience”.  Once you start thinking that every posting is a negative signal about the company in question it really changes the way you read the postings.  And amazingly you can almost always see why this one is a looser.  For example doesn’t this: “report directly to the _ and _ and work alongside _” raise a bit of concern?  I suspect they mentioned that in the posting because it’s a problem and hence it is – more than you know – part of the job description.  I worked for one large institution that had a policy that they would absolutely never pay for job listings – I now think that reflected their deep seated desire to never look so desperate as to be actively seeking a relationship.

Middlemen provide one way to tackle the adverse selection problem.  If I need to fill a position I’d rather, given the above, avoid the job posting markets.  So I go to my social networks, or I go to a professional headhunter.  I find that fascinating.  I’d noticed before how the middleman is two-faced – offering one face to each side of the transaction he is intermediating.  But this high lights another kind of duality in the middleman’s role.  He is at one and the same time offering a service that tempers the market failures due to adverse selection while at the same time his incentive is that ka-ching of closing the deal.

Adam Wiggins of Heroku has spun up a fire and forget charting service.  I’m using it here.  That chart shows the wait to get a license at the Watertown Dept. of Motor Vehicles here in Massachusetts.  They have a take a number queuing system that posts the wait times for all their services on the web.  The charting is rough around the edges.  The wait times are in minutes, and the chart along the bottom is more readable than the main chart.  The main chart apparently plots only a handful of random points as soon as more than a few points are in the range.  Playing with the sliders on the bottom might help.  If I was in more control I’d plot only the points and not the line.  Here’s my Ruby code for doing the polling.  I have a crontab that polls regularly when the office is open.

So far the answer is, show up first thing in the morning, when they open.

Most of my music collection is on a Macintosh in my basement, that machine runs iTunes so everybody on the local network can play that music on their macs. So far so good.

I also have an ok set of speakers wired up the Mac in the basement. So if I play music on that machine it sounds ok in the living room. I happen to do that by running MPD (or Music Player Deamon) on the basement Mac. MPD has numerous remote control mechanisms. Sitting in the living room we use Theramin on our laptops to control it. You can teach MPD about internet audio steams, so that’s how we listen to the radio.

One scenario that’s eluded me until today arises when we watch videos. This blog posting explains how to get your mac laptop’s audio routed to another Mac, like the one in my basement that drives the good speakers. It is pretty complex. First you install SoundFlower so you have some new output devices in your Sound Preferences; you then send your audio to one of those. Second you run the program /Developer/Applications/Audio/AU Lab. You’ll only have that if you’ve installed the developer tools. You’ll run that on both the sending and receiving machine. You’ll create an AU Lab document on each, to route the sound appropriately. You can save those two documents (the sending and receiving one) for reuse. You can have AU Lab running all the time on the receiving machine, so it will pipe the sound as soon as you fire up the right configuration on the sending side. The details are in that posting.  Sadly the time delay for all that is sufficent that the Movie’s audio looses synch; maybe QT Synch can help with that; just how complex can this get?

This can all get wonderfully perverse.  When the sound get’s too loud or soft there are way too many links in the chain that might be at fault.  But like running simulators inside of virtual machines the geeky fun is to make it even more confusing.

So.

PlayDar is reinvention of the ideas around OpenURLs, i.e. you ask your local playdar daemon if it can find some music, given assorted metadata (title, artist, …) and it attempts to find it for you.  It might find it on your machine, but it can also search other places.  For example it might find it in some service you have subscribed to, or on in a friend who’s collection you have arranged to permission to share.  Playdar’s pretty new, and pretty rough right now.

But at the moment there is a web page on showing on my laptop, and it’s an mp3 file which my local playdar daemon found for me.  That file happened be found on the machine in my basement, which is also running a playdar daemon.  So the file is being shipped from the basement playdar daemon to my local playdar daemon which is serving it over http to my web browser.  The web browser is playing the file.  My laptop’s audio is currently routed to an output device provided by SoundFlower.  AU Lab, running on laptop and it’s listening to that SoundFlower audio channel.  AU Lab is then “filtering” that audio, sending to an AUNetSend filter.  Meanwhile back down in my basement another instance of AU Lab is running.  It’s taking it’s input from a sound generator that happens pull the sound off the network via AUNetReceive.  It then pipes the sound to the speakers on that machine.  Those speakers are up here in the living room.  Isn’t that just disgusting!  I’m such a geek.

Sigh, I really ought to find a job.

Try to collect some thoughts about Chrome et. al.  I really shouldn’t be writing this.  I need to find a job!  But I’m hoping that if I write this I can stop thinking about it. That usually works.

First, a few of other people’s postings about it all, one’s I’ve liked:  John  Gruber’s on  Daring Fireball, Anil Dash’s post, and to a lesser degree Matt Cutt’s reaction.  I’m surprised, I haven’t read anything so far that drags out front and center the security and performance aspects of ChromeOS, more on that below.  The whole reframing of this as about the Microsoft/Google rivalry bores me, it’s thought stopping.

One thing I noticed about Google’s Wave was the absense of any other players at the announcement.  No other significant players on the stage or in the PR to give even the hint that their attempt to create a new standard actually had a diverse community of participants getting on board the bandwagon.  That leaves you with the strong scent of “tail light chasing.”

Android has substantial 3rd party partners.  And in fact I think you can argue that Android product team has put a lot of labor into making the offering palatable to various important 3rd party constituencies.  While striking a balance between the handset makers, the telecom companies, and the developer communities isn’t easy the market and technical trends appear it plausible that something quite disruptive (involving open) will play out successfully.  The Android effort, it seems to me, is innovative at the scale of market shaping, much more so than at the scale of the tech.

Android stands on a long long history of failed attempts to create useful handheld devices.  Conventional wisdom has predicted a handheld revolution since maybe the mid-eighties; and an awe inspiring lot of money has been spent attempting to create products in this space.  I think the successes (Palm OS, the black berry, iPhone, etc.) tend to blind observers to how risky and hard this is.  So on the one hand that makes me a bit sanguine about Android and on the other, if I was leading that effort, I’d be reluctant to license much technical innovation.  The user experience for these things is unimaginably hard to get right.

Which is all to say that Google’s efforts to shape the cell phone (aka handheld + telco) industry in their preferred direction look like they are going reasonably well.  Plenty of risk, but roughly they appear to have gotten the balance  right.  I assume their goals here are pretty transparent, they would prefer to create open standards that move that market, which is adjacent to theirs, closer to commodity pricing while assuring that the standards/platforms that emerge to enable create good ecologies for complements for their offerings.

That effort is reasonably critical, but it can’t hold a candle to assuring that the web browser market evolves in what I’m comfortable calling the right way.  While assuring that the telco markets evolve in desirable ways is extremely important for Google assuring that the browser ecology moves forward is a matter of life and death.  The risks here are pretty extreme.  For example if the browser becomes a rat’s nest of security issues then it’s only a matter of time before some black swan drops by and scares off a large population of users for some period of time and Google’s growth stumbles or worse.  If you look at Google impressive efforts to deliver applications with rich user experience via the browser (email, maps, docs, …) from my stand point, i.e. a guy that worked on desk top user interface tool kits for two decades, what’s striking is the lack of head room.  Those applications are topped out, slamming their heads against the roof of what’s possible in the current Browsers.  Which is why on the standards side they are so deep into HTML 5.

I was originally quite saddened to see Google building Chrome.  At the time I viewed it as a kind fork likely to undermine the vitality of both Webkit and Mozilla.  Forking is often quite problematic.  Maintaining common cause is hard and forking makes it harder.  Sometimes it’s necessary, sometimes you have to take a discontinuity; break rather than migrate the existing installed base.  And, installed base can mean a lot of things including the presumptions of the developers currently gathered around the existing code base.

Anyhow, two things changed my mind about Chrome; both Javascript related: i.e. performance and security.  Both are ripe.  Not easy, but the demand is huge, the and the benefits are substantial.  To allocate substantial resources; sums larger than VC money, sums the scale that only a firm like Microsoft or Google might bring to bear you need one more thing.  You need assurances that the problem won’t be routed around before you solve the problem.  You get that when you have an immovable installed base.  I like to joke, sardonically, that if your going to learn a new programming language the one to learn is Javascript because there is no way anybody is displacing it anytime soon.  Notice how addressing performance raises the room while fixing the security problem attempts to tempter the risk of that black swan’s arrival.

Browser headroom and security are problems (risks) for Google, but we all share them.  Reasonably people can argue about the scale of these problem; but I tend to think they are huge and hard.  Which helps to motivate whey Google would tackle with a strategy that is less open and more in house than one might like.  Tackling problems like that in a fully open strategy is harder than doing it in a closed one.  That they took the tack of a semi-open approach is, I think, more about the long term requirement that it be open and less about tempering the distaste at what I presume they saw to be the short term requirement to circle the wagons and tackle these hard problems.

Notice at this point how the Chrome efforts (with security and performance at their design center) are entirely disjoint from the Android efforts (with balancing market shaping goals and hard UI design risks).  To the degree that I’m reading these tea leaves correctly that helps to motivate why the two efforts are largely disjoint.

I notice that like Android, and unlike Wave, ChromeOS product team has brought 3rd party players to the party.  A number of hardware vendors have allowed Google to engage in name dropping.  I wonder if we will see something analogous to Andoid’s Open Handset Alliance?  That kind of thing  de rigueur in the telco industry.  That’s because the power in the telco industry has always been broken more equitably due to the national franchises.  Those franchises have historically been safe.  Each nation’s big phone companies was happy to collaborate on setting standards that lowered their costs and kept their suppliers on an appropriate leash.

The PC industry is a whole different kettle of fish.  Microsoft and Intel have lead standards making in the PC industry, or entirely by virtue of spontaneous emergence (or bottom up) ala Linux and the Web standards.  Gates once complained that Linux exists only by virtue of the commodity hardware he labored so long and hard to assure.  There is some truth to that; and I’ll point out as an aside that Android is an effort to create a similar world of commodity handheld hardware.  So there is a thread here I don’t entirely understand about that.  I don’t really understand how the PC industry reaches agreements about things like what is standard hardware.

Which brings us around to the question of what is an OS?  To me an Operating System is an intentionally designed bridge between hardware and applications.  The goal of that design is to create a space of options, on either side.  Two play grounds (often called platforms) for hardware and software developers (often called firms).  Which raises the obvious question what the hell does Google think it is doing with this ChromeOS?  Sure, I have made the argument that ChromeOS is tackling security and performance so as to address risk and create added headroom.  But, that ain’t a cool new pair of play grounds.  I’m having trouble seeing how the hardware playground created by ChromeOS is substantially different from that we see around netbooks.  If so then that’s a lost opportunity!  The space created on the software side is large, but it’s seems sort of dull – just hill climbing.  These are early days.  We really no nothing, but it seems to me like ChromeOS isn’t distinct enough in the opportunities it is creating for the two key developer communities.  But ask your self, if you had deep pockets would you fund Angel fund a group of clever guys who’s goal was to build something either run on ChromeOS or to run ChromeOS?

I’m a bit conflicted about that conclusion.  I don’t seem to feel that way about Android, and I don’t entirely know why.  Maybe  it’s because Android’s two playgrounds have radically different terms of trade?  Hm.

I find it curious that none of the above observes this thru the lens of what I think is the important shift in the landscape.  I describe this shift by saying that the distribution of new transistors is shifting.  Each year we make N units of new transistors.  Some go into the cloud, a larger are larger share.  Some go into desk tops; and I think that slice of the pie is shrinking (though to be clear the total #/year might still be growing).  And some are going into laptops, handhelds, and cellphones.  That trio’s share is also swelling.  At the same time these devices are less and less autonomous.  They are more entangled, the cloud with the devices, the devices with the cloud.  This increasingly entangled world of devices that are polar opposites in scale is so entirely different than the world of personal computing that I think it requires a kind of brain wash to appreciate it.

That new world is the world that Google thrives in.  Projects like Android, ChromeOS, are both about one side of that bi-polar world.  Their application offerings are more about the entangled nature of this, particularly Wave and Docs.  So any story we tell about ChromeOS needs to be stitched into a narrative about who that new world is going to shake out.  I don’t think I’ve done that here.

Well, that was a morning wasted.

Via Planet Apache Assaf Arkin‘s post lead me to Matt Webb’s delightful talk about design.  Which notes in passing a project to treat the newspaper’s printing plant as a kind of cloud computing accessory.  I’m fascinated by the idea of plugging unique peripherals into cloud computers, or maybe maybe just the network OS.  But it’s particularly delightful to see old media repurposed this way.

at Vimeo

Jane Jacobs is the first place I encountered the  idea repurposing  idle capital equipment during the off hours.  Booking the local school auditorium for commercial concerts.  Letting the library out for weddings.  Letting your firm’s idle cube space out as office space to for small businesses. I used to think about that when ever I gazed out the back window at the idle climbing structure.  These are just coordination problems, and that’s one of the things the net gives us new tools to tackle.

I’m increasingly convinced that the cloud computers are really all about enabling the handling of impulses.  Enabling new kinds of systems that survive entirely because they can capture the value in the rare moments when they are slashdotted.  Mast years as a business model. Why can’t I have the impulse to print up a 1000 copies of a little newspaper and have it delivered to my door tomorrow by the Boston Globe’s printing and distribution infrastructure?

Skyhook is company here in Boston that sells geolocation services based off a database of Wifi locations.  For example, if your walking up my street your phone’s wifi reciever might notice one of my wifi hubs, it then queries the database to find out where you are.  Skyhook isn’t the only organization with a wifi device to location database, it’s just a commercial one.

The original such databases were assembled by volunteers.  The so called War Drivers, who’d drive around town with a GPS and Wifi reciever hooked up to some software.  There are good open or public databases for this kind of thing.  For example WiGLE.

Skyhook sells access to it’s database to folks in the PDA/Cellphone business.  So, presumably they consider it a valuable asset.    I’d assume they seeded their data from the public databases.  I’d love to read a case study of how Skyhook and the public community has managed flow of data between their database and the public ones.  That would be an interesting legal, licensing, ethical, standards story; with a nice subplot about how public goods are taken private by commercial entities and visa versa.  But all that’s not what this posting is about.

I was reading this bit of Skyhook press where in the Skyhook people are pushing the benefits of Wifi v.s. GPS and the CEO happens to say:

Skyhook employees and contractors “wardrive” down millions of miles of roads to correlate location (from GPS) with the signatures of Wi-Fi access points. Morgan said that Wi-Fi beacons are unique even when security is turned on, so that’s not a factor.

Another way that Skyhook keeps its data current is by using the information it gleans from Skyhook users. That’s right: When your iPhone geolocates itself it also sends Wi-Fi beacon data back to Skyhook, which helps keep the system’s location database current.

Ok, that means is that anybody who’s application has a large installed base on the iPhone (or other PDA/Phones) can draft that installed base as war drivers.  I don’t doubt that the Skyhook database is better than the open/public ones.  Given the above, I doubt that will be the case for long.  We are all war drivers now.