Heres something I did not know or expect.  Early in the book Piketty is laboring to help his readers to develop intuitions about income and wealth.   In the colonial era the European nations controlled a lot of productive assets in other nations.  The income from those contributed “nicely” to their national income.  Heck, that’s almost the definition of colonialism. These flows, obviously, weren’t balanced.   Income flowed from the colonies back to Europe.  Little flowed back.  Piketty would like his readers to know that’s not true anymore.

My entire life people have worried that some nation is are buying up America.  Which nation keeps changing though.  To me it has always seems only right that the former Colonial powers would live in fear that the tables might turn   We used to worry that the Japanese were buying up all the good real estate.  We worry now about the Chinese.   Sometimes we worry about oil rich nations.  I was unaware that the French are concerned that California pension funds.  .

There is lots of cross border ownership.  But on the whole both the income flows and the ownership tends to balance out on a pairwise basis.  I did not know that.  I would not have predicted that.  I am not sure I believe it.  I’m optimistic the book will say more.

I’m enjoying his efforts to build out my intuitions about this stuff.  Some examples…

FYI – National income isn’t the same as Gross Domestic Product.  GDP doesn’t include these the cross border income flows.  It also doesn’t include depreciation.  Which I gather we can roughly estimate as 10% a year.  I knew this only in the consequence that recessions are typically followed by an uptick in durable goods.

The pool of wealth is about 50% is residential real estate.

Economists have a rule of thumb that a third of the income flows back to capital as the return investments.   The rest to labor.  But that rule of thumb is likely bogus.  I’d be tempted to say it’s almost a political statement.  But Piketty is more polite, it is based on thin data.

The comment that cross border flows of income tend to balance out these days gives us license to demote that kind of income as a key part of the overall analysis.

What we consider to be wealth is a social construct.  Slaves used to be included.  And we know how the sports team owners feel about “free” agency.  Knowledge is fraught example these days.  It’s thought provoking how much these constructs can change over time.

I particularly liked his aside that when debating the shifting the consensus about what can be owned the agents of owners talk about efficiency, rather than self interest.  People still regularly suggest that slavery is more efficient than a life of poverty.  I’ll note that talk of innovation has starting to displace talk of efficiency.

Public wealth, in the hands of nation states, is close to zero.  Which means we can largely ignore public wealth when thinking about the total wealth of a nation.  He promises to say more about wealth in the hands of non-profits.

Founder reports successful IPO to his VCs.

Each year in Boston we have a road race, a marathon.  All I knew about marathons before I moved here I learned in grade school.  It’s bad news: you die at the end.   But in compensation you get remembered as a mythic hero.   Now that I live here I picked up some random knowledge.  For example it is important not to sprint out into an early lead early.   You gotta pace your self.   Hearing that again this year I thought: yeah! that’s not the advice we give to startups.

Instead we counsel that you gotta get momentum early.  I recall that Steve Jobs advised the Segway team that they needed a huge PR campaign to launch if they ever hoped to trigger the kind of change they imagined.  They didn’t take he advice, and they didn’t trigger the change.  So, see!

Information cascades are one of the many processes that generates power-law distributions. Kieran Healy reports on some fun research into this.  The researchers hacked the early days of a random sample activities at Kickstarter, Epinions, Wikipedia, and Change.org. They blessed some projects with in a small way and then waited to see if how much it helped.  It helped.

Kieran is a professional and his commentary is very astute.  Of course, if you’re interested in hacking, helping, defending systems like these then it’s very nice to have some experiments reported publicly.

Apparently this year’s Marathon winner did take an early lead.  There is a touching story about collective action that possibly explains why that worked for him.  One would assume a marathon is the archetype of an activity immune to collective action, but you’d be wrong.

The web works because it let’s you reach other people.  That contact is the motive force behind the entire net.  All the other drivers are complements, competitors, and parasites.

Which brings me to the semantic web.  The semantic web at best off by one, and at worse it’s entirely wandering off in the wrong direction.   Where web pages are crafted by people to titillate other people – so it’s no wonder they show up – the semantic web exists to empower machines to excite other machines.  It doesn’t make sense, so people don’t show up.  Well at least not much.

According to the study reported here 1% of all the web has a few crumbs of semantic markup.  In an industry where we expect ideas to start bonfires the best we can say for this idea is that it endures.

But on the other hand.

Maybe this is going to change.  And yeah, I’ve and others have said that before.  But the web isn’t the same as it once was.  We are long past the explosive phase transitioning growth and deep into the consolidation.   It’s less about people now, and more about the machines (though we call it big-data now).  There’s an entire industry (aka SEO), and has been for a while, that labors to make the web pages attractive to the machines rather than people.

So it is with interest that I read that report.  Which suggests that Google is starting to bless pages with semantic mark-up with better placement in search results.  That would be a classic standardization move on Google’s part.  If you have market power you can create incentives (force) suppliers to conform to standards in service of your quality/cost metrics.

I’ve predicted this for a while, and mostly I’ve been wrong – since it keeps not happening.  Maybe I’ll finally be right.

Curiously I’d not noticed before a perversity in my presumption that the only way that the semantic web can succeed is if the big machines force the issue.   And that’s that the open world model so beloved by semantic web fans (me included) is totally at odds with this driver.

I found another JavaScript CLI scheme. Skewer which I’ll get around trying sooner or later.  But I got distracted by some adjacent work. Skewer’s scheme for connecting Emac to the java interpreter so it can do remote evaluation etc is to infect the java interpreter with a bit of code that then connects into Emacs using HTTP long polling. Which is good because it works with all the browsers, and is bad because … well actually it’s pretty good. Obviously, that requires that we have a working http server inside of Emacs.

Unsurprisingly people have written http servers in emacs-lisp, one of these is simple-httpd.  It is easily installed from the Melba package repository.

So, what might you do with such a thing? I already mentioned skewer, and I see that somebody wrote a web UI for his emac’s RSS reader.   Somebody else wrote air play server. Those examples are suggestive of what other emacs apps (calc, gnus, magit, erc, etc. etc.) might do. Why? Because they can? I don’t know.

My favorite is impatient-mode, which spontaneously update a web page as you edit your buffer of, typically, html. There’s a video:

You can add filters to the refresh pipeline. So if you want your org-mode files displayed you might refine the filtering scheme. I did this, but I doubt this is the best approach.

(defun imp-htmlize-filter (buffer)
  "Alternate htmlization of BUFFER before sending to clients."
  ;; leave the result in the current-buffer
  (let ((m (with-current-buffer buffer major-mode)))
    (case m
      (org-mode
       (let ((output (current-buffer)))
         (with-current-buffer buffer
           (org-export-as-html 100 nil output))))
      (t
       (let ((html-buffer (save-match-data (htmlize-buffer buffer))))
         (insert-buffer-substring html-buffer)
         (kill-buffer html-buffer))))))

Other things (graphviz?) would be fun too.