29 October 2015

Shave and A Haircut, Two Bits

These are the two bits.

1) The Community Technical Preview of SS with in-engine R connectivity has been released. Runs with the Revolution R version (fee when in general release?). Will the free-as-in-beer version of SS support either version of R in-engine? I have my doubts.
For SQL Server 2016 CTP3, we support ad-hoc execution of R scripts via a new system stored procedure. This stored procedure will support pushing data from a single SELECT statement and multiple input parameters to the R side and return a single data frame as output from the R side.

Not as comprehensive as Joe Conway's PL/R, which lets you create arbitrary functions using R. Well, nearly so at least.

2) An announcement on Planet Postgres that Tom Lane has moved cube. Most folks have no idea who Tom Lane is, and his move isn't germane to this post. The move may be more significant to the PG development community, possibly. But in the text of his new employer's announcement, is this bit:
... as well as the recent of addition of Joe Conway, another highly respected member of the PostgreSQL development community and committer to the PostgreSQL project.

Joe does lots more than PL/R, but moving to an aggresive, by the looks of it, start-up poaching widely, we may well see more robust PL/R support in Postgres. Shut up those MySql folks. Sorry Larry.

28 October 2015

The Global Euro

Let's take a trip through Tim Cook's head. What must be Tim's constant thought? Well, as CEO of the best company on the planet, it must be: how to continue being the best company on the planet over the next decade, or so. Best, of course, from Wall Street's perspective. After that, he'll just pull the rip cord on his golden parachute.

How did Apple get here, one might muse? Was it Steve's genius? Not really. Apple rescued Steve as much as Steve is credited with saving Apple. Next was a failure, pure and simple. Until the iPhone, Apple was just a niche personal computer company. Still is, if you cull out all the non-computer revenue from the filings.

All that wonderful gross margin from iStuff comes from a simple fact: Apple works by cheap goods sold dearly. It has managed to do this by the cult of personality of Steve. The innards of any Apple product are bought in off-the-shelf. Yes, I know: the Ax cpu is wholly Apple's creation of genius engineering, blah blah blah. Fact is, not so much. All the teardowns of the various A chips indicate: added more cache, added more gpus (bought in, of course), and did a good bit of manual layout. The ISA of an ARM chip is set. The node is set by what the foundries can do. If the MO of Apple were not clear before, then the destruction of GT Advanced Technology should make it clear: Apple gives you a pittance, then shoves a telephone pole up your butt. Which fact answers the Big Question: why does Samsung supply Apple? The answer is, they can afford to. Only another economic Gorilla can. It's what the econ types call countervailing power.

It's worth recalling that the first iPhone had a strong advantage for the first couple of years: Apple had captured the supply market for cap touch screens which fact limited competition severely.

(I started this missive while still on the Island, and it's now mid-day Tuesday. I'll wait for the Apple quarterly news before continuing, although the share price isn't moving up.)

Just got back from a teeth cleaning, and Apple is getting cleaned in after hours. I'm going to wait until 8 pm eastern to continue. ... Was down a bit in after hours, now up a bit in pre-market. So.
"The law of large numbers does kick in," he added."Can the iPhone grow another 2-3 years or is it done growing?... I think investors are worried about the next two-year period. 'Can there really be iPhone growth?' Because absent that there's not really going to be much growth from Apple."

In simple terms: China was the last roll-up opportunity. If that sounds alien, check out what's been going on with Valeant; roll-up city. This quarter went OK, and Christmas may as well. Does either speak to the future? No.

Which brings us to Germany, Greece, the Euro, and the Trilateral Commission. One of, if not the largest, paranoid fantasy of the far Right.
On the right, a number of prominent thinkers and politicians have criticized the Trilateral Commission as encroaching on national sovereignty. In his book With No Apologies, former conservative Republican Senator Barry Goldwater lambasted the discussion group by suggesting it was "a skillful, coordinated effort to seize control and consolidate the four centers of power: political, monetary, intellectual, and ecclesiastical... [in] the creation of a worldwide economic power superior to the political governments of the nation-states involved."

Well, perhaps The New World Order is.

The point, of course, is that Germany, through the elegant expediency of forcing the Euro down all those countries' throats, has the benefit of a New European Order, which solely benefits it. While Volkswagen can adjust the price, in Euros, of the Passat across the countries (last issue of report), higher in France while lower in Greece, for example recognizing ability to pay; this is an active step which is totally different from the passive shocks incurred when China, say, pulls the rug out from under the Renminbi. What Apple wants, should Tim ever admit it, is for all countries to do as Bermuda has: a fixed 1-for-1 exchange with the US Buck. Bermuda even calls it a Dollar. That's what Germany got with the Euro. That's what Apple, and the rest of US multis want. Each quarterly they bitch about "currency problems".

A New World (Monetary) Order controlled by Apple and fellow co-conspirators wouldn't upset them in the slightest.

25 October 2015

I Still Hate Neil Irwin, part the third

Spent the last week on the Island, wonky WiFi and just the Wifey's Windoze laptop. No editor or links. Very quiet on the rabble rousing front. So, I get back and The Times does its cynical or ironic or sarcastic two-step. Time to rabble.

First, a long piece on collapsing commodity markets. I'll just leave that for you to peruse at your leisure. No, this missive is all about Neil Irwin and me. Yum.

More than once, I've referenced pieces by the esteemed Mr. Irwin because they provide reportage proving (or, at the least, demonstrating) certain themes of these endeavors. Most often, as this one, they deal with quant (data) issues in macro-economics. Always a thorny problem. Most macro data is generated as sample surveys of business and labor.

Today's piece deals with a subject near and dear to my heart: Dee Feat is in Dee Flation. Not been a missive under that title in some time. Not that there's been a dearth of data, of course, just that the tune remains the same: we're headed down the rabbit (or, rat, as you prefer) hole of deflation. The cash hoarders look likely to get their return on idle moolah. Sigh.

Irwin takes a new path. One I haven't discussed since, I'll guess, grad school. Perhaps undergraduate. The dreaded Phillips curve. He explains, a bit.
How much faith should be placed in a line on a graph first drawn by a New Zealand economist nearly six decades ago, based on data on wages and employment in Britain dating to the 1860s?

The answer, of course, is not much. Phillips, as many macro-economists (esp. of the Right Wing, labour hating, branch), assigns only that sole cause to inflation. But, as the first piece so clearly shows, flation isn't just about wages. Far from it. Just ask your local greasy spoon proprietor why s/he's exploded the price of your 2 over easy.
As the Fed's chairwoman, Janet L. Yellen, put it in a 2007 speech, the Phillips curve "is a core component of every realistic macroeconomic model."

Except it doesn't work. Or at least, it hasn't worked very well in the last few decades in the United States. And it has proved particularly problematic to try to use that historical relationship to predict where inflation is going.

What he neglects, shamefully, is tell the reader about other drivers of flation. That's my task. Again.

To refresh, there are three ways flation (either direction) happens:
- wage push
- cost push
- demand pull

The first happens when workers get paid more than their marginal product (wiki), at least according to even Right Wingnuts who worship at the feet of classical economics. Otherwise, it's just Social Darwinism with capital compressing wages to subsistence, i.e. slavery without the chains. Which worked fine, in a macro sense anyway, during pre-industrial and even early industrial regimes. These days, with such heavy capital requirements, the The Tyranny of Average Cost™ makes it a fool's enterprise. Irwin goes on a long discussion of why Phillips and actual inflation are disconnected, beyond the simple arithmetic.

The second happens when there's not enough widgets to satisfy needs. Petro goes through cycles, which we're always experiencing. And, for the nonce, your 2 over easy. Why the likes of Irwin can't point out this obvious fact is puzzling. Over the decades, and centuries, I'd wager that material shortage is the number 2 cause of inflation.

Finally, that third is what the Right Wingnuts prefer to call "debasing the currency", and why they hate QE and fiscal policy and every other tool used to resurrect an economy. Argentina in the 1980s. But even Left Wing economists know that dropping moolah from airplanes will, if long enough and large enough, cause inflation. This is particularly true when an economy:
- is largely non-self sufficient
- lacks surplus capacity in desired widgets

In other words, the USofA today. Financialized economies, Bermuda is my favorite whipping child, are more susceptible to demand pull inflation, since the moolah dropped from airplanes won't be spent on domestic production. Spent on domestic production, the point of dropping moolah from airplanes, we find more demand for home grown widgets; thus more employment, more investment, more profit and so on. The point of dropping moolah is simply to get folks buying stuff. However, if the stuff isn't created domestically, the follow-on benefits don't occur. And so it was in Argentina.

Simply relying on naive` Phillips is foolish.
If you simply look at the unemployment rate in the United States versus the Consumer Price Index, excluding volatile food and energy prices for every year since 1958, there is nearly no statistical relationship at all, just a jumble of dots. (A best-fit line actually points the wrong direction, correlating higher unemployment with higher inflation, albeit very weakly.)
[my emphasis]

To the rescue comes a Freshwater Economist, Robert J. Gordon of Northwestern. Yes, another nerd who sees the future as distinct from the past. The link is an interesting read, for nerds anyway, but on to the day's topic:
Robert J. Gordon, an economist at Northwestern University, has his own version that he argues explains inflation levels throughout recent decades. But it is hardly simple. Its prediction for inflation relies not just on joblessness but also on measures of productivity growth, shifts in food and energy prices and overall inflation over the six preceding years.
...
or by factors that policy makers have little control over (like what happens to oil prices)

In other words, the sum total effect of those two other, oft ignored, forces. It ain't just the greedy 47%.

The Tyranny of Average Cost, part the third

Once more, The Tyranny of Average Cost™, Volkswagen's emissions venture:
EA stands for "entwicklungsauftrag," or "development order," and signified a major new engine line. Volkswagen deploys its engines and other components not only in Volkswagen cars but also in other brands belonging to the company, like Audi and Skoda. The strategy saves money because costs of development and production are spread across a large number of vehicles. But it means that any problems potentially affect a large number of vehicles.

Once more (the EA189 and EA288 engines): 'we lose money on each unit, but make up for it with volume'. Thus spreads the infection.

17 October 2015

The Walk

No, not a review of some movie about some Freedom Man skipping on a wire between two towers that no longer exist.

This is about walking the walk and talking the talk. More particularly, that Agile doesn't actually do either. The piece is really an exercise in self-promotion, but embodies my experience. It isn't RM-centric, but worth the read. (Disclosure: found via Database Weekly newsletter; you should sign up if you aren't already.)

16 October 2015

The 2 Percent Solution

For some time, both the quant version of these endeavors and the macro-economic version have harped on the quite apparent lack of investing in real assets. Corporations are sitting on $2+ trillion just overseas. QE money has largely gone to financial instruments. True innovation in physical widgets has clearly stalled: the best new thing is Force Touch on an iPhone?? Really?

The CxO class just can't seem to find creative ways to allocate fiduciary capital to physical form. I found this HBR piece recently and was going to write it up. Nah, just go read it. The bottom line, so to speak, is that the Masters of Wall Street are extorting the CxO class to put moolah into fiduciary, rather than physical, investment. The United States of Bermuda.

Well, even Yahoo! Finance (am I the only one to view that as oxymoronic as "happily married"?) now gets it. Real value in an economy comes from physical investment in productive processes. Finance and real estate and insurance (FIRE, as it's known) are just ways to skim moolah off the transfer from savers to borrowers. All their profit comes from extortion. The American Way.

15 October 2015

The Sun Salutation

From the time I exited grad school, I've practiced (and studied when I found an instructor on my wavelength) tai-chi and yoga. Both are Eastern, and tai-chi was originally a martial art. Both have, in Western incarnations, been promoted more as exercise of mind/body. Both are grounded in the same principle: the center (i.e. the space of the pelvic girdle) is all; all energy flows from it and out to the extremities. Golfers have been known to seek guidance from teachers of one or the other.

Tai-chi has a documented history that goes back some centuries, and has evolved into 4 (or 5, depending) schools. Yang is the most commonly taught, and the one I've studied. Again, when an instructor was available. Yoga, on the other hand, goes back millennia, and hatha form is what is mostly taught in the West. Both practices share a common attribute: each instructor popular enough to have a school/academy/franchise modifies "official" forms to better express his/her view of the fundamental principles of the art. And so it is. I first studied tai-chi with a student of William Chen, who taught a Yang form that looks nothing like the standard 24 step Beijing form. The one I teach now came from Art Ziello, from Don Miller (who is the brother of Julian Miller who taught me my first form in Cambridge too many years ago). And it's different from Art's version, which doesn't follow the official 24.

The same process happens in yoga, perhaps more so. Hatha is the main basic school in the West, but each instructor modifies to suit. And, unlike tai-chi, yoga practice is some arbitrary (well, sort of) combination of poses. The most well known documented version is, likely, "The Sivananda Companion...".

"What in Heaven's name does this have to do with quant and normal databases?" I here from the slackers in the back. Well, this. If you've been paying attention to the software world for more than a few days, you're surely aware of the patent trolls, software patentability, and suits from the trolls aimed at various and sundry. The latest (so far as I know) salvo, came last week when the Supreme Court declined to re-open Vringo's loss in the Court of Appeals for the Federal Circuit (CAFC, usually). The CAFC basically told Vringo to shut up and go home by gutting Vringo's "win" over Google. Ever since, the innterTubes has been a Tower of Babel.

The Blue Corner: "All software is algorithms, thus explicitly not patentable under both the Constitution and Federal law. Copyright is appropriate."
The Red Corner: "Software is unique intellectual property and therefore patentable."

Needless to remind, I'm in the Blue Corner and the Vringo, et al, are in the Red Corner. It bears reminding that the first software patents came from a Patent Office that had no experts in software, and were easily gulled (the account goes on a path I don't agree with, BTW).
What we learned is that things that we were doing in previous months are no longer acceptable. It seems that early last week a memo went out from the powers that be to the examiners handling Bilski-related applications, and in the memo it was explained that merely putting "computer implemented method" in the preamble of the claim is not something that will any longer work to overcome a patentable subject matter rejection under 35 U.S.C. 101.

More recently, not so much. The fact that a Supreme Court, deep in the pockets of the Red Corner generally, would refuse to bless Vringo's attempts is of note. Which brings us to today's news.

Bikram Choudhury has made an attempt to copyright his version of yoga, aka "hot yoga". He's just lost. I suspect that lawyers from the Red Corner cabal have already been on the phone offering any and all assistance to right the grievous wrong. All the way to the Supreme Court. You see, the court found that a sequence of yoga poses was not a copyrightable entity!!! This surprises even me. A sequence of words, as poem or novel or essay, certainly is.
"We wanted to establish that you can't own sequences, Mr. Drost, a founder of Evolation Yoga, said. "Traditionally, you take a teacher's training and move forward with it."

I can hear all those Red Corner sphincters sucking half the planet's atmosphere into their lower colons. It is well established that software is a sequence of discrete steps, taken from a previously established set, either the high-level language defined syntax or the cpu ISA. Now, here is a precedent that an real-world sequence, not even the abstract sort of computer languages, isn't even subject to copyright, let alone patent. There's going to be gallons of Chivas downed around Wall Street this week.

The Simple Truth

Up through my stint with CSC, SQL Server was a minor database in my trek. Once, even, the principal datastore. Along about then, I found simple-talk. Then, much of the discussion was driven by database design and development concerns. Celko was a frequent guest essayist.

In recent times, more of the essays have been couched in coder jargon. Sad.

One that's up on the front page, deals with continuous integration of databases. It's interesting, but what I found the j'accuse moment was this bit from a comment by rogerthat:
The worst times we've had with failing database builds were a result of using code first development (program design refactors database design, somewhat automagically). ... Experience leads us to design the database to handle the cited requirements while still keeping the database flexible (and no, this does not include unrealistic character data lengths).

While rogerthat doesn't use the N word, it's clear from the comment that normalization is the key. There is hope for humanity.

09 October 2015

Captain Obvious, part the second

This endeavor's been bleating about the need to use SSD as a matter of course from the beginning. More than half a decade. It is with some amusement that I read this quant/R related post. He who laughs last doesn't get it. Gad. Quants are just as bad as "database developers" who continue to dump flat files in SQL databases. Gad.
a huge increase in performance to the application overall

Well, yeah.

06 October 2015

The Tyranny of Average Cost, part the third

Another tidbit, this time from AnandTech.
Yet more worrying for manufacturers, the costs of preparing chips for new nodes isn't just rising but rising quickly, with mask sets already over a million dollars and expected to grow even further thanks to the high costs of developing masks for current multi-patterning technologies.

The Apple paradigm, selling only to the top X% (thought to be 20-ish), can only work for a slim minority of vendors. Moreover, it won't even work for that slim minority if they become the *sole* vendors from some external production process. In simpler words: the Apple paradigm only works, for Apple, iff there are other vendors to support the capex of said external production process. "We lose money on each widget, but make up for it with volume." Without all those, alleged, money losing smartphone vendors, Apple wouldn't have suppliers to eviscerate.
By splitting up a chip in this fashion, the number of transistors laid down on the leading-edge node would be held to a minimum, resulting in a smaller module that would be cheaper to design and cheaper to produce than a full SoC, all the while the other modules would be relatively cheap to produce and cheap to design (if not largely held over from existing designs to begin with).

There was a time, back in the late 70s and early 80s, when IBM held a massive tech lead on the seven dwarves (may have shrunk by one or two, don't recall exactly) with their "thermal conduction module". This was before VLSI and chips as we know them today. It appears that Marvell has discovered old wine and fashioned a new bottle. Smaller, yes, but exactly the same idea.

There will always be some new way to spread total cost, esp. capex. But, the bottom line, so to speak, is finding expanding demand for cycles. WinTel did that for decades. Now, it's the smartphone vendors and telecoms making the yin-yang of sink and supply. Life changes not so much as it appears.

04 October 2015

Hubris

Volkswagen demonstrated, once again, that The Masters of Universe make money the old fashioned way, they steal it. So, what does the Angela of death have to say?
"I believe the reputation of the German economy and the trust in the German economy has not been shaken by this to the extent that we are no longer considered a good business location," she told Deutschlandfunk, according to the text of an interview due to be broadcast later on Sunday.

Of course, Germany is a good place to do business. Government is the lap dog of business. Germany didn't define fascism, Mussolini did, but they perfected it. Government for, and by, capital. So long as Germany can export, particularly in the fixed rate Euro, it can sit fat and happy. Without export, the German economy collapses. Much like China.

02 October 2015

No Soup For You

Out of curiosity, I went looking at the MS/SQL Server site to check up on the state of the R-in-SS implementation. Not there, yet.
...still work in progress, not available yet.

Sigh.

01 October 2015

Told Ya So, part the second

Gentle reader may remember my musings that the ad blocker spewing was so much baloney. The cure is simple: price based on actual views. Seems Sergei and Larry are regular readers. Google henceforth will do so.

Now, was that so hard?