Tuesday, January 28, 2014

The Siege: Victory of the German Maidens

So there was a thing at the Modern Languages Association annual conference this year. It was about adjunct teaching blah tenure lines blah interviewing blah. Blah. My take is that, strip the precious specificity of academia away, and you're seeing a fairly typical problem of the politics of skilled labour.

Wait? You don't consider professors teaching English at a university  to be "skilled labour" in the same sense as the honest hands of toil making cool things with Very Big Machine Tools? 

A photo essay on cognitive dissonance: 

1: Cause:

2: Effect.

The "Help!!" sign is just about the definition of a shortage of skilled labour, and the skill in question is close reading. True, it is shelf schematics, as opposed to Beowulf, and it is so that buyers will know what is going to be on the shelves when they make their purchase. SAMS work doesn't have as much social capital as lecturing the children of judges and doctors on Point Grey -to put it mildly!-- but any adjunct in this country would kill for the lifestyle of our company's roving SAMS expert. 

In the face of the triple crisis of higher education, employment and demographics in this country and on this continent, it is worth contemplating the intractable mystery of skilled labour in our modern economy. Somehow, we got the story of skilled labour radically wrong. Our story is that an increased supply of skilled labour will drive the economy forward. When in fact, it just...

Oh. Wait. That's not a mystery at all.

Seventy years ago, labour worked differently. Voracious factories absorbed every hand they could, and then some more, and the Flak batteries in the park were crewed (in part) by the League of German Maidens. 

And in Belfast.. . In Belfast. . 

"In the county Tyrone, in the town of Dungannon
Where many a ruckus meself had a hand in
Bob Williamson lived there, a weaver by trade
And all of us thought him a stout-hearted blade.
On the twelfth of July as it yearly did come
Bob played on the flute to the sound of the drum
You can talk of your fiddles, your harp or your lute
But there's nothing could sound like the Old Orange Flute.
But the treacherous scoundrel, he took us all in
For he married a Papish named Bridget McGinn
Turned Papish himself and forsook the Old Cause
That gave us our freedom, religion and laws.
And the boys in the county made such a stir on it
They forced Bob to flee to the province of Connaught;

Anyway, in the land of anger and hate, there was this:

Wikipedia (also below)

It looks a little strange, doesn't? 

Here's a plane, English as roast beef:

The Stirling doesn't look like that. It's got .. Teutonic lines, frankly. Like this.

Or even this, I fancy. 

It is also not long for this world. On the night of the 22/23 November, 469 Lancasters, 234  Halifaxes, 50 Stirlings and 11 Mosquitos, a total of 764 aircraft, attacked Germany, 670 taking pictures of Berlin. 1,285 tons of high explosive and 1,491 tons of incendiaries were dropped. No less an authority than Wikipedia has determined this to have been the single most effective attack on the capital of the Reich. 26 aircraft lost - 3.4% of the force, below the 4% cutoff calculated as an unacceptable attrition rate. But 5 of those aircraft were Stirlings, 10% of the engaged force, this for the very good reason that the Stirlings were flying almost 3000 feet below the Lancasters. AA often rather unfairly dismissed as ineffectual, as Sebastian Cox notes, literally decimated the Stirling force. Ten percent losses over 40 missions was a death sentence to the Stirling crews and prefigured a spiralling decline. The effectiveness of the force was predicated on the crews' ability to use their weapons, and without experienced crews to train them, that effectiveness would never be achieved. 

Sometime in the aftermath, Air Chief Marshal quietly stood the Stirling force down. Ten squadrons were removed from the line. When Bomber Command returned to Berlin on the night of 26/7 November, it would be with 422 planes dropping a total of only 1764 tons of bombs (443 Lancasters and 7 Mosquitoes sortieing.) 

The official historian of the RCAF gives the clearest account of the sequence of events, evidence, I think, of the embarrassment of the British aviation establishment. Our detailed understanding of military aircraft tends to come from the manufacturer or the veterans. The former prefer to present a well-burnished story of the procurement process ruining their perfect vision, while the latter's relationship with their aircraft is often interesting. The Lancaster was a better bomber than the Stirling --as complicated as this story is on examination-- but it carried its crews on death rides into Germany. The Stirling force would spend the rest of the war as improvised transport crews and glider tugs. Inglorious, but safer work, and they would be left to sing the praises of a surprisingly manoeuvrable machine, a sentiment on display at the Wikipedia page. 

Well, then. Let's be college students again, and steal (non-computed data)  from Wiki. 

Short Stirling
Avro Lancaster
Boeing B-17
Boeing B-29
7 (First and second pilot, navigator/bomb aimer, front gunner/WT operator, two air gunners, and flight engineer
7: pilot, flight engineer, navigator, bomb aimer/nose gunner, wireless operator, mid-upper and rear gunners
Crew: 10: Pilot, co-pilot, navigator, bombardier/nose gunner,top turret gunner, radio operator, waist gunners (2), ball turret gunner, tail gunner
·         Crew: 11 (Pilot, Co-pilot, Bombardier, Flight Engineer, Navigator, Radio Operator, Radar Observer, 2 Gunners,  Central Fire Control, Tail Gunner

87 ft 3 in
 69 ft 4 in
74 ft 4 in
99 ft
99 ft 1 in
102 ft 0 in
103 ft 9 in
141 ft 3 in
22 ft 9 in
20 ft 6 in
19 ft 1 in
27 ft 9 in
Wing Area
1,460 ft
1,297 sq ft
1,420 sq ft
1,736 sq ft
Aspect Ratio

Empty Weight
46,900 lb
36,457 lb
36,135 lb 
74,500 lb
Loaded Weight
59,400 lb
68,000 lb
54,000 lb
120,000 lb
Max. Takeoff
70,000 lb
72,000 lb
65,500 lb
133,500 lb
4 Bristol Hercules II, 1,375 hp
4 Merlin XX, 1280hp
4 × Wright R-1820-97 "Cyclone" turbosupercharged radial engines, 1,200 hp
: 4 × Wright R-3350 Duplex Cyclone turbosupercharged  2,200 hp (1,640 kW) each
Max speed
282 mph at 63,000 lb at 13,000 ft
287 mph
357 mph
Cruising Speed
200 mph
182 mph 
220 mph
Rate of Climb
720 ft/min
900 ft/min
900 ft/min
21,400 ft
35,600 ft
31850 ft 
Wing Loading
44.9 lb/ft²
38.0 lb/sq ft
69.12 lb/sqft
0.093 hp/lb
0.089 hp/lb
0.073 hp/lb

    The comparison here isn't as useful as it might be. You really need to rigorously reference weight and altitude when you give performance numbers. Otherwise, there is a temptation to cherry pick the numbers, and long ago unfair comparisons are not always caught in the data. (Did the B-29 really have a lower service ceiling than the B-17?) V(cruise) is even more problematic. The air force can tell you what your machine is supposed to do when you crack the throttle, but it cannot tell the gaskets how much gas they should leak. What should leap out of the data is the sheer hugeness of the Stirling. In spite of being (slightly) older than the Lancaster or the massively refurbished B-17G, it is a bigger plane. The Lancaster and B-29 certainly shine comparatively as weightlifters, but let's look at actual planning data for a second.

     Here's a photographic reproduction of an actual raid planning guideline, complete with misspelled "Sterling," brought up from the bowels of the Air Ministry by Lionel Lacey-Johnston and reproduced from his 1991 Pointblank and Beyond

    The implications of the Lancaster's superiority as a load-carrier over the Halifax and especially the Stirling are clear. Again, we might ask: how did the Lancaster get so good?

    Well, here's another Air Ministry document, this one for public consumption. It's a detail of the wing structure design of the Short Stirling. 

    The Stirling was famously crippled by its wing. It was originally borrowed from the Short Sunderland, much as the B-24 borrowed the wing of Consolidated's last flying boat. 

    Unfortunately, the Sunderland looked like this (just plain clipped from Wikipedia):

    General characteristics
    • Crew: 9—11 (two pilots, radio operator, navigator, engineer, bomb-aimer, three to five gunners)
    • Length: 85 ft 4 in (26.0 m)
    • Wingspan: 112 ft 9½ in (34.39 m)
    • Height: 32 ft 10½ in (10 m)
    • Wing area: 1,487 ft² (138 m²)
    • Empty weight: 34,500 lb (15,663 kg)
    • Loaded weight: 58,000 lb (26,332 kg)
    • Powerplant: 4 × Bristol Pegasus XVIII nine-cylinder radial engine, 1,065 hp (794 kW) each
    • Guns:
    • Bombs: various defensive and offensive munitions, including bombs, mines and depth charges carried internally and, some, winched out beneath the wings. Manually launched flares, sea markers and smoke-floats.

    And, just for reference, its successor, the Short Shetland, looked like this:

    • Guns: (as planned)
      • Three turrets, each with 2 × 0.5 in (12.7 mm) Browning machine guns in nose, mid-upper and tail positions, and 1× 0.5 in machine guns in port and starboard beam positions
    • Bombs: Up to 4,000 lb (1,800 kg) of bombs or depth charges

    112ft wingspan on the Sunderland, going up to 150 in the Shetland! 

    Now, I am not privy to the thinking in the deepest recesses of Rochester, but I can see the argument for copying a successful wing. The problem is that this is in no way the "same wing." The Sunderland, like most flying boats of its day, carried its bombs in the fuselage, and winched them out onto the wing in flight in order to drop them. It's a bad idea, you see, to put the cracks of a bomb bay door in the bottom of a flying boat that is going to be crashing into the water at 70mph. 

    So can you see the  logic behind redesigning the wing to take multiple bomb cells, each able to hold a 2000lb armour piercing bomb, just when you are in the middle of developing the Sunderland's follow-on? I can.

    It's not a bad plan. In 1936, and even in 1939, there is a very real possibility that Britain's next enemy will be most effectively attacked by having squadrons of bombers fly over their battleships and dropping concentrated salvoes of bombs on them. It did not work out that way, of course. By 1942, it was clear that the war the Stirling was built for was going to be fought by carting massive loads of bombs over too-be-identified-later German industrial targets and dropping them. We have already seen how the Lancaster (and Mosquito) solve this problem: with massive longerons and spars capable of taking a concentrated load of bombs distributed equally around the aircraft's centre of gravity along the fuselage. The baroque complexities of the armament distribution through the wing cells is not exactly compatible with strength and simplicity.

    Naturally, this is not what the voice of the firm will say in 1942. It will instead suggest that the Air Ministry was "presumably" thinking of conserving hangar space. The splatbook literature will take this factoid and run with it, and, if I recall correctly, it even features in John Jame's still unparalleled social history of the RAF that attempts to reconstruct the history of its buildup in terms of -gasp- its workforce! Wikipedia suggests that the limit was an attempt to cap Short Bros.' ambitions. The force of the Air Ministry's sharp "Now cut that out," is, however, perhaps underplayed. After all, even the Air Ministry could not say for sure in 1939 that Rochester was wrong. 

    The problem here is that it is somewhat weird and contingent that there is even a four-engined bomber under development in Britain in the spring of 1939. Short Brothers being one of the problem children firms of the mid war, get little credit for their achievements before it. Whatever one thinks of the operational fiasco associated with them, the Empire Boats line was a massive industrial success. The Stirling made a credible alternative to the Ministry's first choice, the Supermarine 317. 

    Now let's think about that for a moment. Vickers was a good engineering firm. It bought Supermarine because Supermarine was good at what it did, too. The story as it is told on Wikipedia is that Reginald Mitchell's death brought everything to a halt, but the subsequent history of the Spitfire is vindication enough of Joseph Smith as a designer. The retreat from the 317 is not to be explained in terms of an individual tragedy. It is a recognition of just how hard it was to design and build amodern four engined bomber that did not turn out to be oversized and underwhelming, the kind of monster that in retrospect leaves one asking, "What were they thinking?" * We are looking at the limits of engineering management, circa 1939, here.

    From any reasonable perspective, the Stirling was a grand success on the morning of 14 May 1939. It was not a secret, having been revealed by the French and American press. Albeit that this was still a rumour amongst less accurate ones, the fact that Short also had a contract to build a four-engined long range airliner made the existence of this "Stirling" seem pretty likely. What I am saying is that there was no really compelling reason not to fly the Stirling amongst the other spectacular prototypes exhibited in the "Parliamentary Garden Party" thrown by the Government that June in a desperate attempt to prove to the world, and especially Hitler, that the British air force really was rearming with effect. I don't know if the Stirling would have been exhibited that day, and it is a little hopeful that it would have finally got the whole "reap the whirlwind" thing through Berlin's head. Then, of course, the prototype Stirling collapsed its undercarriage landing from its test flight.

    So much for that idea.

    Whirlwind reaping was further set back by a spectacular low altitude raid in August 1940 that damaged the factory and destroyed 10 completed Stirlings on the ground. I would repeat Wikipedia's idle prattle about how this bombing set the Stirling project back by a year, but since everyone knows that strategic bombing was a pointless waste of resources, I won't muddy the water with pointless facts. What we do know is that the first Stirling attack on continental targets took place on 10/11 February 1941. 2,383 were built, and, as noted, they were withdrawn from combat just as the weight of the strategic offensive began to tell.

    It was a painful story, complicated by the choice of locations to build the aircraft. Rochester looks like a bad choice of location, the tip of Kent being far too close to enemy air bases, but it is not a good idea to waste factories, and Sunderlands were built there in numbers. Stirling production was devolved to the Austin plant at Longland in the Midlands, and then to Belfast, where Short had cooperated with shipbuilder Harland and Wolf to build an aircraft plant with an inglorious record of producing some of the RAF's least successful bomber projects (1,2).  What could you do? Belfast somehow managed to be full of skilled labour with nothing to do, and when they didn't have anything to do, they took out their boredom on Catholics.** Not a good plan for a country that aims to ally itself with the United States. (Theme music that's not "The Wearing of the Green.") 

    And so the machines of Belfast were flung at the League of German Girls, and the Giant Wurzburg-directed blind fire predicted brackets gripped the giant planes and threw them to the ground like falling leaves. 

    I've offered the idea that this is a story about labour. Belfast might serve as a paradigm for our world: a black hole of skills training demonstrating that a supply of skilled labour is not the same as a demand for it. The best efforts of these able, if fractious men were extraordinary machines, byzantine assemblies of remarkable design.

    The machines that met them could be described in the same way, with this crucial difference. They were built to meet the unschooled talents of the girls of Germany --among others, if I am going to be accurate, as opposed to romantic--. Stripped of their caissons, they were "static" guns --that much easier to cater to, they were much less flexible in operations. The optical sights and predictors that ought to have allowed them to fight as individual guns, and, for example, shoot tanks, were omitted. Too expensive to provide to every barrel, the guns were instead fought as batteries under the direction of a master predictor and sight. The gunners and Flakhelferin need only swing their guns onto their pointers and fire and feed fixed rounds into the hungry maws of the guns. 

    That being said, what I have described is not what actually happened. Anyone who has dealt with the electronics of the era would know differently. The death of the Stirlings tell us that these problems, of unsourced buzzing hums and the acrid smell of burning insulation were solved, battery by battery, German Maiden by German Maiden.  

    There is, I suggest, something to learn about the relationship of technology and labour here..

    *Although the linked article manages to make it sound as though the USAAC planned for a plane that would roll out of the hangar it was assembled in and promptly break through the tarmac, sitting with its landing gear mired in the California mud for some three months before it could be rescued.
    **Tragically for everyone who has lived the song, the Irish Rovers' "When the Shipyards Go Back on Full Time" is not on Youtube. Perhaps it cuts too close.


    1. Look, are you taking the piss about image scaling?

      Never mind. The thing about the Stirling that sticks in my mind is that supposedly the main electrical distribution board is directly behind the RAF roundel on the fuselage, thus giving the fighter an absolutely ideal mark...

    2. Sorry 'bout that. Here's the thing. The Stirling was a huge failure of design, by the standards of the all-up-weight/structure weight ratio. This doesn't look like something that we would expect from the shop that gave the world the Empire Boat and Sunderland.

      What went wrong? Well, all we have is evasive "explanations" from the shop that boil down to it being all the Air Ministry's fault for cropping the wings.

      If there is more to the story than that, and I think there is, it's in the engineering drawings. Reading the damn things being a lost art, I thought I would give the Interweb the full resolution.

      Looks like I shouldn't have tipped the camera, though.

      As to why it matters, well I'm demanding that the world give us a better standard of history of technology than "It's gonna be alright! Google has self-driving cars!" Understanding how and why Rochester failed is important. I can't lay everything off on make-work for Belfast, because the Stirling wasn't designed with the Belfast works in mind. But I do not think it a coincidence that the Lancaster came out of an engineering town, while the Stirling came out of the dying dockyard towns of the southeast. That's a boat up there, that is, and not in a good way.

      1. Or the wing's trying to do too much, accepts sub-optimal aerodynamics because of the bomb cells, and has to be produced at a high rate when it's designed to be an exercise in fuss-and-fit and big margins for a long service life.

        There were 42 Empire mail boats built (over 5 years), not quite 800 Short Sunderland flying boats built (over 9 years), and there were about 2,400 Stirlings built over five years. (Compared to 7,400 Lancasters.)

        I suspect that somewhere in there, no one stopped and said "wait, wait a minute, we designed this thing to take the North Atlantic for a couple decades and keep passenger service safety margins, and what we're sending it out to do gives it a service life that might be under a hundred hours, almost certainly under five hundred, this isn't the wing for that".

      2. I suspect you're right, Graydon. In that sense, we've got a bit of "failing forward" going on here. Most of the RAF's heavy bombers have been built in batches of a hundred or less because that's how many heavy bombers the RAF has needed for 87 of its 96 years. (The same could be said for the Sunderland/ maritime patrol role.)

        So the Stirling came along at exactly the wrong time. Though that's what everyone fated to be born into such times, or something like that.

        I know that I'm reaching here, but I really do think that the Stirling has something to say about the way that we treat our reserves of skilled labour. In theory, Belfast was exactly such a reserve, kept precious and free of the pollution of paid work until it was time for the state to come along and order a few thousand bombers. And yet it was a complete bust, while the Reichsverteidigung managed to conjure up a fiercesome army of artillerists from what must have seemed the bottom of the barrel of the nation's pool of labour --no offence to the teenage girls of Germany! Not to repeat myself, or anything, but it's in the grind of full employment that we are going to see automation advance. Not in this relentless effort we are making to prime technological progress by supply-side means.

      3. Well, if you want to argue that the ideal state for an economy is a modest persistent real labour shortage, I will agree enthusiastically. People know they can quit and get a new job tomorrow, and this affects everything else. (I think the history of the western world from at least the Black Plague forward supports this view, too, along with allowing the observation that concentrated capital wants to make labour as cheap as possible and that this impulse is systematically destructive to the economy.)

        In terms of a skilled labour reserve, I don't think there's really any such thing in the general case.

        Automation or not, work has a context. That context does or doesn't have effective feedback. (There's nothing about free markets that guarantees feedback at all, never mind *good* feedback, and vice-versa with government work.) You've always (as long as there's more than four people) got a system, but it might not be a helpful thing. (and people *hate* any mention of system or process or optimization, because those things exist to tell them that they're doing it wrong or that they don't have a job. It's not easy stuff to introduce.)

        So I'd say in this case you've got a bunch of people obviously capable of building aircraft being somewhat confused about what aircraft they were building (which, to be fair, is in part because the RAF was confused about what aircraft it wanted or needed; the specification wants catapult launch and utility as a troop transport as well as the bomber requirements) and not getting good operational feedback. (The Lancaster benefits from the Manchester *not working*, to pick up your fail-forward theme, and the Wellington benefits from there being a war on, so the cost of the radical and excellent construction technique is acceptable; nobody would have tolerated the cost of the geodesic structure as a commercial cost.)

        Whereas the German Maidens are benefiting from a simpler problem -- ballistics isn't as tough as building aircraft -- combined with really high motivation (maybe they won't bomb *my* house...) and the professional artillerists approaching the problem as one that they had to systematize as much as possible. (Instead of Stirling's boat-like desire to rely on long apprenticeships and custom, in a context of rapid and peculiar technological change.)

    3. Just to be clear here, I am questioning whether there exists such a thing as a skilled labour reserve, and I say that as someone who has to train and train again, trying to reduce the essence of what I do. Holy crap is that hard! In the end, you let the once-a-month, never mind once-a-year contingencies slide. They'll have to learn that by encountering the problem and then standing in the figurative smoking rubble and asking, "Now, how could I have handled that differently?"

      Or as Nathan Rosenberg and Alasdair MacIntyre argued, the core of technological praxis is irreducible to rules. There is an experiental component, so that all skill comes out of "learning by doing."

      To dig down for a science fiction-y analogy, we suffer from the "Spectre General" Fallacy.* Somewhere out there, there is a body of skilled labour adapted to do a job that doesn't exist yet. So just invent the job, and the labour will be there. The job of the educational industry is to produce the labour so that the inventor/entrepeneur's genius can be actualised.

      Put it that way, and it seems clear that the Spectre General Fallacy is ideological in character. To get someone who will be good at a job that technology has not invented yet, we train people to the peak of whatever technological praxis already exists. This certainly might have an outcome in the economy, but more obviously by repressing wages than by unleashing the genius of the supply-side !

      *Not the band with the awesome metal hair. The Theodore Cogswell novella.

      1. (As a side note, why does your sidebar calendar undergo javascript St. Vitus dance?)

        I think there's a general case for organizing the application of labour; using it has an experiental component, sure, but I don't think most collective endeavours get that far, the usual organizational response to a new technology is to change as little as possible. (It's rare for the individual incentives to have anything to do with applying the new technology well. This is why startups are necessarily small. And why Toyota can derive substantial competitive advantage from something as basic as LEAN.)

        The only problem with invoking the Theodore Cogswell novella is that by some terrible alchemy, the people involved *are* good at fixing technology they've never seen. Not going to believe that one.

        So, sure, "skilled labour reserve" is highly questionable; either they don't know how to do this at all, or they don't know how to do it fast enough, or flexibly, and so on.

        "We have a culture and praxis and active skill set for solving problems" -- the period example has to be the Rolls-Royce aero engine shops -- is also real, though, and that's the thing you both want to preserve and are an idiot to leave idle.

        Whether that's harder to do, or just doesn't have the incentives, so it looks harder, I can't say. I don't think it's very hard to do as a thing, but this is a minority position.

    4. About the sidebar calendar, I do not know. I haven't added one, so perhaps blogspot, or some interaction of blogspot and your browser, has randomly attached one? (I'm not seeing it from my viewpoint. Or maybe I'm not seeing the St. Vitus Dance that is there. Or maybe it's because we're cutting cost by eating old rye.)

      The Spectre General Fallacy is precisely that there exists such a thing as a bunch of people who are good with a technology that they've never seen. It's the training that matters, and not the experiential learning that comes with working with them.

      As for skill embodied in a shop (Hello graduate student Bruno Latour!), I'm, uhm, I'm in a place. I work with, and amidst, a group of people who are very good at moving large amounts of stock through congested, undersized supermarkets. Thant's not an argument for setting out to build small supermarkets.

      I notice that some of our floundering competitors (who ever thought that frigging Walmart would ever belong in that category?) have decided that it is. I think that they'd be better off trying to raid our labour pool. But for that, they'd have to unclench the hand that grips the wallet and pay what they're worth, and Walmart seems unable to grasp the possibility that UFCW-scale wages actually buys something.

      This is why I like my letter from Mauritius so much. "We have a labour shortage here, and people are striking for better wages. But we shouldn't give in, because they're too poor and sick to work harder if they get them. Oh, sure, if they were paid more, they'd buy more food, but it would be bad food, so they still wouldn't work more. QED." Any argument is solid, as long as the conclusion is that you're not going to need to give something up.

      1. (by "Calendar" I mean "everything in the sidebar below (and including) "Blog Archive"; it bounces up and down, visually, which I think means it's being continuously re-rendered; at the very least, something in there uses up a lot of CPU. Other blogspot blogs I've been to don't do this.)

        As for skill embodied in a shop (Hello graduate student Bruno Latour!), I'm, uhm, I'm in a place. I work with, and amidst, a group of people who are very good at moving large amounts of stock through congested, undersized supermarkets. Thant's not an argument for setting out to build small supermarkets.

        It's not an argument for building small supermarkets any more than having a good piston engine shop was an argument for not building gas turbine aero engines, but I note that the skills transferred enough that Rolls-Royce's aero engine works is still standing. Much as there's an argument for a general skill of management, I think there's an argument for a general skill of solving (at least categories of) technical problems. It's just not ever fully embodied in individuals, because it's significantly organizational, and even if the individual knows how to do it, they can't do it by themselves, the organizational expression is required. (and you get awful messes if someone who is highly skilled at participating in such an organizational expression of competence gets put in an environment without that organization.)

        This is why I like my letter from Mauritius so much. "We have a labour shortage here, and people are striking for better wages. But we shouldn't give in, because they're too poor and sick to work harder if they get them. Oh, sure, if they were paid more, they'd buy more food, but it would be bad food, so they still wouldn't work more. QED." Any argument is solid, as long as the conclusion is that you're not going to need to give something up.

        I think this stems from misunderstanding the overall goal. (There's a digression about what "the purpose of a system is what it does" says about people who want to own latifunda-model enterprises, but in this company I believe I can take that as read.)

        If you believe that the purpose of the organization is to make you rich, or distinctly relatively more successful, or whatever set of terms one wants to attach to the basic primate desire to be obviously high status in ways the other primates either must or can be compelled to acknowledge, you're making an intractable mistake in the context of an industrial or post-industrial economy. (Something to which social norms haven't really adapted, so far as I can tell.)

        Status -- whether it's social or derived from greater profit than competitors -- comes from relatively superior performance *at something else*. Once you start treating the problem as "deliver me status" (or, lately, the generalized status proxy money) you get stuck in a place where you can't look at the thing you need to do relatively better directly, you always have to squint at it through dollars and this doesn't work very well. (In large part because if you're trying to maximize profit you're stuck destroying value; value's benefit-to-cost, and you can't maximize profit without either raising cost or reducing benefit. Customers notice.)

        So I think the writer of your letter from Mauritius sincerely thinks "labour exists to deliver me status", does not know this, can't examine the axiom, and proceeds to completely logical conclusions given their axioms. It'd be a lot easier problem to address if it wasn't axioms, people aren't generally willing to admit they have those, never mind alter them.

    5. As a minor quibble, the repair guys in the novella did know about most of the technology from manuals and a few examples, IIRC.

    6. In Cogswell's The Spectre General, a bunch of effectively Neolithic (they have flint tools!) Imperial Space Marine 'Ercs have handed down their skills, via instruction manuals, through the generations so that their descendants, so that they can repair spaceships better than the actual mechanics who work on the actual spaceships of the successor state, who are incompetent because reasons.

      It's a great story. I loved it as a kid, and I love it now. It is a real statement of the importance of the mechanic in a complex, industrial system.

      The problem, and the starting point for dissenting from the SGF is that we have good reason, from experience and from philosophy, to find it implausible that anyone can learn to excel in a skill from simple instruction. We can move on from that basic point, via MacIntyre's reading of The Nichomachean Ethics to a moral aporia, or we can refuse to trod that high ground and ask what it might mean for technical education within industry.

      If people can be taught to reproduce a technology that we do not yet have, then the solution to technological progress is easy. We just have to train new generations of technicians at the highest level of scientific abstraction. New, life-changing products will then flow from their l33t engineering physics skillz.

      If, on the other hand, "learning by doing" is the key to advancing one's techne. we have a shop in which people get better by trying to advance their trade. We rescue Aristotle's account of how one becomes fully a moral person (taking care to get rid of the incidental apology for slavery), if that's our bag, and come to a better understanding of how technological change happens.

      Graydon wants to make this about shops --organisations. If you keep Rolls-Royce working, better Merlin engines are going to keep on rolling out.

      Why? This is what Rosenberg calls looking into the black box of technological change. Instead of technology being channeled into the real world of human doings by an inventor/entrepeneur, it emerges from shop floor praxis. Learning by doing leads to innovation by doing.

      Perhaps --probably-- some shops are better at this than others. But as a counterpoint to Graydon's Rolls-Royce, I would advance Armstrong-Siddeley, a shop of bumblers, one failure or undershoot after another, and then --bang!-- the world's first successful axial jet turbine. Innovation in the aviation world, Lutz Budrass suggests, is through "failing forward." This is surely, in part, because defence departments are so much more willing to tolerate ambitious failures than the private sector. Perhaps it is even an argument for seeing wars (at least, wars with a serious agonistic component) as driving technological change.

      Or maybe it's all about money.

    7. Learning by doing *doesn't* lead to innovation by doing. It can't. Practice makes permanent, and you'll find with most shops what you've got is a machine to *resist* change. (Some of what I've done professionally has involved getting people to change how they document. It's notable that no amount of measurable less effort suffices as a reason for the change.)

      There's a tension between "do the job" and "don't get yelled at", and then there's a whole raft of questions about "job". (Because someone does have to pay for whatever value you manage to deliver if you want to keep doing this for long enough to say "culture" with a straight face.) I'd say Armstrong-Siddeley had to be fairly far off toward the "do the job" given their diversity of products and a market stance of selling quality to quality; we're not (unlike Rolls-Royce) trying to optimize our production, we're trying to optimize how closely we can match what the customer wants, sort of thing.

      There's a huge difference between "we've always done it like this?" and "why do we do this? do we need to do this at all? Who wants this? do they pay?" as default questions, socially accepted as the appropriate answers to "why do we..." questions from new people in the shop.