A Real Printer's Devil

In my earlier posting on my grandfather's article I should have perhaps explained the term "printer's devil" which appears there.   It refers to a printer's apprentice.  The following (taken from an unpublished autobiographical note by my grandfather) gives an idea of what it was like to be a printer's devil in the first half of the last century:

Walter Brelstaff went to Providence Infants' School in 1908, afterwards to Northgate School which he left in 1916 to start work as an errand boy a J. T. Stokeld & Sons, local printers. He was for some time a real printer's devil, having to wash the large and small composition rollers used for automatic inking on the presses. There were three machines for printing larger sizes of paper and two smaller machines. It was a very dirty job, using paraffin to remove the ink. A black apron was poor protection. It was some time before he could get rid of the soreness and traces of ink. There was only a cold tap to wash hands. The large lead-lined sink was also used for washing type taken of the presses. Paraffin again. A large thick board sloped down into the sink to support the 'formes' - iron and steel frames into which the type was securely locked.

On the web you will find several very fanciful, contrived and almost certainly incorrect derivations proposed for the term "printer's devil".  It most probably comes from the common use of "devil" to refer to a dirty or mischievous child.  My guess is that it originated back in the Middle Ages, probably fairly soon after Caxton introduced the printing press to England.

A Starfish Decoy Site on the Cleveland Hills

On the moors above Guisborough stands this brick and concrete bunker which,  as children, we always referred to as the Troll's House.  My father said that it was built during the Second World War to hold oil that was to be used to set the moors alight in order to distract German bombers from dropping their bombs on the nearby industrial town of Middlesbrough.  Well, a little Googling reveals this to be approximately correct, but the full story is somewhat more exciting.  Bob Hunt's Portsdown Tunnels web-site has a good introduction (apparently taken from an article by Fred Nash of Essex County Council):

In the autumn of 1940, the mass bombing of Britain's cities - and in particular the devastation wrought on Coventry - turned the attention of Colonel J F Turner's Air Ministry department to the possibility of providing decoys to divert the bombers away from large urban targets. What was needed was the replication, as far as was possible, of a city under incendiary attack. It needed fires on a truly legendary scale. And thus was born the fabled Starfish site, the largest, most sophisticated decoy built in the war.

Within days of the raid on Coventry on the night of 14/15 November, work started on the construction of these huge fire-based decoys around Britain's major towns and cities. From Southampton to Glasgow, from Birmingham to Crewe, tons of wood, and indeed anything combustible, was gathered to make improvised bonfires to try and draw some of the bombers away from the centres of population. But although immediately successful in some places, notably Bristol, Derby and Sheffield, these makeshift fire arrangements needed to be improved to guarantee a long-lasting effect with different kinds of fire to give a realistic variety of colour and intensity. If the Luftwaffe bomber pilots were to be deceived into thinking they were attacking a blazing town, the effects needed to be dramatic, varied and sustainable.

Tests and experiments followed. Using steel tanks, troughs, pipes and grids, fuel could be made to pour, spray or trickle at timed intervals. Different kinds of flammable materials, variously soaked with boiling oil, paraffin or creosote, could create the random fire effects of burning houses, factories and power stations. Finally, four kinds of fire source which could be "manufactured" were decided upon, and construction work began.

There is also some interesting material on Starfish sites and other types of decoy sites by Huby Fairhead at the Norfolk and Suffolk Aviation Museum.

It was the design and construction of these simulations of burning towns and cities that was known as Project Starfish.   I guess that the Guisborough Starfish site was one of the simple, first generation sites. Setting alight the heather moors would have been the obvious way to start a big fire quickly in the Middlesbrough area. However, my father says that the site was never used and the moors never were set alight, so either the German bombing of Middlesbrough wasn't as heavy as had been feared or else the site quickly became obsolete and was replaced by a more complex decoy somewhere else in the area (possibly a more easily accessible lowland site?).  Another possibility was that, while the moors were not set alight, the site did use some of the more sophisticated pyrotechnic devices and that these were not as noticable from Guisborough.  But this is really just speculation on my part.

This shows the view from the south with the North Sea in the distance.  In this view, Middlesbrough is hidden behind the left end of the bunker, but there is no doubt that the occupants of the bunker would have been able to see clearly if Middlesbrough was ablaze.  [Note added 2007-09-18: Middlesbrough itself would be hidden by the Eston Hills, but the glow would still be visible.]

This shows the main entrance which, originally, would have had a blast wall immediately in front of it.  The blast wall, together with part of the sloping wall to the right of the door, may have been demolished to provide rubble for repairing the nearby Kildale Road.

And this is a view from the north taken by my father. It was seeing this photo on his Flickr site a few months ago that renewed my interest in the history of this building.

According to Huby Fairhead, Project Starfish was considered a great success:

The decoy towns were bombed about 100 times, drawing some 5% of the bombs intended for towns and cities.  Official figures declared that Decoy Sites saved an estimated 2,500 lives and avoided 3,000 injuries; four civilians were killed through raids on decoy sites.  These 1946 figures seem to be on the low side.  It should be noted that, in most case, only bombs dropped on sites could be counted as being “dropped on the decoy”.  After the war, however, it was agreed that bombs dropped within two miles of a site should have been counted to allow for inaccuracies due to Anti-Aircraft fire, fighter aircraft, weather conditions or poor bomb-aiming and navigation.

You can find Starfish decoy bunkers in other parts of the country.  Here is one near Glasgow photographed by Russell W. Barnes, and here one at Kislyth (also in Scotland) photographed by Gordon 'The Historian'.  I even seem to remember seeing something on a walk past Hardwick House near Mapledurham  (just 4km from Reading) that I now realise could be one of these bunkers.  Sometime soon, I will take a walk out that way to check it out.  [Note added 2008-04-18: I eventually did.  I doesnt look like a starfish site, just four air-raid shelters associated with the hall.  See here for more details.]

Update on Tue 2007-12-11 by Tristram Brelstaff

For more wartime sites in and around Teesside see here.

Update on Tue 2010-10-12 by Tristram Brelstaff

Here is an article on a Starfish site at Arborfield near Reading.

Avoiding the Age of C++

While I am on the subject of Luca Cardelli I should point out the slides to his autobiographical talk An Accidental Simula User which provide an excellent background on the motivations and significance of his work.  I cannot resist quoting one sentence:

Java allowed me to move from one type safe language (Modula-3) to another, narrowly avoiding the entire age of C++!

Lucky man!

You might notice that Cardelli ignores the common advice about not putting too much information on overhead projector slides.  However, I think it is better to regard his slides as a colourful magazine-style article meant to be read on-screen and, as this, they are rather good.  (You might also note that Cardelli's web-site would never win any awards for web-site design but, after all, it's the content that matters.)

Emperor Dragonfly

This afternoon I went for a walk round the Reading University lakes.  It has been a few weeks since I was last up there and the summer is definitely coming to an end: lots of greenery, few flowers, and the long grass has been mown and is strewn untidily,  slowly turning brown on the slopes leading down to the water.  The lakes seemed very muddy, maybe there is some building work going on somewhere, and there were only a few sad mallards moping around the margins.  At this subdued time of the year, one of the few places were you can still find lots of flowers and insects is along the western edge of the top lake.  Here I came across an emperor dragonfly (Anax imperator) calmly devouring a bug.  As I watched, the bug's wings and hind legs disappeared into the dragonfly's maw.  The sun came out for just a couple of minutes enabling me to take a few decent photos.  On the walk back I was fiddling with my camera, looking through the photos I had taken, when a passing woman asked what there was to take photos of on a dull day like this, and I told her about the blue dragonfly.

"Hiring the Creator as a Consultant is not considered Part of the Scientific Method"

I have long admired the work of computing scientist Luca Cardelli.  I even have a copy of his magnum opus A Theory of Objects (coauthored with Martin Abadi) and, which is even more impressive, have actually managed to read about a third of it.  (A Theory of Objects is to object oriented programming what the Principia Mathematica was to mathematics.)  While most of his work is rather technical, he has produced several sparkling semipopular talks and articles, such as this paper in which he compares the ways a biologist and a computing scientist might go about understanding a tamagotchi.  I particularly like the following section from which the quote about "Hiring the Creator" comes:

Here are some typical questions that arise from attempts at understanding principles of organization. We begin with a couple of questions that are not normally asked by biologists, but that we need to considered because they would be first in the mind of technologists:

Q1: Who created it? We actually known the answer to this question (Aki Maita [pictured]), but it does not help: hiring the creator as a consultant is not considered part of the scientific method. Moreover, how could something so unique and sophisticated as a Tamagotchi have suddenly appeared seemingly out of nowhere? It is such an unlikely and unparalleled phenomenon that we have to question whether we could ever actually understand the mind of the creator, and whether the creator herself truly understands her design (“Aki's own Tamagotchi seldom lives longer than its baby stage.” - Apple Daily).

Q2: Where is the documentation? Well, there is no documentation, at least no design manual that explains what its principles are or how it works. Even if we could acquire the design manual from the creator (by industrial espionage), it would be written in the Language of the Creator, i.e., Japanese, and would be of little use to us. Now, turning to more scientific questions:

Q3: What is its function? What does a Tamagotchi compute? We have here a relatively primitive information processing device, but there is no easy way to explain what its processing function actually is. In fact, its function is not quantifiable; it does not appear to compute anything in particular. And how can we hope to understand its design principles if we cannot say what it does?

Q4: Why does it have 3 buttons? There surely must be a deep reason for this: 3-button devices are comparatively rare in technology. Did it evolve from archaic 2-button devices of which we have no record? Is 3 just the closest integer to e? Is there some general scaling law that relates the size of a device to the number of its buttons? It seems that none of these questions can be answered from abstract principles.

Therefore, principle-driven understanding fails.