tentacle tentacle tentacle tentacle tentacle tentacle
Rob Sherman · 473y 10m 14d ago

Ramsay Island, on a seabird’s birthday, 2017

Rob Sherman · 220y 3m 43d ago
There the musky fishy genital smell
of things not yet actual: shivering impulses, shadows,
little amorous movements, quicksilver strainings and

each winter they gather here,
twenty seals in this room behind the sea, all swaddled
and tucked in fat, like the soul in is cylinder of flesh.

With their grandmother mouths, with their dog-soft
   eyes, asking
who's this moving in the dark? Me.

Alice Oswald, Dart
An Etiquette For Minor Time Travel

An 'Iron Duke' class locomotive, c.1850.

As a digital artist, I rarely use modern computers to interrogate modern life. Most often, it is the past that benefits most from the application of digital tools; though we can also learn a surprising amount about modern technology, and its effects on us, in the reverse.

My Twitterbot project, The Lost Pubs, tried to symbolically combat the frightening rate at which historic pubs were disappearing from the UK by generating a new, fictional establishment once every six hours. Conversely, my practice-based PhD sought to solve some knotty contemporary problems with the creation of interactive, AI-driven characters; not through powerful new algorithms or hardware, but by drawing lessons from the history of 18th-century popular magic, as another time and place in which ordinary people were attempting to live and work alongside beings that were not, biologically, there at all.

The Living With Machines project also embodies this reciprocal relationship between the present and the past, and the digital residency was an opportunity to work with such ideas in a new context. The datasets on offer each suggested fresh approaches for projects; but appropriately enough, it was the StopsGB set which resurrected an older one.

Several years ago I had wanted to create something that captured that frustrating but intoxicating feeling we have all experienced on trains: the reason why we scurry straight for the window seats when we climb aboard. It is a feeling provoked by those moments - moments we have experienced perhaps many hundreds of times - when we look up from our phone, laptop, book or attractive fellow passenger and catch a glimpse of something outside the window: two motorists either fighting or embracing at a junction; a terraced back garden filled with American car parts; a cloud of purple smoke curling around the tower of a village church.

These vignettes are too brief and detached to be called ‘stories’ in their own right: but they are cross-sections of something story-like, filleted glimpses of scenes into which we have been afforded a slightly perverse perspective as we slice past. However, our question - what’s happening here? - can never be answered; the drama shuttles off, back the way we came, tugged offstage as we chutter onwards.

Something in this experience seemed interesting enough, a challenge to recreate. I wanted to try and make something whole out of these fragments, to nurture those momentary frustrations rather than trying to remove them, as I might do in conventional storytelling. I fiddled up the idea of a visual poem: of vignettes running past the reader at the speed of an express train, some too fast to read, others gone before their significance could really be grasped. Sometimes, though, the poems would slow (a pause at a station, or a delay on the line) and the reader could focus on them properly, even if they could never alight into their detail as they might wish.

Living With Machines, and the StopsGB dataset - a georeferenced and metadata-filled chronicle of every UK passenger station - helped to resurrect my enthusiasm for this project by providing it with something beyond artistic novelty: a historical resonance.

If the voyeurism and tantalising mutations of the landscape provided by trains still feel addictive today, how much more transformative must they have been during the Victorian ‘railway mania’ of 1840-1880, under the influence of which many of these stations were built? With the cuttings and embankments fresh-dug and denuded, it was nothing less than an entirely new relationship with the time, space and stories of English life.

Whole topographies were levelled; new vistas and views opened up. Towns and villages were cut in two, and a local, specific and slower scale of living was sutured with something more national, more universal and much more rapid. The railways bored their way through the heart of previously self-contained, private places and contexts, with their insular, local stories, allowing the idling passenger to peruse landscapes and lives without committing to any one of them, and dismissing them with only a blink.

Of course, the network was not a passive witness of the communities it passed by: it sowed the land beside the tracks with new infrastructure, drifts of invasive plants carried on the turbulence, and new, prescriptive realities. The standardised ‘railway time’, pegged to the Greenwich Meridian, swiftly came to replace the ‘local time’ of each station stop, for thousands of years tied to the particular passage of the sun across that particular parcel of sky.

The tensions and interactions between the track and the communities it incised were, of course, pointedly messy: but while we often frame popular and literary reaction to the railways as Luddism by all but name, the truth - as with almost all technologies - is far less clear, and far more interesting.

A screenshot showing, on the left, the original spreadsheet of the StopsGB dataset, on the right the cleaned JSON of stops on the Great Western Railway.

Of course there is shared blood between the modern computer and the rail network, a common ancestor in clockwork. Both are machines that operate according to precise increments of time ticking up their smooth operation: and this project provided an opportunity to reunite them, and for each to comment upon the other. My visual poem would be a digital one, built using the popular videogame development engine Unity, and the StopsGB dataset would provide its spatial and temporal structure. While I briefly considered building a tool that could create randomised rail journeys from the entire dataset, given the limited scope of the residency I would have spent most of my funded workhours tinkering (and repairing) the basic code. Instead, I used the station data to construct a single journey along an existing trainline.

I decided on the original Great Western Railway and a date of 1850, when there was, finally, a single, continuous railway line from Penzance in Cornwall all the way to London Paddington, in a patchwork rope of competing companies, track gauges and (I imagine) painful ticketing issues. What’s more, I know the line well: my grandfather spent his career working the last steam trains out of the Swindon yards in the 1950s and 60s, and I now live near Exeter, where my favourite swimming beach is backed by the original Victorian tunnels sewn through the sandstone cliffs. Lastly, the differences in longitude between the eastern and western ends of the line meant the greatest possible difference between local times and the ‘national time’ kept by the railways. In 1850, these two times were still in dispute, the standardisation still not ironed out: station clocks still often had three hands.

Unfortunately, the dataset was not really created with this kind of specificity in mind. While I could isolate individual stations by their latitude and longitude, and the Wikidata code for their operating company, there was no easy way to automate the extraction of individual lines. Instead, I spent some time manually cross-referencing the data with Wikidata and the National Library Of Scotland’s Ordnance Survey archives to isolate those stations on the original line that were open in my chosen year of 1850. These were then imported into a JSON along with some additional data for each station: the terrain, approximate urban density, and the distance between stations.

This JSON file became the backbone and the metronome of my project in Unity: a subset of 68 points of station data that would drive a ten-hour, real-time recreation of a journey on this train line, by the tick of the computer’s relentless internal timetable, roughly sixty times a second.

A screenshot of the static 3D model of the train carriage.

Unity is mostly used for visually sumptuous 3D experiences, but I wasn’t planning to stretch these capabilities. The basis for the poem was a very basic scene, the audience’s viewpoint sitting inside a boxy, toy-like recreation of a train carriage. There is a certain degree of verisimilitude: the audience can use the mouse to turn their ‘heads’ slightly, though there is no recreated cabin, and they can never turn away from the frame of the window itself. Pressing a mouse button lowers the window itself and leans them out, where they can look up and down the carriage and into the void beyond. Devising this, and the rattles, shakes and lurchings of the carriage in motion, was the product of a delightful use of project funds: a research trip on the heritage South Devon Railway, from Buckfastleigh to Totnes, and my son’s first travel by rail.

The train carriage itself stays still, within the scene: it is the railway line itself, the world beyond, that moves from right to left, past the window frame and the simulated camera. This includes each of the 68 stations, sliding into view as a set of simple green arches, their names cropped from the JSON file and displayed beneath a large, pale-faced clock which shows both the current time - matched to the Greenwich Meridian by the computer - and the local time of the station, based on its actual longitude. In Penzance, at the westernmost point of the line, this is a difference of nearly 25 minutes. As the journey progresses back towards London, the difference shrinks as the train approaches the ‘official’ present.

An example of one of the station stops in the experience, with a generated advertisement.

Building this 3D application, I understand how people can lose years of their lives, and square metres of their garden sheds, to model railways. The temptation to chew up all of my time tinkering with the rigid, mathematically-precise scheduling of the stations passing by the carriage, based on scrap-paper calculations of the train’s top speed and the distances between them, was a strong one. However, a trainline (and a computer) is nothing without a cargo of human messiness; a poetry that doesn’t always cleave to such a timetable.

This project is not designed as a simulation of a 19th century railway. Such experiences exist, and took a lot longer than two working weeks to put together. I wanted as minimal a recreation of the physical phenomena involved as possible. The 3D environment does not extend to the landscape passing by the train windows. Instead, in the long tradition of concrete and kinetic poetry, in which the movement and spatial qualities of language add to their meaning, that landscape is recreated in words; lines of poetry moving past the window in parallax to give an impression of speed and volume. Farther-off sights move more slowly: descriptions of the weather in huge, grey, ponderous letters crawl over the distant horizon. Closer vignettes - a scruffy man in a meadow, making a shape with his outstretched hands, perhaps - whip past in an instant, a simulation of fifty nine miles per hour with a good head of steam, and are gone before the audience can really grasp at them.

If they do not touch the controls, the audience’s view settles into something approaching the traditional. The train window becomes a framing page, the views outside lines in temporary, haiku-like stanzas. Each line is a self-contained vision, but come together to create, in the words of the Imagist poet Ezra Pound, “an intellectual and emotional complex in an instant of time”. If the audience lean themselves out of the window, this two-dimensional plane is complicated: they see the sliding layers of words moving past in the simulated breeze, and may well catch other lines, hidden from within the carriage: the faint drama of the tiny creatures moving between the rails underneath, or tantalising, perpendicular scenes taking place on bridges over the track.

Generative text is something I have worked with before, as a halfway house between manual authorship and something completely automated. I have no interest in the more hands-off, data-heavy approaches of large language models and neural networks, which silo language into black boxes of training corpora. They seem to outsource most of what might be actually, humanly enjoyable about this process. Instead I used a familiar tool: the Javascript library Tracery, authored by the artist Dr. Kate Compton, which allows you to generate texts using lists of vocabulary placed into syntactic rulesets called ‘grammars’. It was the creation of these grammars, and the vocabularies to power them, that occupied the majority of the second half of the project.

I created a handful of generators, placed in the 3D scene to correspond to different layers of the landscape from the trackside to the horizon. Each had its own Tracery grammar to produce appropriate lines of poetry. These grammars began on paper: a series of drafted micro-poems that I completed on Post-It notes over a series of sunny mornings in the grounds of Exeter University. They were a first pass at trying to define what sort of vignettes might appear on such a journey through the English countryside, across ten hours one day in 1850.

What emerged from this exercise were indistinct fragments, each hinting at something familiar - ‘a ghostly moon glimmering like a tuppence’ - or something peculiar, tantalising or remarkable: ‘fifty slugs cannibalising a trifle’; ‘ten arms laid smoking on the gravel’; ‘Irishmen building a building in the shape of a tomb’.

To create the poetry generators, I first wrote large numbers of drafts on Post-It notes to tease out structures and types of vocabulary that I wanted to include.

From these specific, manual stories I could start the slow, inefficient toil of extrapolating structures for the grammars - the #adjective# #person# #verb#s the #object# - and the reams of vocabulary to fill these templates. Lists of celestial bodies, ingredients, materials, objects, types of room, types of building, locations of massacres by British troops, adjectives and verbs positive and negative were paired with constructions that generated lines not unlike those ones I had written by hand: moments that hinted at something more, some greater depth inaccessible to some stranger merely assing through at speed. As in any story, the most tantalising - the most coarsely narrative - were in those grammars where a #person# #verb_interact.s# with a #person#. Such things are the basis of any story, and will always draw the eye.

An example of the poetry generation code, written in the Tracery Library.

I could continue to work on these grammars indefinitely, and there is certainly still work to do to remove repetition, and some of the odder or more-awkward constructions that inevitably arise with generative techniques. Of course, my generated lines are protected from close scrutiny by their fleeting existence: each passed in front of the train window, at the simulated speed of the train itself, and then finally destroyed and deleted from the program’s memory once out of view.

However, as promised in my original plans, there are points in the journey when things slow down enough to really dwell on the words. The train stops at each station on the line for a randomised period of minutes, where the audience can see generated headlines from generated newspapers, and generated travellers embracing generated relatives. And this train, like all trains, is not immune to hold-ups. Each section between stations has a semi-randomised chance for a delay, in which the scene slows to a stop, sitting on the tracks for anything from a few seconds to ten minutes. The moving lines of poetry slow and stop as well, the constantly-shifting verbal contours of the landscape fixed in place, framed by the veneer of the window, in an entirely random composition.

Eventually the train creaks into motion again, and down the line the audience sees the source of the delay: itself the product of yet another Tracery grammar. Sometimes it is as prosaic as another train passing, or a broken signal limp on its pole. At other times, however, the audience is viscerally reminded of the raw, new gap between the local and the national in 1850. The difference between ‘railway time’ and local time can cause misunderstandings and accidents. Somebody tries to cross a level crossing at 11.50, ten minutes before the 12.00 train arrives: instead they collide with it thundering past, keeping London rather than Bristol time. The debris and evidence of such accidents must be cleared before the train can progress, crawling back up to top speed, past body parts strewn; overturned carriages; screaming horses. Such unexpected connections are why an ‘etiquette’ is needed. A world in which people thunder through old time and old space needs some way to manage the dramatic interactions between them and their new, discomfited neighbours.

Like most things I make, this was an ambitious amount of work for the timeline of the residency: I’m glad that I can hide behind its ‘prototype’ status on Github. The 3D scene design is more or less complete, and there is little that could be added to it that would improve the experience overall. The real work lies in iterating on the grammars that generate the poetry, both in the range of vocabulary that each grammar has access to and in the types of clause, sentence and paragraph that each produces. The more-than-occasional awkwardness, repetition and dullness in the generated lines were the inevitable result of the speed at which I was working, and can only be resolved with careful editing and manual creativity: a disadvantage of such a laborious approach for some, I suppose.

There is also an interesting tension here that I have not had the time to explore: a tension between the manual work of a storyteller, writing one specific set of stories, and a programmer specifying the rules and raw materials by which a machine can produce models of such stories, with no direct intelligence or close reference point behind them. It’s a generalisation and automation of process that the railways themselves introduced to the countryside in the 1800s; one that, like the railways in their own time, is vulnerable to knee-jerk criticism despite its surprising beauty and craft.

The next question is that of display. The work was deliberately created in portrait aspect to mirror the shape of a train window, but also to make it easy to recreate as a smartphone app. I like the idea of this endlessly variable train journey existing on somebody’s home screen: something to be distracted by, rather than from.

I’m also interested in the potential of public display. There’s a bare patch of wall in my university’s Humanities department café where a mounted screen would fit quite nicely. However, any relatively-public space where it can sit unobtrusively, piquing sporadic curiosity, would do: beside the men’s toilets on Platform 6 at Exeter St. David’s, for example.

Thinking about it, perhaps it would be better if it was slightly tucked away, only occasionally attracting the eye of a passer-by. The project isn’t really designed as a single story, to be experienced from start to finish; all ten or eleven hours of it. It isn’t something to consume. Like the view from a real train window, it’s a tool for occasional contemplation, and for reflecting one’s boredom; for fleeting double-takes, before the specifics of one’s own life, one’s own stories - this side of the glass - begin to intrude once again.

You can learn more about the project, and the Living With Machines research, here.

A few years ago I happened to visit Longleat Park with my family, in my Dad’s Orinoco-green Land Cruiser: all coffee-cream arse-heat and a roof like an observatory. Longleat, at its own centre, is a stately home; though its movement through time has been anything but stately. From its owner’s pornographic impastos in the drawing rooms to its transformation into a Middle-English safari park in recent years, I get the sense that the National Trust regularly, and collectively, wakes up sweating at the thought of it.

We were there, along with a parade of other brightly-coloured vehicles, to do something which I to this day find not only absurd but grotesque; we joined the slow queue along the tarmac as it wound past the entrance gates and into the woods, a queue no different in composition from the one in which we had previously sat, for an hour, on the A303. It stunk of caramel and onions inside the car, and I could see the snot-glint of grease on every haunch of leather; but of course we could not open the windows.

I had one of those apocalyptic, drifting instances when I realised and recognised what everything around me was constructed from, its heredity; from the seatbacks to the diesel in the engine, and the fat child’s proto-bicep pressed against the glass in the car behind us.

Luckily, this granularity did not last. The landscape we were moving through was former hunting land; curated by ewes, thick as a club sandwich, as unbroken as garnet. The woods were sparse and very old, and there was a tiger, predictably, shivering in its colour almost like a fox. The lion’s den was empty for the winter. The macaques had tortured all of the squirrels into the distance. There may have been a rhino but I may also be dreaming that, now; I see it parked beside a hedge full of rosehips, as appropriate as a battle-tank in Wiltshire, the syrup dripping off its armour.

The whole day was, by that point, made glum and gross, and we hadn’t even reached the fenced island of the gift shops yet. Longleat is not a circus, and the animals looked healthy, not badly-conserved. If anything, they looked faintly embarrassed to be there.

We reached a curve in the road; on one side were the woods, and the Park, and on the other an open, massive, slow basket of field, thick on its far side with ferns and the chain-link fence, keeping out the greater part of England. As we sat there idling, waiting for unbelievable tigers to cross the road, there was something happening, reconfiguring, in the dogtooth disruption of those ferns. I had been so used to the limited, blaring palette of Longleat’s charges that I had almost missed it. Eventually it resolved itself into a small, close-knit herd of deer, wandered through some gap in the fence from their territory outside the Park. In the centre, protecting and protected, was a not-so-enormous stag, with antlers in a petrified, almost-Celtic diagram. His fawns and does were flush against his coat, and even looking straight at them and knowing them I could barely pick them out in his fallow; could scarcely tell what was eyeball and what was autumn hide.

We had startled them, and they were rammed right up against the boundaries of our experience. Everybody in their cars turned away from the exotic obvious and watched, steering-locked, as the wild family unused to fences negotiated the length, looking for their way back into anonymity. They were the colour of the ferns at that time of year; the colour of the diesel in the engines, and old computers and brown meat around the wishbone and ash and bare branches and every grey sky of England that we had come there to forget. And yet we watched him and them, at the very edge of their freedom, resplendent in their camouflage.

The plane on which we all sat shifted, and I - and I think some others in their cars, behind and in front - felt a form of moral vertigo, the distinct feeling that humans feel when they see a thing within its nature, unaware, while we crawl along on wheels and eat coconut from the bag. Its a pre-Catholic, primeval sort of guilt; a recognition that past the glare and bombast of what we think we might like, the tropical and the dazzling, there might be a native alternative, a dunner dinner for our eyes; a distant clutch of ferns amongst seventy thousand others, chevroned like a prison uniform; and a doe’s eyes like her own pelt, hiding on the background.


My parent’s cat has just learnt (or, perhaps less impressively, just grown) to jump clear from the kitchen floor and into the fridge. I have an irresponsible habit of leaving its cheese-thick door ajar when constructing lunch, and for the past week or so with almost no fail I have heard the corn-grind of his nascent toenails on the lacquered wood, hoiking himself up and into the cavity.

At first it was laughable that this compact oik, this tiny hoiker, no longer than an Evian, with a rib-cage like an arm of bangles, should be able to leap eight times his own height. But, of course, on comes life: filling him up, changing his schematic, upgrading him, hardening his bones and his claws. Every day his face grows more and more concave, more SETI-like, though his enormous eyes, the colour of lager, never cease dominating.

He pips the pip of the pipistrelle when he is pleased to see me or is hungry, and his tail lashes like a cat’s. His tongue, when he deigns to display it, is as sweet as a banana chip. He eats like his own tapeworm, and has found the longest stretch of unbroken space in our house, an isosceles from the tip to the tip, and he runs it daily, religiously, back legs pulling forwards and to the right, threatening at every moment to throw him out of control. With all of these disparate likenesses in him, it is becoming no trouble to rally some of them to his cause and pull himself up into odd places, especially if they contain nice smells.

Usually I catch him before he can haul himself into the body of the fridge proper; he manages only a hit of melt-water, floating with spring onion skins, before he is deposited back onto the kitchen floor. Other times, however, my sandwich is complicated, and requires numerous trips, strategies, avid concentration, and I forget about him. And when I finish and come back to it all I can see of him is those eyes, rustic, hot and alien, weirding me out from amongst the darkness between the tofu and Crunch Corners.

In some ways this is adorable; a precious sight of a precious thing being obliviously capricious, reveling in breaching the rules which are only enforced half-halfheartedly. I sigh, remonstrate him all up on one level, in the same high, fairytale registers in which his natural prey speaks; this being the only way to get him to look at me, to turn his amber instruments my way. In that moment I feel as if I have just won some research time with an expensive telescope.

In other ways it is truly unsettling. Here on the greasy, misted shelves Teddy lurks (that is his name) in the vivarium: amongst dripping packages, the eeking plastic and the compartments, and in so doing takes on all the attributes of a commodity himself. He is lost amongst the other discrete objects in the fridge’s depths; or rather, is not lost but merely indistinguishable. It is no longer important that he belongs out in the warmth and our attention. He is no different in weight, distribution and unit cost from a pack of four chicken legs, a few tins of sweetcorn, a Tupperware of cold rice, or indeed an Evian.

If I squint, he disappears entirely.

What is worse is that these commercial attributes are not imagined; they are part of him already. It is only the sterilising airs of the fridge1 which bring him into focus. He has, of course, always been a commodity; there are many kittens like him, and there will be many more, and yet all of their owners will believe them unique. He is mobile within bounds: when needed, he is stopped and ceased to make an opportunity for photographs, a feature for guests, and a giver of pleasure. He was bought, quite materially, for a large amount of money. We were even fooled by his packaging, his blue fur like something synthetic and highly flammable, framing those unbelievable, high-resolution eyes. He is maintained with unguents, pills and pastes. He will grow obsolescent before other, less-rarefied cats.

I do not like this feeling. I do not like it revealing his ubiquity, his seriality, his place on the shelf. He was bred to be beautiful and gawky and loving. Soon we will clip his testicles off, as one takes off the ugly head of a carrot.

I do not like this feeling, and so I pluck him out the fridge always, as soon as I can, before he can charm his way into this carton of blueberries, fuzzy with their own, coquettish fur. I put him down, trill at him, have him trill right back, and shut the door. He goes to sleep in a perfect cursive. His eyes shut, impossibly. He loses his charge so easily.

  1. And isn’t it horrifying that it is always running, running, running, always chilling, just in case we may need one thing or another from it at any moment? Have you ever seen the back of a fridge? Have you ever felt the heat it exchanges at its back doors? Have you run your hands over the enormous swaddled pipes, like the legs of divers, the fins and panniers and heatsinks and gills? There isn’t a worse appliance in the house. It is doing it even right in this moment! Straining! Sweating! I hate it! 

Rob Sherman · 7y 20m 63d ago

"Suddenly it was very pleasant, sitting by the lochside across the way from Quinag, on a weekday morning, drinking strong coffee in weak sunlight. Kings do not have it better."

Andrew Grieg, At The Loch Of The Green Corrie