The Tape Load and the Patience It Taught

The Commodore 64’s built-in tape routine moved about fifty bytes a second. A game filling the machine’s usable memory therefore took something like a quarter of an hour to arrive, assuming nothing went wrong, and something did go wrong often enough that every household with a Datassette owned a small screwdriver for adjusting the tape head’s azimuth alignment — the angle of the head against the tape, which drifted, and which you corrected by opening the deck and turning a screw a few degrees while a game half-loaded. On the Spectrum the failure announced itself with R Tape Loading Error, four words that a whole country can still recite.

This was the normal condition of playing a game in Britain in 1985, and there is a strain of writing about it that treats the wait as a moral education. I want to argue something narrower and, I think, more useful: the tape load was a genuinely terrible piece of engineering that produced three real effects on how people played, two of which the medium has lost and one of which it should never have wanted back.

The numbers, briefly

Turbo loaders fixed the worst of it. Commercial publishers stopped using the ROM routine almost immediately and wrote their own — Novaload, Freeload, Cyberload on the Commodore, Speedlock and Bleepload on the Spectrum — pushing throughput up by a factor of five or more and dragging a load down to three or four minutes. They also, cannily, played music while they worked, which is how the loading screen became a stage rather than a wait.

The disk drive was barely a rescue. Commodore shipped the 1541 with a hardware fault in the C64’s serial implementation that dropped it to a few hundred bytes a second, so a floppy drive costing more than the computer loaded a game in a couple of minutes. Epyx sold the Fast Load cartridge in 1984 to fix Commodore’s mistake, and sold a great many.

So the tax varied — fifteen minutes at worst, three at best, plus a failure rate. Now the effects.

The counter and the multi-load

The three-digit counter on the front of a Datassette was a memory device with no connection to anything. It counted revolutions. It had no relationship to time, position or data, and it reset when you felt like it — which meant its numbers were personal. You wrote them on the inlay card in biro: 000 for the game, 247 for level two, 512 for the bit you never reached.

That mattered because of multi-loading. A game too big for memory arrived in pieces, and each piece came off the tape when you got there — so reaching level three meant the machine stopped, told you to press play, and made you wait again, mid-session, having already earned it. Dragon’s Lair on the C64 in 1986 is the notorious case: a laserdisc’s worth of animation redistributed across a cassette, so a player paid minutes of load for seconds of play, over and over. It is possibly the worst value-per-minute anything in the medium’s history.

Multi-load also produced the era’s strangest design pressure. A level break was expensive, so designers fought to keep the game in memory — and a very great deal of what people admire about eighties design is downstream of that fight. Tight loops, reused assets, levels built out of recombined tiles, a single screen of rules explored to exhaustion. The constraint that made playing miserable made designing rigorous, which is an uncomfortable thing to be true and is true anyway.

It made you commit

You got one game per evening. That is the whole thing, and it explains more about eighties play patterns than any amount of talk about difficulty.

When entry costs ten minutes, a player does not browse. You choose before you commit, and having committed you are in — you will play the thing you loaded, because the alternative is another ten minutes. So you learn it. You get past the bit you are bad at, because quitting is more expensive than persisting. A modern player with a two-hundred-game library and instant switching bounces off a game at the first hostile twenty minutes, and this is entirely rational behaviour on their part, and it means a great many well-made games never get the third hour where they become good.

I do not think this makes eighties players better. It makes them trapped, and trapped is a decent proxy for committed if you squint. The medium replaced the trap with a market, and the market punishes any game whose argument takes ninety minutes to state.

It made failure expensive, and so made you study

Impossible Mission on a Commodore 64: thirty-two rooms, a search you cannot rush, and a load between you and another attempt. So you did not die at the design and shrug. You sat there, and you read the room — where the robots patrol, what the somersault clears, which terminal is worth the walk — because the alternative was watching a tape counter climb for four minutes. The somersault and the search puzzle hold up now largely because they were built for a player who had been forcibly made to look at them.

This is the effect worth keeping, and the good modern designers keep it deliberately. FromSoftware’s whole method is the reintroduction of a cost on failure — the run back, the lost souls, the bonfire reset — and the result is exactly the eighties effect: players who look at an encounter, because the alternative is expensive. The industry spent two decades making death cheap and then rediscovered that a consequence is what turns a fight into a text you read. Difficulty is a design choice, and this is the honest case for it: attention is bought with cost, and nothing else buys it.

The tape bought it by accident, and charged the wrong currency. FromSoftware buys it on purpose and charges in fiction. The second is better in every respect.

It made the manual hour

Here is the effect nobody talks about, and it is the strangest.

You could not stare at a loading bar for fifteen minutes. So you read the box. And publishers, knowing exactly what that window was worth, filled it — which is why eighties manuals carry fiction rather than key bindings. Elite shipped in 1984 with The Dark Wheel, a novella by Robert Holdstock, a real novelist with a World Fantasy Award to his name, commissioned to write a piece of prose that would be read entirely during tape loads. A universe compressed into 32K came with a paperback because the compression left nothing on screen to explain it, and the compression left a fifteen-minute reading window to explain it in.

Look at the feedback loop. The hardware could not render a world, so the world went in the manual; the hardware was slow, so there was time to read the manual; the manual built the world in the player’s head, so the sparse graphics on arrival were legible. The tape load was load-bearing in the rendering pipeline. Three lines and a triangle became a Cobra Mk III because you had spent eleven minutes reading about one.

That loop is gone and nothing replaced it. Modern games explain themselves during play, which means explanation competes with action, which means it loses. Tunic’s central trick is to notice this and invert it: scatter the manual through the game as collectible pages, and the reading becomes the reward. It is the only game in years to have found a legitimate use for the artefact, and it did so by making you earn it.

What patience actually was

Strip the nostalgia and the tape load taught nothing about patience. It taught helplessness with a soundtrack. The virtue people remember was a hostage situation, and everyone who says otherwise has forgotten the specific misery of a R Tape Loading Error at minute nine of a game borrowed from a friend who wanted it back on Sunday.

What the tape did was impose three conditions — one game at a time, a real price on failure, and a quiet window with reading material in it — and those conditions produced a kind of attention that the medium’s own improvements have since dismantled, one convenience at a time, each of them individually correct.

The conditions are reproducible without the transport. A game can be long enough to need commitment. Failure can cost something. There can be a text worth reading and a moment engineered to read it in. The C64 catalogue that still holds up mostly holds up because it was designed for a player who had been forced to pay attention, and a designer today has to earn that same attention with nothing but the work. That is a harder job, and the games that manage it are better games than the ones that had a broken tape deck doing the job for them.

Where to play them: emulation gives you the whole archive with the loads set to instant, which is the correct setting and quietly removes the thing this essay is about. Turn the load speed back down once, on one game, to see what the design was shaped around. Then turn it off again.