Realtime Games Software

Realtime Games Software were Ian Oliver, Grahame Baird and Andy Onions, and what they were good at was one thing: solid three-dimensional graphics on machines with no hardware for it.

Amiga Computing met them, and recorded where:

I first met Realtime at a ZX Microfair, a computer show so downmarket it bears more of a resemblance to Petticoat Lane than the Mayfair… Ian Oliver, Grahame Baird and Andy Onions are no more pretentious than the tatty trestle table they stood behind.

They had doctorates. The same profile notes “the men with the PhDs” and that “their mathematical skills have shown in a wide range of 8-bit games — Z80 stuff for the Spectrum and Amstrad CPC”, and quotes the entry they wrote for their own university rag magazine:

Realtime (Games Software) Limited — Great guys to go out for a drink with.

Fast facts

The three: Ian Oliver, Grahame Baird, Andy Onions.
Press evidence in this library: CRASH, ACE, Amiga Computing and Amiga Format, 1985 to 1989.
Speciality: filled-vector (“solid”) 3D.

Confirmed CRASH scores, from the issue 48 index:

Game	Score	Issue
Starstrike II	96%	28
3D Starstrike	93%	11
3D Tank Duel	83%	7

“You can’t do that on a Spectrum”

That sentence is the company’s entire reputation, and Amiga Computing reports it as the standing reaction to their work — games “which elicited the comment ‘You can’t do that on a Spectrum’, when others, like Ultimate, have fallen by the wayside.”

The technical claim is specific and worth stating properly. Wireframe vector graphics — lines only — were achievable on 8-bit hardware and several people managed them. Filled polygons are a different problem: every frame you must sort surfaces by depth, decide what is hidden, and shade the visible faces, all inside the time between two television frames, on a processor doing about three and a half million cycles a second. Starstrike II took 96% doing it.

The 16-bit crossing that worked

Realtime is one of the few 8-bit specialists whose skill transferred cleanly upward, because the thing they were expert in was exactly what the new machines were about to make central.

ACE, August 1988:

Realtime produced 3D Starstrike, 3D Starstrike II and the 8-bit conversions of Starglider, before bursting into 16-bit prominence with the oh-so-impressive Carrier Command.

Amiga Format in December 1989 places them in the short list of British studios who defined the vector generation — Braben with Virus, Singleton with Whirligig, Argonaut with the Starglider sequel, Paul Woakes with Mercenary, and “Realtime, authors of 3D Starstrike, wrote Carrier Command”.

Carrier Command is the payoff: a game Amiga Format describes as asking you “to conquer an island network using an aircraft carrier, fighter planes and amphibious tanks”, built out of the same filled-polygon technique, now with hardware that could carry it.

One line in ACE’s 1988 piece deserves keeping for a different reason:

Their next project for Telecomsoft will again use solid 3D — a field that Realtime pioneered and BT are swiftly monopolising.

That is a games magazine noting, in passing, that the British telephone company was cornering the market in vector graphics — see Firebird for how Telecomsoft assembled that position, and Hewson for what it cost a rival.

Carrier Command on a Spectrum

The 16-bit crossing above is the usual way to tell it. The more interesting fact is that Carrier Command went back down — and worked. ACE, July 1989, reviewing the Rainbird Spectrum release at £14.99 cassette, £15.99 disk:

The marvellous thing about Carrier Command is that the gameplay is just as brilliant on the Speccy as it is on ST and Amiga. However, it is not all sweetness and light, as it will only work in 128K mode.

The scores across formats make the point: ACE 927 for the Atari ST version, ACE 860 for the Spectrum. A game built to show what 16-bit machines could do, ported down to an eight-year-old 8-bit design, losing about seven per cent.

The line that landed on MicroProse

When British Telecom sold Telecomsoft in 1989, MicroProse inherited the catalogue — including this one. ACE, in the same July 1989 issue, put it about as bluntly as a magazine can:

MicroProse must be thanking their lucky stars they now have the publishing rights to excellent strategic software like Realtime’s brilliant Carrier Command. It wipes the floor with anything the Tetbury based outfit have come up with themselves in the last two years.

“The Tetbury based outfit” is MicroProse’s British operation at 2 Market Place — the same address that had been selling Ultima and F-15 simulators. A simulation specialist bought a telecoms company’s software arm and acquired, in the process, a strategy game better than its own.

Battle Command, and the move to Ocean

Realtime did not stay with Telecomsoft’s successor. The follow-up to Carrier Command went to Ocean, and ACE tracked it from announcement to release.

December 1989, on the sequel then in development:

Realtime Games Software’s long due, but eagerly awaited, follow-up to the gaming classic Carrier Command should be out next March. The game is to be called Battle Command… keen to stress that “it is not a tank simulation”.

The specification, as the studio gave it: around 30 mission scenarios before a “final encounter” the team would not describe, 18 months of development on mainly 16-bit formats, and improvements to what the magazine called “the unimprovable Carrier Command” — “far more complex solid 3D graphics with highly detailed objects, and a better gameplay balance of strategy and shoot-’em-up”.

By May 1990 it was “perhaps the most eagerly awaited tank game” of the year, and ACE names the publisher plainly: “Ocean’s Battle Command, the successor to Realtime Games’ Carrier Command”. It shipped on ST, Amiga and PC that summer, with Spectrum 128K, C64 and Amstrad CPC versions following.

⚠ Worth noting against the Telecomsoft material above. Realtime’s association with BT was a phase, not a home. 3D Starstrike came out through Realtime’s own arrangements, Carrier Command through Rainbird, and Battle Command through Ocean — three publishers in six years, with the studio’s value staying constant throughout. That is the practical version of the point below: when your asset is mathematics rather than a catalogue, you can change publisher without losing anything.

Why it mattered

Realtime is the counter-example to the story that the 8-bit generation could not cross to 16-bit. Most could not, and the reason is usually framed as failing to scale up from bedroom to studio. Realtime crossed for a different reason: their asset was mathematics, and mathematics does not become obsolete when the hardware changes. It becomes cheaper to execute.

A publisher whose value was a licence had to renegotiate it. A publisher whose value was a programmer had to keep the programmer. Realtime’s value was knowing how to sort polygons by depth quickly, and that was worth more on an Amiga than it had ever been on a Spectrum.

Solid 3D, from a trestle table