While centered on vector games, it also mention analog computers and the PDP-1
In addition to this, a small primer to x/y display types. We may discern the following fundamental types of analog displays (in increasing order of complexity and cost, and also chronologically):
X/Y displays, also known as point plotting displays or “animated” (sometimes also “painted”) displays. Sometimes also referred to as random-scan displays, as in random addressing of display coordinates as opposed to sequential. While all random-scan displays are x/y displays of sorts, these are the original ones: A single dot, given by x and y coordinates is activated. Often, but not necessarily, there is also a z component for an intensity. (Intensities are often just handled by varying activation time.)
These are the least expensive type of random-scan displays as there is no special hardware involved, other than the hardware for converting digital coordinates into analog deflection values, and cooling coils in order to stabilize the image, as the beam moves from dot to dot (which may involve quite swift movements). A display operation was simple fire and forget, with any recurring dots having to be refreshed by the program in a loop. (Hence the name “animated display”, as a program would animate a dot, which would then gradually decay over the next few seconds.) On the plus side, as there are no storage requirements, while any type of memory was still very expensive, this allowed for rather high resolution (like 4096 × 4096). On the other hand, these displays were typically slow, especially for the need to pause between any consecutive dots to allow the beam to stabilize for a sharper image. The CRTs were often repurposed radar tubes as used for “plan position indicator” or PPI scan-radar. Typically, these used a phosphor of long sustain/persistence, which also helped in stabilizing the image.
A prominent example for this is the DEC Type 30 Visual CRT display for the PDP-1, as used for the original Spacewar! game. (This one is a bit special, as it used a dual phosphor coating, known as P7 phosphor, which provides a bright blue initial activation of short persistence, which could be used to trigger a light pen, and a long sustain phosphor of yellowish green, which decayed over several seconds. This also added a time component to the display and dedicated photographic filters were available to isolate either the particular moment or the entire timeline.)
Vector displays in the stricter meaning, also known as calligraphic displays. These were essentially the same as the dotted displays described above, but using a more complex control hardware including storage. These could actually display strokes as in “vector” by continuously illuminating a path from stored point to point. Stored coordinates and their properties were visited in a loop by the controller, liberating the CPU from having to refresh these coordinates in the main program. This also involved a set of commands for storing or clearing a dot or stroke, which may involve an entire language used by the display controller. Some, but not all tubes also involved multiple cathodes for RGB color. Later, more common tubes, which could have been also used for raster displays using a different yoke and controller, were used. The controller might have been part of the display or a separate unit, or even a card in the main unit.
Prominent examples are the display controllers for the PDP-8, PDP-11, PDP-12, etc, and – of course – typical vector arcade games.
Finally, there were also storage tubes, which were storing the image by reactivating the phosphor autonomously based on electromagnetic properties of the tube. No special hardware is required to refresh the stored image, as it’s the static activation of the phosphor itself, which is used for the refresh. Tubes may include multiple planes, which may be addressed and cleared individually, which also required special hardware to control these tubes.
This was typical Tektronix territory – and rather expensive (probably more epensive than the computer which was used to generate the diplay commands).
I find these often confused. It’s also important for what you could do in terms of generating an image (AKA platform considerations). The first two types are prone to flicker. Point-plotting displays also are not usable for any kind of filled areas, as every point has to be displayed (and refreshed) individually. Since this is a slow process, you want to keep the total number of dots displayed low. (However, if you optimize for this and do not require multiplication and division, like in rotations, you can still run a game at 60 fps, even on 1960-hardware.) Vector displays are more versatile and speedier, but the feasible number of coordinates is still limited, otherwise, your image will flicker, when the display list becomes too long and the refresh loop can’t keep up with the decay of the phospohr activation. (This combines with another limit set by the very construction of the tube, where swift and fast strokes often rquire higher voltages.) Storage tubes, however, have no such limit to the complexity of the static planes, only the initial build-up of the image requires runtime. If your display supports multiple planes and your scene consists of mostly static backgrounds and chrome, with just some elements being subject to change on a live plane, the sky is your limit.
I wanted one of these so bad: https://www.youtube.com/watch?v=h1aMnTzTXEo