Why not make computer chips bigger rather than transistors smaller?

Her name was Hopper, not Hooper. Her lecture included showing the audience a “nanosecond” (11 inches of wire) and contrasted it to a “microsecond” (a 982-foot spool).

Another aspect would be that it would be a temporary fix.

Let’s say that Intel decides to make a Big Boy processor, but wants it to fit on motherboards that fit in standard desktop boxes. For argument’s sake, let’s say that with all the wires and connections and heat removal, the chip can realistically only be 8x the size of current chips and can hold ~8x the number of transistors and the like.

Because of greater distances, not enough research into larger sizes and ineffient use, I’ll SWAG that its 4x as fast. Nice, but not revolutionary. However, after your standard 36 months, small processors have caught up to it speed-wise.

At this point, for the Big Boy to keep up, it has to either grow in size (and now start to require specially designed boxes), or to start to miniaturize, which is what I think this is trying to avoid in the first place.

So you essentially only get a one-time significant expansion in size. Another way to look at it is if since the introduction of the 386, if the chip makers used only size expansion, a processor with modern speeds would be quite large and may not fit within a standard enclosure.

And the “bug” was found in Eniac, which had a crew of women swapping out replacement tube-laden PCBs on a routine basis.