2

The 8-bit microprocessors invented in the seventies, had a 16-bit address space. It didn't take long for memory demand to exceed this, with the result that bank switching was a fact of life for the remainder of the lifespan of those architectures.

Many different bank switching schemes were used. The one I'm interested in is the flexible and logically simple one where the 64K address space is divided into pages of size X, each of which can be independently mapped onto any X-sized and aligned section of physical RAM. This was used by the CoCo 3 (8K granularity) and according to the answer to this question, also by the Cambridge Z88 and Enterprise 128 (16K granularity): Which memory management is easiest to program - CoCo 3 or C128/Apple //e?

I imagine someone must have been awarded a patent for that scheme at some point, but Google searches on relevant keywords are just turning up patents on somewhat related but different things. I might guess it would be patented around the early eighties, but one thing I have learned in studying the history of computing is that most things go back a lot further than you think.

Was that bank switching scheme ever patented, and if so, when?

7
  • 1
    @TimLocke I'm sure it would, if anyone can find the patent! As I said, I have only been able to find patents on somewhat related but different things.
    – rwallace
    Jan 17 at 17:34
  • 2
    Many companies used it in their computers so I'd doubt the idea was patented, or perhaps it was patented decades before on early computers and the patent had run out. I think I've read DEC used it on the PDP-8.
    – Tim Locke
    Jan 17 at 17:38
  • 4
    @TimLocke -indeed, Wikipedia and other sources indicate that the PDP-8 had a memory expansion unit that could address various numbers of 'fields' depending on the model. See, e.g. gunkies.org/wiki/PDP-8_Memory_Extension_units
    – Jon Custer
    Jan 17 at 19:41
  • 2
    @JohnDallman With patents, the weirdest things have happened. So, I wouldn't be surprised to actually see a patent matching the OP's criteria. But anecdotally, around 1984 my brother designed a Z80 system with bank switching, and wasn't sued, while he had to fight against graphics-card related claims a few years later. Jan 18 at 11:06
  • 1
    There are at least related patents such as Tandy’s own EP0123248A2; there probably are patents with claims covering this type of banked memory access, but they might not have been enforced (especially in the 8-bit micro world, consider the time taken to grant v. the lifetime in the marketplace). Jan 18 at 12:52

1 Answer 1

8

For the origin of fixed size pages, and independently mapped pages, you need to go back at least as far as the Manchester Atlas machine of 1962. A historical perspective of both the Atlas ad the earlier Mark I can be found here.

In that article it mentions that Atlas memory was divided into fixed size pages of 512 words each. The idea of separating virtual address space from physical address space and mapping one onto the other is mentioned in the article as having been discovered during the earlier Mark I project, going back as far as 1949.

As to why these concepts were not patented or otherwise protected, that might have to do with British intellectual property law at that time. And any attempt to patent paging or page mapping after 1965 would have shown the Atlas to be prior art. (Thanks, tofro).

1
  • 3
    In a strict sense, this doesn't answer the question. It simply lists an example machine that used banking technology that apparently wasn't patented. That's true for a lot of machines, including the ZX Spectrum. BUT: It shows that anyone trying to patent banking technology would really have a hard time because of the "prior art" argument.
    – tofro
    Jan 18 at 14:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.