PLEX86  x86- Virtual Machine (VM) Program
 Plex86  |  CVS  |  Mailing List  |  Download  |  Computer Folklore

big endian vs. little endian, why 1954


(Eric Chomko) wrote, in part:

I would *like* to say that it is because they manufacture computers in Israel and the Arab world, where the language in use is written from right to left, but numbers are still written with their most significant digit on the left, so that in these countries they make computers that are intuitive for them.

big endian vs. little endian, why 1957
Yep- I have 8 Powerpc sbc's running the lab, console access is by rs232, 9600N81. The great thing about rs232 is its nice & simple, which is important...

Unfortunately, that is not the case. It is true the Intel 8087 was designed in Israel, but that isn't why the 8008 and successors were little-endian.

As has already been noted, some computers with a 16-bit word length, when handling 32-bit integers, placed the least significant word in a lower memory location, so that they could somewhat more simply fetch the least significant part first, and then carry later. The Honeywell 316 computer is one example of this.

But such computers still stored character data in 16-bit words so that the first character was in the leftmost or most significant part.

Then, one fine morning, someone at the Digital Equipment Corporation got a brilliant idea.

big endian vs. little endian, why 1955
Because the 8008 was designed by semiconductor designers not computer peoplem, and they did things to minimise the silicon used. In this case...

We're already making old-fashioned 12-bit and 18-bit computers. Now, the world is going to the 8-bit character. Let's make a 16-bit computer that is not old-fashioned - people who want an old-fashioned 16-bit computer can already buy a Honeywell 316 - we'll make one that is new and exciting.

And they did. The PDP-11 architecture was very innovative.

And somewhere along the way they had *another* brilliant idea. No, we won't add extra transistors to our design so that we can add 1 to the address, and fetch the least significant part first. Because we can be consistent *without* adding an extra transistor! Yes! We'll just put the second character of two characters in a 16-bit word in the most significant part, so that the four characters in a 32-bit number will be consistently backwards!

And, lo and behold, the idea of consistent little-endian computing was born. That the idea was so bizarre and confusing that it was not understood by the engineers who then made the PDP-11's big-endian floating-point unit does not deprive DEC and the PDP-11 of the glory - or, perhaps, the blame - of inflicting this upon the world.

John Savard Usenet Zone Free Binaries Usenet Server More than 120,000 groups Unlimited download


List | Previous | Next

big endian vs. little endian, why 1955

Alt Folklore Computers Newsgroups

2004 8051 unit sales 1953