I'm a fervent user of the Alternate Multiplex Interrupt Specification for 86-DOS systems, a method using a common set of multiplexer interfaces over interrupt 2Dh. It was originally developed in the early 1990s by Ralf Brown, who is better known for compiling and releasing the great Interrupt List.

Now the website of Ralf Brown's Files only contains two downloads specifically involving AMIS: AMISLIB version 0.92 dated 1995-09-24 and the AMIS v3.5 specification dated 1992-09-13. The contents of the specification proper appear to be found in the Interrupt List too, with the addition of some programs' private functions (function numbers >= 10h). The Interrupt List (release 61) also appears to be the only place to contain the new AMIS v3.6 function 06h, Get device-driver information.

All in all, among these files the AMIS versions referenced are v3.0, v3.3, v3.5 and v3.6. The question is: Were there any other historical versions? And is there any material on the known versions that I did not list here?

  • 1
    Have you looked at the Wayback archive?
    – Jon Custer
    Nov 9, 2021 at 23:25
  • 2
    @JonCuster I did. A quick vgrep shows that the earliest page they have (from '99 IIRC) is the same as the current page. Might have missed something but it looks identical. Nov 10, 2021 at 3:28
  • 1
    | sed 's/vgrep/vdiff/' (Little typo but it's been bugging me) Nov 10, 2021 at 12:43
  • 1
    I figured you had, but always a good thing to ask. And I say that because I totally spaced on the Walnut Creek products, of which I have a number. Well, it has been a while since I actually laid eyes on them, but they are around here ... somewhere. Sigh.
    – Jon Custer
    Nov 10, 2021 at 22:39

1 Answer 1


Were there any other historical versions?

Yes, there were, but the history of AMIS is just a little too old to be extensively archived.

The initial proposal dates back to September 1991 in release 27 of the Interrupt List; this is mentioned in the announcement of version 2 of the proposal:

In the September 1 release of the interrupt list (INTER27), I made a proposal for an alternate multiplex interrupt with a standardized method for sharing multiplex numbers and a set of standard function calls. After some discussions and useful suggestions, I now have the following revised proposal.

Version 2 suggested using interrupt 0x7D.

Walnut Creek’s May 1992 Garbo archive CD-ROM contains release 29 of the Interrupt List, which documents AMIS version 3.2.1. This already has two known signatures for software not written by Ralf Brown.

Walnut Creek’s June 1992 Simtel-20 MS-DOS archive CD-ROM contains AMISLIB 0.83, based on version 3.4 of the spec; however the library doesn’t include the spec itself. This contains a changelog covering version 0.80, 0.81, 0.82 and 0.83 of the library, suggesting that at least those were distributed. The same CD-ROM contains release 30 of the Interrupt List, which documents AMIS version 3.4.

I get the impression from the announcement of AMISLIB version 0.81 that the first publicly-available version of AMISLIB was 0.80:

I have just released an update (v0.81) to yesterday's preliminary version of AMISLIB, a library for implementing TSRs which comply with the Alternate Multiplex Interrupt Spec I've been propounding for the past several months.

Other archive CD-ROMs likely contain other releases of at least the Interrupt List, with different versions of the spec, although since version 3.4 was released by June 1992, finding older versions is likely to be difficult (Walnut Creek at least only started in 1991, and I don’t know how many DOS-related archive CD-ROMs they published before mid-1992).

And is there any material on the known versions that I did not list here?

AMISLIB 0.92 itself contains version 3.5.1 of the spec; the differences with version 3.5 are that bit 6 of byte 0 in the hotkey list is used, and the list of known signatures is significantly expanded.

  • 1
    Thanks, good work! I collected the relevant files from those two collections in pushbx.org/ecm/test/20211110
    – ecm
    Nov 10, 2021 at 12:28
  • 2
    While looking at the RBIL 29 and 30 releases I stumbled upon the old int 7Dh mention, which actually survived up to RBIL 61. It states: "[obsoleted proposal] - ALTERNATE MULTIPLEX INTERRUPT Note: this interface has been moved to INT 2D; there are no known implementations on INT 7D".
    – ecm
    Nov 10, 2021 at 12:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .