Can an API be copyrighted?
Alan Klietz
For example, assume that I write MY-DOS, which is functionally
equivalent to the published specifications of MS-DOS. If I try
to sell MY-DOS am I in deep weeds, legally speaking?
Note that I'm talking about *programming* interfaces, not GUIs.
--
Alan E. Klietz
Minnesota Supercomputer Center, Inc.
1200 Washington Avenue South
Minneapolis, MN 55415
Tel: +1 612 626 1737 Internet: al...@msc.edu
Fax: +1 612 624 6550
Richard Brett Frankenberger
>Is it possible to copyright an application programming interface?
>For example, assume that I write MY-DOS, which is functionally
>equivalent to the published specifications of MS-DOS. If I try
>to sell MY-DOS am I in deep weeds, legally speaking?
>
>Note that I'm talking about *programming* interfaces, not GUIs.
>
>--
>Alan E. Klietz
DISCLAIMER: Don't confuse me with a lawyer, and don't confuse the following
with legal advice.
Essentially, BIOS is just an API. You make the certain BIOS calls and BIOS
responds in a certain way. Currently, lots of companies (Pheonix Technologies,
American Megatrends Inc, etc.) make BIOS that are compatible with (have the
same API as) IBM's PC BIOS, but do not infringe IBM's copyright. (IBM
tried very hard to protect its BIOS - lots of people lost lawsuits against
IBM after trying to write a compatible BIOS, but I believe phoenix technolgoies
was the first company with the resources to fight off IBM's massive legal
staff, and after they won, then lots of people came out with compatible
BIOS's. )
Also, DR-DOS ... did they get a license from microsoft? If not, then that
pretty well proves that you can clone DOS legally. Also note that a micro-
processor's instruction set is basically an API, and Intel's 80x86 line
has been cloned (AMD has a license, but CyRix did it without a license by
not using any of Intel's microcode, but duplicating their instruction set/
API).
From these examples, I would tend to conclude that an API (at least a low
level one) cannot be copyrighted. I tend to think that it might be the same
for, say, the microsoft windows API (It just wouldn't be proftiable to clone
windows since microsoft practically gives it away anyway).
- Brett (rfra...@cs.umr.edu)
Bruce Hayden
>In <1993Jan31.0...@uc.msc.edu> al...@af.msc.edu (Alan Klietz) writes:
>>Is it possible to copyright an application programming interface?
>>For example, assume that I write MY-DOS, which is functionally
>>equivalent to the published specifications of MS-DOS. If I try
>>to sell MY-DOS am I in deep weeds, legally speaking?
>>Note that I'm talking about *programming* interfaces, not GUIs.
>This area of copyright law is a bit fuzzy these days. On the one hand,
>recall the Lotus v. Paperback case, where Paperback's only sin was
>making something that looked the same from the user's point of view
>but which (we assume) was written without reference to the source code
>of Lotus. Paperback was driven out of business.
The judge in that case tried to distinguish between tying up the
interface, and what Paperback Books was doing. I think that was
one of the real weaknesses in the case - you can't make that distinction
and thus the judge started us down the slippery slope of copyrighting
interfaces.
>On the other hand, there are cases going the other way.
Quite a few more, especially in the last year. I believe the 2nd, 5th,
9th and Federal circuits have jumped on the Altai bandwagen. One of the
externalities screened by Altai are interfaces.
>That's the fuzzy area for copyright. If you want a more complete list
>of things to worry about, there is also patent law and the Lanham Act.
>If somebody has, say, a patent that is automatically infringed by any
>computer that is running something execute-alike with MS-DOS,
>then people selling execute-alikes have to worry.
>And if the "look" of MS-DOS indicates its origin (e.g. Microsoft)
>then execute-alike software might confuse consumers (Microsoft's lawyers
>might argue) leading to a charge of false designation of origin ...
I have questions about the validity of any such patent. The Lanham ACt
possibility is more promising. I wonder if you could make an arguement
that conformance to an internal interface specification would tend to
make the modules on the other side of the interface believe that the
module was actually the copied module (and thus a false designation
of origin). Before we get too carried away, I should remind you of
the reverse Lanham Act finding in one of thee reverse engineering
cases this fall (Sega???).
Bruce E. Hayden
(303) 758-8400
bha...@csn.org