Re: Where we went astray-- Antitrust and Microsoft

On Sun, 2 Nov 1997, John H. Lederer <johnl[_at_]ibm.net> wrote:

<Good points about marginal costs snipped>

> I have considered two modifications of software copyrighting:
>
> (1) require disclosures of the source code. much as we effectively or
> particularly require disclosure of other copyrighted material. This
> might have a very salutary effect. For instance, IBM abandoned
> supporting Win32 programs in OS/2 as they had Win16 programs because of
> the conviction that Microsft would make chnages in the Win32 API's that
> would be hard to keep up with. I think, that with disclosure of the
> source code, it would be much easier to make functionally equivalent
> OS's.

This wouldn't work because much of the value of a piece of software is in the trade secrets it embodies.

You simply are nuts to attempt to "reverse engineer" the internal components of an entire OS like Win95, so the only way you are going to reproduce the software (in an economically viable way) is to have a well-defined, relatively static API, or to have access to the source code.

You do not need to give out your trade secrets in your OS to provide a complete description of your API (the interface to the internal operations of the OS) to outsiders. The problem is that Microsoft has apparently used information not published in its APIs when writing its own applications. They used trade secrets in their OS to benefit their application software, and this issue is whether this is anticompetitive...

Forcing MS to disclose their source code **would** solve this problem, but I think it would destroy the software industry since once you see how a person does something in software, unless they have that idea patented, its usually trivial to reproduce legally (that is produce with the same performance and functionality).

Forcing the publication of internal source code would also be a disaster for any software publisher as it would encourage other folks to write software that relies on or uses internal hooks that are not designed for others to use. If you've written any significant amount of software, you'll know exactly what I'm talking about (program to an interface, not an implementation).

> (2) change the term of copyright for software

I'm all for that. Copyright is intended to incentivize (and reward) authors for their work. Most software is of little economic value to the author after a few years. BUT, I'm really not sure that shortening the term would make a big economic impact for that reason. HOWEVER, This would help greatly the free software movement, which is more responsible for the success of the Web and the Net than any of the latecomers like Netscape or Microsoft.

     -Gabe

(speaking for myself only)

Gabriel Wachob, J.D.                 http://www.aimnet.com/~gwachob
COO, Findlaw                         http://www.findlaw.com