Re: open source windows

viro_at_parcelfarce.linux.theplanet.co.uk
Date: 09/11/03 

Date: Thu, 11 Sep 2003 11:25:44 +0000 (UTC)

In article <3F6057FE.5050100@yahoo.co.in>,
dilipkc <chhetri_inside@yahoo.co.in> wrote:

>off-course the effort is required, so my question now is which one
>requires more effort,
> a) Writing wine to run windows application.
> OR
> b) Modifying Linux kernel to give Windows API and Driver interface.

... and then maintaining resulting pile of *** through the kernel changes.
I take it that you volunteer for that job? Fine, it's your funeral, after
all.

Let me spell it out for you: there is no such thing as stable in-kernel
API. Of course nobody changes the interfaces (and all drivers that happen
to use them) out of sheer masochism, but when changes are needed, they
happen. Third-party drivers are either updated by their maintainers or
they stop working. It's as simple as that.

Keeping an API (A_B_I in that case) stable across kernel versions is not
an easy job and unless you know somebody who could do that and would care
to waste their time on that, it won't happen.

Since I don't need windows drivers (and don't trust hardware vendors to
hire decent programmers) - why the hell would I spend my time supporting
badly-written crap I don't have to use? And AFAICS it's the common situation
among the kernel folks.

Unless you are willing to do that yourself or pay a *hell* of a salary
to somebody who would do that for you - forget about it. And it won't
be cheap.

> If both are comparable then it's certainly wise to implement case "b",
> as it will enable you run Windows application better than wine.

Since I don't run Windows applications, why would I care about it? You
want it, you write it... Or pay up.