On Tue, 18 Dec 2001, Patrik Stridvall wrote:
Note that I said "_small_ improvements" because of the
modular nature of Wine. If the improvements are big, the DLL
separation would allow them to keep those changes proprietary.
I don't think small improvements is a problem anyway and
beginning an implementation of say DCOM is probaly not a
small improvement and DirectX certainly isn't.
But if it's not, it's within a small constant factor to
replace the rest
with prorietary code and not release anything. See, if the
DLL is mostly
implemented, the changes are not that big, so they should be
contributed
back. If the DLL is mostly stubs, just rewrite the entire
thing, you're
not wasting that much effort in the first place, and you can choose
whatever licence you want. DirectX and DCOM fall in this cathegory.
I have use the Crypto API as an example several times that to illustrate
that the DLL:s isn't a relevant protection boundary for the LGPL,
even Alexandre seems to accept this.
In any case if the DLL is mostly implemented there is very little
reason for a company to try to enter the market, it is simply to
dangerous, so I don't see the problem.
I fail to see _any_ moral/ethical/philosophical problem with
this. Do you?
Perhaps you have been more enlightend now.
Focusing on obscure issues only is anything but enlightening. :)
Well, the concrete issues like do we scare potential companies
that wants to help way quite hard to argue about in a objective
way and I do not really like subjective arguments.
I leave them to for example Gavriels that IMHO has done a good
good job at that.
I have been trying to approach the problem from the other side,
that the LGPL doesn't provide much protection at all.