Improving copyright

[Niall Douglas]

On 14 May 2004 at 19:22, João Miguel Neves wrote:

> > Which is precisely the same logic why people continue to use
> > Windows. Unless there's an obvious and compelling reason to change,
> > the incumbant has a major advantage.
> > 
> Fortunately those reasons are appearing and growing (viruses and
> license restrictions).

On a technical level, Windows should be more secure than Linux as it 
uses a fine-grained ACL security system. Unfortunately "should" isn't 
"is" :(

Both systems are equally broken with respect to good security. See 
http://www.eros-os.org/

> There are some developments like the ndis wrapper driver (which I
> think is a bad idea, but we obviously have very different views on how
> quick users can drop a platform).

Techie types can drop platforms pretty quickly but non-techies will 
put up with surprisingly large amounts of pain before they'll change. 
If you feel out of your depth about something, you'll tend to 
overvalue the familiar - it's why many MacOS fans are so diehard 
about the issue.

> > 2. Few Windows games will run under WINE. And this is unlikely to
> > change as DirectX demands a certain driver architecture Linux isn't
> > moving towards.
> 
> You do now that there is a DirectX implementation of Wine and that the
> latest work in X.org is improving in taking advantage of the hardware.

Absolutely. However unless they implement binary compatibility with 
Windows graphics card drivers they are dowsing a fire with petrol. 
Graphics cards evolve so fast even Microsoft has trouble keeping 
DirectX up with them.

As I mentioned before, the Linux driver model isn't the Windows one 
by any measure. Therefore to use Windows drivers they'd have to do 
quite a lot of emulation which attracts a heavy speed penalty. Not 
what gamers want!

> > 3. No matter how wonderful WINE becomes, running Windows
> > applications under it will always be a bit of a black art. Most
> > people just want to shove in the CD and make it go. This is why
> > Linux is unsuited to replace Windows in non-server arenas
> > 
> I don't get why this is impossible? Binfmt and things like nautilus
> together with wine can do that behaviour without any problems. I don't
> understand why you say it isn't possible.

Are you seriously claiming that Joe Soap who doesn't even know what a 
command line is will ever be able to run Windows binaries 
transparently on any Linux ever?

Of course it can be done - like Apple did for legacy MacOS app 
support. But it's very tough - like pushing square pegs into round 
holes. And far more effort than simply cloning Windows. If you want a 
comparator, look at cygwin on Win32 - nearly a Unix environment, but 
no Joe Soap will ever be able to run an unmodified Linux binary on 
Windows via cygwin.

> > > And why do you think reactos will go further
> > > than wine on implementing win32?
> > 
> > Actually ReactOS incorporates WINE, they simply provide a "native"
> > implementation layer - instead of calling Linux functions, they call
> > reimplementations of the NT kernel API.
> 
> So they'll also have to help develop DirectX for Wine to that. Nice to
> know...

It certainly is most interesting. I see no point in upgrading from 
Win2k ever except I have the horrible feeling MS will make the newest 
MSVC require a newer version. If I could break from MS operating 
systems I would.

Cheers,
Niall




-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 208 bytes
Desc: not available
URL: <http://lists.fsfe.org/pipermail/discussion/attachments/20040514/ffb753e1/attachment.sig>