Xorg 1.5 missed the train?

[Les Mikesell]

Bill Crawford wrote:
>
>> What is it that would suggest that it is finalized to a manager that might
>> want to commit resources to writing a driver his company will have to
>> support?
> 
> So why should "Fedora" commit to supporting "binary driver Foo" before
> their release is ready?

That has nothing to do with what I said.  I'm suggesting that fedora 
should ship with interfaces that are publicized as standard, and allow 
time for changes in this standard to propagate before shipping something 
different from the standard.  This has nothing to do with supporting 
anything or anyone.  It is common decency in interaction.

>> No, it wouldn't be the same if that label had been applied and announced
>> publicly in time for others to coordinate with a shipping date.
> 
> You keep saying "publicly" but for the company concerned, who will
> have to make their driver work with the ABI in question, it being in
> the X server code base and discussed on the mailing lists (which they
> do have access to) is reasonable enough information for them to go on.

Do you have the authority to speak for them?  It just does not sound 
like a reasonable business decision to expect anyone to make.

>> Errr, how is that different from breaking?  Interfaces work or not.
> 
> It's very different. The driver concerned will continue to work with
> the server ABI it was built to work with. Noone has mandated that
> every X11 server in the world be updated tomorrow night at midnight!

If you mean that, you shouldn't be shipping one that makes this demand.

> In other words, it's old news to most, no one is forcing you to use
> this new server and "break" your drivers. Of course, you *could* have
> used hardware with open source driver support (that is updated already
> in the new X.org code base), but noone is forcing you to do so.

Unless you install fedora, which doesn't mention that it shipped a 
pre-release.

>> All I expect is a reasonable chance for others to coordinate.  This is like
>> shipping a power cord with a new plug style before announcing the matching
>> standard for the socket where you are supposed to plug it in.
> 
> Not so. In particular, noone is forcing you to change the existing
> sockets, and the new design was settled on quite some time ago. The
> people producing the plug are well aware of this process and have
> chosen to not update their plug design to match yet, ... and I have
> really had enough of these analogies even if you haven't.

It's not typical to update products to match a new standard before the 
standard in question is finalized, wireless-N notwithstanding.

-- 
   Les Mikesell
    lesmikesell at gmail.com