RE: [AD] Proposal to drop DOS port from Allegro 4

[Shawn Hargreaves <shargreaves@xxxxxxxxxx>]

Clayton Weaver writes:
> I don't even use DOS at all. I don't know anyone personally 
> that does. I think you'd be hard pressed to find a non-programmer 
> (ie, computer geek) that does because he wants to. 

For desktop use on a PC (whether amateur game coding or just
game playing), that is certainly true. But the world is a much
bigger place than just people using PC's for fun. Just because
these uses don't happen to apply to you at the moment, that is 
no reason to discount the many applications in other areas such
as embedded systems, scientific research, and so on.

> The silly thing about wanting to keep DOS support as 
> functional as the rest of the library is that you are very 
> potentially limiting what the library can do.

This argument would be a lot more convincing if someone could
give specific examples of things that are limiting about DOS.
ie. a feature that is important to support in Allegro, not
possible to do in DOS, and cannot be optionally exposed just 
on the platforms that do support it.

Please be specific. It is very hard to come up with counter 
arguments to vague phrases like "potentially limiting" :-)

> Those continuing to compile only for DOS will find that 
> most people do not play their games.

But not everyone is making games, and not everyone cares about
getting other people to play them. For instance the people who
made http://www.pinball.wms.com/pinball2000/home.html wouldn't
have used Allegro if it didn't support DOS.

> In conclusion, DOS developers: stick to Allegro 4.

But then, what about the 90% of new additions that also apply
to the DOS version (eg. pretty much everything except the 
hardware driver layer), which DOS developers would be unable
to take advantage of? 3.12 is also a perfectly good library,
but it would be shame if DOS developers were unable to upgrade
past that to get new things like the Unicode support, 
especially considering that the Unicode routines are completely
platform independent, and that it took exactly zero effort to 
make them support DOS along with the other platforms!

I think this whole thing is being blown up way out of proportion
(and voting on it is _not_ a good idea: the DOS port would 
deserve to continue even if only 1 person was using it, as long 
as this did't cause any problems for the other platforms).

Rather than kicking up such a big fuss about removing features,
why not just concentrate on adding new things to the platforms
that you do care about? (so things move forwards, rather than
back). If these features can be added in a portable way, make
them portable. If not, add a caps flag to indicate what platforms
support them. If you can't be bothered to implement something
on DOS, don't implement it. Maybe someone else will care enough
to implement it themselves, or maybe it will never get 
implemented. This makes no difference to you if you don't use
DOS, but leaves the DOS people free to make things work if
they can be bothered. ie. better to spend your time improving 
the features that you use, rather than removing features that 
you don't use...



-- 
Shawn