Re: [Prboom-users] Fw: Sniper's rifle for Boom?

[Colin P. <cp...@cp...>]

On Tue, Feb 19, 2002 at 10:50:07PM -0500, mwm...@co... wrote:
> On Tue, Feb 19, 2002 at 06:27:31PM +0100, Florian Schulze wrote:
> > From: "Feher Tamas" <et...@fr...>
> > > Dear Mr. Schulze,
> > > I wonder if Boom can be made more "useful" for standalone gameplay
> > > as well as LAN deathmatches. My idea is: it would be nice to
> > > implement a sniper rifle, as an alternative to BFG9000.
> 
> Dear anyone concerned,
> 
> I wonder if I'm the only one who thinks brainstorming stuff like this is
> way beyond prboom's scope?

I'll second that. The source is out there, make your own mod and DIY if
you want the feature badly. PrBoom is not and (hopefully) never will be
a mod like that.

> 	I mean, c'mon, there are still bugs to fix; other than the
> perennial (and probably futile :) ) striving for bug-compatibility with
> Doom *AND* with Boom *AND* with Marine's Best Friend

Not futile I hope :-)

> , screwups with thing flags when reading DEH patches comes to my mind
> (try playing Strain on any map with Holobots or BFG-shooting
> troopers-- look out for attacking "Unknown gettable thing" errors).

Ah, what you want is full compatibility with dehacked, (supporting
dehacked is not an original Doom feature IMO, it was a hack and
technically illegal at that - I've never been a fan of it).
Unfortunately my dehacked knowledge is severely limited hence the lack
of progress in this area, but it is on the agenda.

> And then there's the automap's grid not rotating along with the map
> itself (press TAB then G then R then turn around a few times to see
> what I mean).

Yes I guess that's a bug (altho i can see uses for this, e.g. you want
the grid lines lined up with the player, so you can see the line
straight ahead of you).

> and maybe a 'debug' code to make the automap ignore ML_SECRET.

Hmm, that sounds like a logical extension of IDDT & friends.

On Wed, Feb 20, 2002 at 02:49:50PM -0500, mwm...@co... wrote:
> On Wed, Feb 20, 2002 at 10:59:17AM +0100, Florian Schulze wrote:
> > Could you please explain what you mean with more efficient audio and
> > visual routines? We know PrBoom needs a faster computer than the
> > original Doom or Boom, but it seems to efficient enough
> 
> Yes, but it's also slower than l[sx]doom, especially at resolutions of
> 640x480 or higher.  Personally I suspect this is SDL's fault, and not
> much to be done about it unless Colin Phipps's (unused?) X11-specific
> code is to be merged back in.

I was planning to investigate this, in fact I have the guts of a 486
scattered over the desk to one side of me now where I was going to
resurrect the machine I developed lxdoom on. Unfortunately it seems to
have decided to punish me for neglecting it by developing weird
incurable faults.

When we switched to SDL we did continue using lxdoom's old video
conversion routines for a while on the grounds that they were a bit
faster, but then dropped them on the grounds that it was Someone Else's
Problem to improve SDL. In fact I think most of the performance loss can
be blamed on things like this where we opted for portability over
complexity, like dropping the optimised i386 assembler rendering
functions and some of the loop unrolling (which only really helped old
processors, but on those old processors it makes a real difference).
There are some issues with SDL's palette support which they seem to have
tried to address with more recent versions but we aren't taking
advantage of yet - probably with some work we can get much closer to
lxdoom's performance.

The interesting question is whether people think it's more worthwhile
boosting performance for old machines, or should we go down the zdoom
path of improving performance at hires (zdoom goes to a lot of effort to
optimise the actual rendering process for hires, hence why it can run up
to far higher resolutions quite smoothly).

Actually, the optimisation I'm most interested in at the moment is for
the recent trend of monster levels, where the game code itself starts to
become a bottleneck...

-- 
Colin Phipps <cp...@cp...>   http://www.cph.demon.co.uk/