Ok, I cou= ldn’t take it anymore and decided to completely break apart the program into sepa= rate component parts and start from scratch. I have learned much in the past 3 years, and couldn't stand looking at the bad programming any more.

&nbs= p;

So, here = is a list of changes for beta 2.

&nbs= p;

0. The ra= ster games are gone. They may make a comeback, but for now AAE is vector only. T= his lets me focus on making the best vector emulation possible.

&nbs= p;

1. AAE is= now a basic driver driven application. All rom, sample, artwork, key configs and dip switches are loaded centrally. This m= akes for less work and more uniform settings.

&nbs= p;

2. There = is now one central loop for the application, with an init, run loop and shutdown routine for each game. much nicer.

&nbs= p;

3. The bu= ilt-in GUI is now game 0, and is treated just like any other emulated game. Now I = can update it to reflect the new direction of aae. (When I get time)=

&nbs= p;

4. Overha= uled the menu system, adding floating point support. I am actually quite proud of it, with only about a page and a half of code. No spaghetti code!

&nbs= p;

5. Rewrot= e the vector drawing routines, adding gain support and reordering the drawing so dimmer lines are always drawn first, with brighter on top. Solids are now d= rawn first as well, so that lines composed of more then 1 color will always be on top, just like the real thing. With the gain support you can now crank up the brightness to see 0 brightness lines, just like on= the real hardware. (Like in MAME tm.) Also with the gain control you can = now over bright the vector lines, to reflect more real world looks. As a side note t= he draw routines are about 2/3 faster, due to removing most of the gl calls from the rendering loop.

&nbs= p;

6. Added = phosper "Glow" effects to more simulate the= real thing, and made it variable. This goes through quite a lot of GL gyrations,= so don't expect it to run on low end hardware. It takes about 1.5 to 2 ms to composite and render on a 7600GS, not too bad.

&nbs= p;

7. Added = shader support for the above effect and a multitexturing shader for overlays. Shader 2.0 is the only thing I am supporting, so this will immediately stop older cards from using this progr= am. Anything decent made in the last 3 to 4 years should run fine, so I don't s= ee any need to change this. Time marches on, and a decent NVidia card is now about $80 bucks in the US. from newe= gg.com

&nbs= p;

8. Lots o= f bug and memory management fixes, reduced the memory footprint by about 25%.

&nbs= p;

9. Remove= d the non-working 6809 core. I still need to find one for St= arWars support, and I don't want to use MAME's. I am not ready to GPL my code yet.= ...Arrggh.

&nbs= p;

10. With = the rewrite of the GL code, everything is now rendered to a second FBO buffer, = and then blit to the screen. This takes more time, = but is faster at higher resolutions, less pixels to be rendered to the framebuffer. Makes compositing th= e image better to, since that is all now centralized in the GL core. All rendering is done internally at 1024x1024, no matter what resolution you= 217;re running.

&nbs= p;

11. Added= rom audit funtions, -listroms<= /span> lists the roms required, -= verifyroms (obvious) and -listromstotxt, which lists all required rom sets to a text f= ile. Tried console window support, hated it, so this is the best I can do for ri= ght now.

&nbs= p;

12. Added Quantum, although until I can figure out rawmouse support, the mouse code sucks, horribly.

&nbs= p;

Still working on sorting out all the GL changes and making e= verthing work, at least for the most part. All drivers have to be rewritten, so now I am running a total of 5 games for testing ;(

&nbs= p;

Lots of b= ugs still need to be squashed as well.

Expect a = release in 2 weeks +!!!

&nbs= p;