Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - willrandship

Pages: 1 ... 21 22 [23] 24 25 ... 208
331
TI-Nspire / Re: On-calc Nspire emulators
« on: June 11, 2013, 03:18:36 pm »
128 BYTES. Not KB. Not MB. Bytes.

Those 128 bytes of RAM were used for two things: Self-modifying code, and data storage. For example, Pitfall uses 2 bytes for the score, 2 bytes for the time, 2 bytes for the lives (2 lives, but it's faster than using an AND) ONE byte for the current area, etc.

However, they had a far more manageable 4 KB of ROM, so there was plenty of room for routines.

They had no framebuffer. The screen was drawn by hardcoded routines, with specific hardware support for certain registers.

There were 5 special display registers.

2 "Player" registers which were 8 bit, and would be automatically displayed at a certain portion of the scanline. (You chose location, color and graphics before the line drew)
2 "Missile" registers which were 1 bit. (Think 1-pixel bullets) also with configurable colors per-line.
1 "Ball" register. Effectively a missile register, but with more features. Player and Ball registers could be reflected or repeated in certain locations. Missiles could not. Balls could also be stretched in two directions so as to be on more scanlines without the effort of drawing them there.

Other than those, there were 40 playfield blocks, at 1/4 the resolution of the rest of the screen. These all had the same color, unless you changed it between each scanline, and when redrawing one, you had to redraw all of them. The purple and grey blocks on that image are playfield blocks.

If you want more info on this stuff watch this video. Very informative.



So, a quick explanation of EVERY graphic on that screenshot.

The background color is being changed every few scanlines, thus the shift from lighter to darker browns.

The Purple and grey blocks are Playfield blocks, with the playfield color changing occasionally.

The player, key, and little box thing are all drawn using the player registers. Between each scanline, the graphics are changed, and between each sprite, so are the colors. You will never see more than 2 unique 8*x sprites on any line, but space invaders for the atari shows how you can get multiple duplicated sprites.

The black lines at the edge are actually a bad thing. It means their routines took too long, and they missed part of the scanline.

332
TI-Nspire / Re: On-calc Nspire emulators
« on: June 10, 2013, 07:08:41 pm »
The atari may be painful, but it is still quite weak. I wouldn't be surprised to see a decent 2600 emulator for the nspire.

The painful parts have to do with the other chips, like the graphics chip and the sound generation hardware. The CPU is a 6507 running at just over 1 MHz, with 128 Bytes of RAM, so it's not exactly powerful. IIRC it's the same CPU architecture as the NES but with some features trimmed.

If you can find the source, this would be a good emu to start with.


333
TI-Nspire / Re: On-calc Nspire emulators
« on: June 10, 2013, 12:59:11 pm »
Perfect examples of that happening on more recent games include Harvest Moon 64 (cheapest is $55, cartridge only) and the Zelda Master Quest Gamecube Disc. ($35) Where other games on their consoles go as low as $10

334
TI-Nspire / Re: On-calc Nspire emulators
« on: June 07, 2013, 09:53:35 pm »
If you want an emulator, make it yourself.

335
TI Z80 / Re: TinyCraft II (name subject to change)
« on: June 04, 2013, 05:44:01 pm »
I was under the incorrect assumption that the sectors were smaller than that. Pity.

336
TI Z80 / Re: TinyCraft II (name subject to change)
« on: June 04, 2013, 05:05:41 pm »
@streetwalker fragmentation isn't necessary for that idea, but it would make it easier.

Think of it like this:

Archive: 20000 bytes of program source.

Current offset for reading: 0

The program reads in the first 1000 bytes or so, and allows you to edit them. You insert 24 bytes.

Archive frees 24 bytes at the end, and shifts the data down until it reaches your insertion.

Your insertion is placed.

Done.

Note that this would be one of maybe 25 steps in a buffer of changes that would need to be tracked, and would be written if you ever filled said buffer or left that area of the file.

This is how good PC hex editors work.

337
TI Z80 / Re: TinyCraft II (name subject to change)
« on: June 03, 2013, 05:53:13 pm »
That doesn't sound impossible to me....

I meant a text editor that loads parts of the files into RAM at a time (Specifically, the part you're looking at) and automatically writes back modifications either once a block is unloaded or when the alteration buffer is full.

338
TI Z80 / Re: FileSyst
« on: June 03, 2013, 05:39:36 pm »
How have I not seen this project until now?

Request if it's not already there: Archive support.

By the way, don't be surprised that your code runs faster than TI-OS. There are many areas where their code could be optimized heavily.

339
TI Z80 / Re: TinyCraft II (name subject to change)
« on: June 02, 2013, 10:49:26 pm »
In other words, we need a direct-to-archive editor, that Axe supports :P

340
TI-Nspire / Re: On-calc Nspire emulators
« on: May 30, 2013, 02:54:42 pm »
to search any particular site in google just use

site:url.com stuff

and it will restrict results to that website.

341
Humour and Jokes / Re: 9001 signs you're addicted to calcs and Omni
« on: May 01, 2013, 04:40:40 pm »
5061: You've actually eaten Omni peanuts, and realized they are made from styrofoam.

342
Humour and Jokes / Re: If there was tech support in botswana...
« on: April 30, 2013, 07:03:41 pm »
No hacking necessary, really. I don't think it's that hard to screw with the nameserver.

343
TI Z80 / Re: GLIB a graphics axe 3d librairy
« on: April 30, 2013, 07:01:17 pm »
If we implement the in-place sort, the ID list becomes unnecessary, as it's implied that it goes {0 1 2 3 .. highest ID}

344
TI Z80 / Re: GLIB a graphics axe 3d librairy
« on: April 30, 2013, 09:45:54 am »
That's the idea. See, right now, you are creating a list. That's one n. Then, you are sorting that list using an n^2 method. After that, you effectively sort the resulting ID list as you parse through it, presumably with a "Go until you find the right one" approach, which , when sent through the list n times for n vertexes, is sum(1,n).

Altogether this comes to n+n^2+sum(1,n). For a 6-member list this is 6+36+21=53. n^2 is 36.

345
TI Z80 / Re: GLIB a graphics axe 3d librairy
« on: April 30, 2013, 09:26:42 am »
You don't need to draw during the sort. Just sort the list in-place as opposed to creating a new, still unsorted list.

Are there any assumptions we can make about the output list?

Pages: 1 ... 21 22 [23] 24 25 ... 208