The Future of Homebrews

[gdement]

Right now I'm mostly interested in homebrewing for the 7800. That system has a lot to prove and a lot of firsts lying ahead of it. The NES doesn't feel like much of a frontier as it's already had a huge number of games and plenty of industry support during its lifetime.

[sandmountainslim]

Well, one way out of this would be Teams. We already see that coming on the VCS, where the trend seems to be having more and more artists involved in projects,

1009461[/snapback]

Teams are a good idea Now that XYPE is on hiatus have you ever thought of calling a new team/label "OCELOC" ? That has a ring to it! No charge for coming up with the idea

[Bruce Tomlin]

Pretty much, yeah. I've never actually done any SMS code myself, so I don't know, but Charles MacDonald says that you can indeed access all Genesis buttons from SMS mode. You can even switch the VDP to mode 5 (Genesis mode) while still otherwise in SMS mode, so you can also write pseudo-SMS games that use some of the added graphics capabilities of the Genesis VDP.

I looked into the tech info for the SMS last night. It really is more like the ColecoVision than I thought.

 

Apparently the SMS VDP has a full TMS9918A implementation (for backward compatibility with earlier Sega systems), adding a mode 4 which is the normal SMS display mode. This is not the MSX mode 4; the SMS is based on the 9918, but in a different way from MSX. It also has esentially the same sound chip as the ColecoVision, but it was merged into the chipset in the redesign from Mark III. One thing I found interesting was that the 16K VRAM was implemented using two 8K SRAMs, arranged as 16-bits wide. This plus the higher speed of SRAM gave the extra memory bandwidth to display four times as much data. (mode 4 always used 4-bit pixels)

 

The Genesis/MD adds a mode 5, which is the standard Genesis graphics mode (I didn't look into the details on that), but removes the TMS9918A compatibility. There were very few SMS games that didn't use mode 4, and the most notable is F-16 Fighting Falcon. I always thought that the reason it didn't work on the PBC was because it was the only 32K card game. Nope, it's because it used a TMS9918A graphics mode!

 

The Genesis/MD is put into backward compatibility mode by using the "M3" pin (which apparently refers to the Sega Mark III) in the cartridge slot. This disables the 68000 and changes the cartridge connector to the Z-80 bus. Both the Z-80 and the 68000 can use both modes 4 and 5, but nobody bothered to use mode 4 on a 68000 game. There was a proof of concept demo a few years back, but that's it. And the M3 pin was apparently removed from the Genesis 3.

 

So with a bit of code patching to handle the different ports, and remove the BIOS usage, you could make a ColecoVision game run on the SMS, but not via the PBC or SMS emulation mode on a Genesis/MD. At that point you would have to do major work to a game to convert the graphics to mode 4. You could also convert a pre-M3 game to Coleco, if you could live with 1K CPU RAM.

 

FWIW, the NES PPU also was clearly inspired by the TMS9918A, but went in a different direction.

Edited February 6, 2006 by Bruce Tomlin

[mos6507]

I'm not knocking that, but my experience with hobby projects is the more people that are required, the greater chance of failure.

 

Picking a platform which even when pushed to the limit can never require insurmountably high quality or quantity of digital assets is most convenient.

 

Well, one way out of this would be Teams. We already see that coming on the VCS, where the trend seems to be having more and more artists involved in projects,

[dark willow]

I think for the foreseeable future (certianlyl the next 5 years or so) the 2600 and the NES will be the dominant homebrew systems, since there is a still a sizeable user base of people to produce *for* and since so many people grew up with these systems. THis is especially true of the NES hardware which was (and still is) cloned/copied so heavily in Asia, and I expect we will be seeing many more hombrew titled for the NES come from that part of the world in years to come.

 

I think a good point is made about homebrew beign possibly less likely move up the technological chain to systems like Xbox/PS2/PSP because of the time demands for making such complex games. However that some games for the Jaguar have been developed may be the precursor to us being wrong on that, though I'd be *very* suprised if the amount of work produced is comparable to that of the 2600.

[Bruce Tomlin]

After researching the SMS and Genesis (the summary of which I posted above), I am seriously considering going straight to the Genesis when I am through with the ColecoVision.

 

I mean, as long as I'm moving up to a system with enough color and memory space to make some nice looking games, I might as well go all the way and get a 68000 as well, for which it's quite easy to write code in C. In fact, I could even use Apple's own MPW compiler as long as I don't need to run it on an Intel CPU. And I'm quite at home with looking at disassembled 68000 code, having cracked much copy protection and hand decompiled much code back in my older Mac days.

 

Why should I force myself to dick around with the limitations of the NES video chip, 6502 assembler (which I've always found annoying), and dual-bus cartridges (with a wierd pin spacing, no less), when I can go straight for something much better? Why should I drive a Yugo if I can get an Escalade for the same price? Yeah, 20 sprites per line, baby.

 

Another part of my choice was realizing that my Multi Game Hunter copier unit (which I got used without the SNES parts) could be a decent development system.

[ZylonBane]

The higher you go on the console food chain, the less chance you, as a homebrewer, will be able to produce something that doesn't look like crap. It's as simple as that. A 2600 coder working by himself can produce a game as good as anything Atari put out because those games mostly were created by a single person. But beyond that point, games were created with ever-larger TEAMS of people. To make a decent Genesis game, you pretty much need a dedicated artist, programmer, and music/sound designer, at the bare minimum. Or if you choose to do all those things yourself, you'd better be a genius or have five years blocked out to make your game.

[+swlovinist]

I think that Atari 2600 will stay the leader of homebrews, just mainly to the sheer availability of product. Although the 2600 is a monster to program for, it still can be done with a small operation or even one person. You can still buy up extra carts and systems for very cheap(when compared to other systems). As for other systems(Nes) you have to deal with lockout chips and more complicated programming. I would agree with Colecovision systems scarcity being a factor, as well as not being as durable as the good ol 2600. One system that is a huge suprise to me is the Vectrex, which for only having under 40 games released offically, it has a huge homebrew base.

[BassGuitari]

I think it'd be cool to see more homebrews for the Odyssey 2, the few that are out now (Amok, Planet Lander, KTAA, and Pong) are awesome. Besides, even the biggest Odyssey fan has to admit that there are few really good games for it.

 

And Blackbird, I personally would love to see a homebrew for the Channel F! I like the Channel F, but it needs all the help it can get. There's fewer decent titles for it than even the Odyssey.

[ZylonBane]

Besides, even the biggest Odyssey fan has to admit that there are few really good games for it.

There's a reason for that.

[Bruce Tomlin]

The higher you go on the console food chain, the less chance you, as a homebrewer, will be able to produce something that doesn't look like crap. It's as simple as that. A 2600 coder working by himself can produce a game as good as anything Atari put out because those games mostly were created by a single person. But beyond that point, games were created with ever-larger TEAMS of people. To make a decent Genesis game, you pretty much need a dedicated artist, programmer, and music/sound designer, at the bare minimum. Or if you choose to do all those things yourself, you'd better be a genius or have five years blocked out to make your game.

I know you get a real kick out of telling other people what they can't do, but seriously, there is no law that says a Sega Genesis game has to be graphics heavy. Deep down inside, it's basically a character-cell graphics machine like the NES, just with more capability (more and bigger sprites). Just because it can do games with giant sprites doesn't mean every game has to be a giant sprite game. Plus, being able to develop in C should actually make it easier than if I had to do everything in assembler.

 

I'm not sure what your definition of "decent" is, but let's make a deal: I don't force you to play my game, and you don't force me to make yet another bleeping Mortal Kombat clone.

[supercat]

Besides, even the biggest Odyssey fan has to admit that there are few really good games for it.

There's a reason for that.

1020876[/snapback]

 

The biggest reason for that was probably that the O2 has a BIOS internal to the 8048 which cannot be easily accessed; this increases the difficulty of reverse-engineering to the point that few outsiders bothered to produce carts for it (none that I can remember being aware of at the time).

 

From a 1979 technical perspective, the 2600 has the O2 beat on playfield, color, and sound, but the O2 had the 2600 beat on sprite graphics (the O2 could display four programmable 8x8 sprites, twelve independent sprites that used shapes from a fixed library, and four groups of four 'tethered' sprites using that same shape library (unlike the 2600's "cloned" sprites, the groups of four sprites on the 2600 had independent shape and color registers). The O2 also had more RAM than the 2600.

 

Prior to the emergence of Activision, the top three consoles would probably have ranked "(1) Intellivision; (2) Odyssey2; (3) Atari 2600". It was only after Activision came on scene with things like the VDEL sprite trick that the ranking changed to "(1) Atari 2600; (2) Intellivision; (3) Odyssey2".

 

As it happens, I don't think that even a super-genius programmer would be able to push the O2 fully to the level of the best 2600 games. On the other hand, the O2 has games like UFO!, Killer Bees!. or even K.C. Munchkin! which the 2600 would be hard-pressed to match.

[mos6507]

If you release a game with more minimalistic graphics than most of its catalog has, people will just not accept it that well. I think homebrewers require some minimum adulation to feel like their efforts were worth it.

 

I think that's why the Jag homebrew scene is lackluster at best. The Jag isn't very good at 3D, but it is a 2D powerhouse (look at Rayman). No offense, but regardless of how good the Gorf Classic homebrew may play, it is not exactly a technical showcase.

[jaybird3rd]

If you release a game with more minimalistic graphics than most of its catalog has, people will just not accept it that well.  I think homebrewers require some minimum adulation to feel like their efforts were worth it.

1021193[/snapback]

I agree, and that's one reason why, when I sat down to decide what my homebrew platform of choice would be, anything from the 16-bit era onward was pretty much off the table right from the beginning. Those are the systems that are powerful enough to really show off the value of professionally-created art, and any efforts of mine would fail to hold up alongside commercial releases (one requirement of a successful homebrew title) without a professional artist, which I am not.

 

In my opinion, the most interesting homebrew platform at the moment is the Atari 7800. It has the smallest library of any of the major consoles of its era, and since many of its titles are mediocre at best and cover only a few genres, it has the most untapped potential. Its MARIA graphics chip is powerful enough to create appealing visuals, but isn't so powerful that a professional artist is needed to create graphics that match the best commercial 7800 titles and push the hardware to its limits.

 

As for homebrew development in general, I think it's likely to continue as an exciting but very small subset of independent game development. It's more of a labor of love than development for profit-generating platforms, and is a haven for those who want to have fun creating games but do not have the resources to produce the kind of audiovisuals that games on modern platforms need to get any attention in today's world. Great independent games are released for modern systems every day that barely get any notice, but those same games would be a much bigger deal on something like the 7800. Even if one-tenth as many people see them, those people are a larger percentage of the user base and in general are more enthusiastic and supportive (although the foaming-at-the-mouth lunatics in the Jaguar community are a notable exception).

Edited February 21, 2006 by jaybird3rd

[ZylonBane]

Its MARIA graphics chip is powerful enough to create appealing visuals, but isn't so powerful that a professional artist is needed to create graphics that match the best commercial 7800 titles and push the hardware to its limits.

Of course, this is helped by the fact that it's pretty apparent Atari rarely hired professional artists themselves. Granted the late-era stuff like Ninja Golf had well-done art, but most of the catalog has that amateurish MicroIllustrator look.

[jaybird3rd]

Of course, this is helped by the fact that it's pretty apparent Atari rarely hired professional artists themselves. Granted the late-era stuff like Ninja Golf had well-done art, but most of the catalog has that amateurish MicroIllustrator look.

1021910[/snapback]

Too true. That's why the 7800 is a much less intimidating platform for homebrew games. The standard that has been set by many of the commercial releases is pretty low, and would be a lot easier for a homebrewer to surpass than the standard for, say, the Sega Genesis games (which often had teams of artists contributing to them).

[Bruce Tomlin]

If you release a game with more minimalistic graphics than most of its catalog has, people will just not accept it that well.  I think homebrewers require some minimum adulation to feel like their efforts were worth it.

 

I think that's why the Jag homebrew scene is lackluster at best.  The Jag isn't very good at 3D, but it is a 2D powerhouse (look at Rayman).  No offense, but regardless of how good the Gorf Classic homebrew may play, it is not exactly a technical showcase.

I don't think Gorf for the Jaguar is an example of "it's too hard to make a game look good enough for a more powerful system", but of "something that could be played on MAME being ported to a system that few people have".

 

Seriously. Gorf is a pre-crash arcade game, and as such is trivial to run the original ROM with MAME. The Jaguar is a niche system, and dare I say it, it's a "fanboy" system too. Okay, so Minter's games were good, but the system didn't have many other compelling games, and mostly only Jag fanboys are likely to have one hooked up and ready to play. And since they're the only target market (whether the game author realized it or not), they are the ones who will look at it and go "WTF does this need to be on the Jag?"

 

On the other hand, you have me, already trying to cram a 1986-era game (and an original game, not a port) into a ColecoVision. I would do more colors if I could, but the CV doesn't have enough graphics power for my own art skills. I'm no serious artist, but I originally came up with some sprite ideas that would have worked better with 16 colors. It took some decent art "skillz" to work around the limitations.

 

I've posted elsewhere today what my reasons were for choosing the Genesis, but the simple fact is that I wanted a system that could do 16 colors from a palette, and that was reasonably easy to build cartridges for. The NES cartridge bus is too complicated (two data buses AND a lockout chip, combined with a weird pin spacing), and the SMS isn't common enough. Those were the only two 8-bit systems from that era worth looking at. The Genesis came next, and has a lot of what makes the Coleco easy for me to program. Plus, the linear addressing means that no bankswitching will be needed to get past the 32K limit, which I will probably hit on the Coleco.

Edited February 25, 2006 by Bruce Tomlin