Slashdot Mirror
Andre Lamothe Launches XGameStation
TheAdventurer writes "Andre Lamothe, author of many popular video game programming books, has released his XGameStation. The initial offering, the XGameStation Micro Edition, is a retro level hardware platform, similar to the old Atari and NES systems, designed to teach enthusiasts and students the elements of console hardware design and effective low level programming skills. The unit comes with an e-book written by Andre on how to develop on the platform using its assembly language IDE (included) and how to make your own extensions to the device. It is priced at $199."
Cheers,
Erick
http://www.busyweather.com/
This is a good way to get a new generation of engineers going. I know many of us probably got our start with games, computers, etc at an early age. It seems like it was easier back then because the systems weren't as complicated as they are now. Back then we had nice breadboards, through-hole components, eeproms, etc that made hardware hacking and learning easier. Now with all the SMT, it's hard to jump in and learn from scratch.
A new device such as this with eduction in mind from the start surely sounds like a nice hobby experience.
Buy your tech gear for much less.
If you cannot do that sort of stuff on a modern computer (which you can, to an extent), what's the point in learning about it? Is it just for people who want to learn ASM, with the prospect of being able to use it in embedded devices or something? These days alot of the graphical things are handled by the GPU anyway, meaning much of what you learn is rather useless in terms of modern game programming, plus any of your creations have a rather limited distribution market. Though it is rather cool if you DO want to make things in ASM larger than your normal bits of inlined stuff, I guess.
A bit of background on Andre....
I'm posting this AC because I know this will get modded down as flamebait.
I've been working in the game industry as a programmer for 15 years, and have never regretting purchasing a single technical book (and I've bought hundreds) until, one day when I was at Barnes & Noble, buying six by Andre Lamothe thinking they might be useful.
Not only are his books terribly tied to specific platforms (he wouldn't know how to program if sample code wasn't handed to him on a platter by Microsoft--most of his sample code is taken STRAIGHT out of Microsoft's examples), but he can't write a paragraph without factual errors to save his life.
I remember a whole section where he was talking about 56 Kbps modems having a total of 56 Kbps of bandwidth split between incoming and outgoing data, and if you sent more, you could receive less. And another where he dismissed NAT because IP addresses on the internal network would conflict with those on the Internet and the idea would never catch on (he'd never heard of reserved netblocks, apparently).
He has become just a name and a marketing tool; do NOT rely on him for actual game programming tips and information.
and more powerful as well.
Around $25 for the system on ebay, then a quick search for a bootdisk.
Plus, it can be made bootable, so other people can play your game without any modifications to the DC.
~~~
Click here, you know you wanna!
You haven't done much PS2 programming, have you? I have, and yes they do.
word.
Game code still uses alot of assembly, unless you want it to go at 10 FPS. Letting API's handle stuff isn't always freat, theres overhead and at leasy 99% of the time bulk. (You rarely use 100% of the features of anything) Reinventing the wheel is bad, but I have code I've made myself that I reuse in games, and just play around with it a bit so it so it perfectly fits the games needs. For other stuff though, I agree with you, unless your programming for some special case (Embedded stuff, robots, games etc) then there is no need for ASM, and API's are great.
The XGS was changed beyond recognition. Before, it had a Super NES style GPU. Now it has one more similar to that of the Atari 2600, where the CPU has to output each pixel one by one.
The design of the xgamestation is almost *exactly* like the arcade hardware that we're using at work.
This little bit of hardware might seem to be underpowered, but I can tell you from experience that this is the kind of stuff that a lot of professional game developers are using today.
With the arcade business being what it is today, the challenge is to fit the most amount of game into the smallest (meaning cheapest) hardware. The game I'm working on right now runs on hardware (similar to the xgamestation) that only costs about $40 for each unit. It's basically a 66mhz Z80 cpu, 4mb of texture memory, and an fpga that is really good at moving around memory, and it's a childrens redemption piece (ticket spitter) that's going to sell like hotcakes.
When I applied for my current job, I was one of the only guys who had any non-pc game development experience (through the sony net yaroze). This little box should open the door for a lot of aspiring game programmers.
Between Torque and the XGameStation, young programmers have no excuse for not having a kick-ass game for their demo reels.
for great justice, this sig has been moved
He may have built this more for people who want to learn for the sake of learning, not for any particular glory or power. True, APIs are great, but you cannot be a truely great programmer unless you know what the API is doing, and could at least theoretically craft it your self. Doing it all your self once for the sake of learning it will make you a better programmer when you retrun to useing the APIs because you will know how the internal algorithm is working, and can make your code intergrate better.
The article says "built in programmer".
I knew the Japanese people were small, but this is ridiculous.
You can learn to program on DOZENS of the best "classic" arcade platforms (as well as more modern stuff like Dreamcast) and it won't cost a dime. Download MAME and an assembly language manual for the machine of choice and have at it.
Spend $199 and hours of time to learn obsolete skills? Games aren't programmed in assembly anymore... things have changed in the last 20 years.
It would appear you have completly missed the point here. First they do plan on making compilers for C and basic.
second on of the main ideas here is to learn how to be a GOOD programmer, the limitations of the system will force you to code well. this wont be like a desktop where you can have super sloppy code, but Ghz of power to force it to run.
Third, one of the things this is designed to teach is the hardware side of things. Its designed to be a completly open system to let you hack, tinker, and create. It comes with a book that teaches you the entire EE side of designing such a system. The idea being that if you dont like this system (or just think it might be fun) you can build your own system, and make it do whatever you want.
Yeah, that's where I'm heading too. Even the $80 Linksys WRT54GS has a faster processor, right? I don't know how economical these Single-Board-Computers are, but you can get pretty fast CPUs for a reasonable amount of money, and low-power enough to still not require a fan. Are these sort of things in the realm of hobbyists, or does it become to expensive/difficult to engineer single quantities of these?
200$ is a bit expensive for a hobby video game programming system. Why not just pick up a dreamcast for cheap and use it?
There are plenty of resources for it and some good examples of homebrewed games and applications already out there.
Not to mention it's a lot more modern and you can use some nice rendering hardware.
Linkage:
Dreamcast Programming
Dreamcast Homebrew
Dreamcast Emulation
At my job, most of the code I write gets changed around dozens of times before the product is out the door. That's just the nature of the business.
First they want a space ninja, then they want a caveman lawyer. The game you start programming the first month of the project is almost guaranteed to be an entirely different game from the one that you finish.
Frustrated with all of the changes, I asked my boss once for the design document, and he pulled out a barely legible scrap of paper.
That goes for the low-level hardware stuff too. Even the hardware might change mid-project! My sprite code stopped working one day after we got new boards from the hardware people. I called them up, and sure enough, they changed around the hardware! (Thanks for telling me beforehand!)
Long story short; assembly programming is a skill that every serious game programmer should know, but you never know where you're code is going to end up, so it only makes sense to write well-documented C/C++ instead of the complicated ASM equilavent. (yes, I know you want to prove how clever you are with your 100% ASM programs, but in a professional environment you need to make sure that your code is understandable and modifiable by everyone on the project.)
Even on the low-level system I described in my previous post, we still use C.
for great justice, this sig has been moved
The only thing that really gets me on the thing is this price. Would it be cheaper to homebrew your own console? I mean you can still get 65C02s and memory in such small amounts shouldn't be that expensive, right?
Anyway this thing is off to the right start I think. It even has a demoscene, which to me shows how interested people are in this thing. I might get one of these in the near future and mess around with it.
Read the Quake II source (sans obsolete software renderer) and tell me "game code still uses alot of assembly." Even the math library is C. And that was seven years ago.
"[A] high IQ is like a Jeep; you will still get stuck, just farther from help!" --Just d' FAQs, c.g.a
"And phone lines have send/recieve pairs"
. ht m
There are "complete and total morons," and then there is you. Phone lines do not have "send and receive pairs". There are just two wires (one pair) in most telephone systems in the world. Some European systems may carrry a high voltage ring signal on an additional wire, bur that's not necessary.
May I suggest you take a quick google before you "spew" yourself:
http://electronics.howstuffworks.com/telephone2
-S
"on of the main ideas here is to learn how to be a GOOD programmer, the limitations of the system will force you to code well."
To get the best performance out of limited hardware you are going to have to be a very BAD programmer. Out of code ram for that special effect loop? just change a couple of bytes in a similar one and call it. Need to make sure to update the sound DAC every 100 cycles? put in a call to the DAC check between every other call and in long loops.
The system is great for learning what hardware and APIs do for you, but to be a good programmer you need to work with a complicated system. Different tasks require different skills and game programming have many different tasks. If you program for any of the current consoles you are going to have to learn about writing efficient code, because you will always need to do more with less cycles.
i've always been interested in homebrew game development.... but $199? for that price, i could buy a gamecube, max-drive-pro, nintendo ethernet card, and datel keyboard. the ideal gc-linux setup, ready for all the low (or high) level programming you want. plus, the hardware's a lot more capable.
Because he is shipping ebooks with it that "assume" the reader knows nothing about harware design, and if you go through the book from start to finish, you'll be able to make the unit yourself. Of course, you'll learn how to make that system, and maybe it'll teach you how to make your own system, as well.
Thanks,
Leabre
Look at titles for portables, the smaller handhelds (e.g. a gameboy), a great many of the uprights in the arcade...
The fact is that a lot of code is still being written for devices/systems where your app will be seriously starved for resources. This is where "game code still uses alot[sic] of assembly". And you can bet that even though the bulk of the codebase may be in C that someone has taken the time to disassemble it and make sure the compiler is doing the right thing -- and then re-writing portions in assembly or reworking the code to make the compiler perform better on it.
I want a new world. I think this one is broken.
Seems at first to be an Atari 2600, little more or less. I was going to ask why, when emulators for the 2600 are exceedingly close to the real thing due to the thousands of games by unconnected programmers that they must successfully run, anyone would take this approach. But on closer inspection there's 128K of "fast 15ns static RAM" in there, and he mentions that it's fast enough to use as a conventional graphics framebuffer. So you don't necessarily have to take the pure Atari 2600 approach.
Also, the CPU is much faster -- it's probably inaccurate to say 80 times faster because I don't know whether the instruction set is more or less primitive than that of the 6507, but I doubt it could be much more primitive.
That's probably the reason why there is no sprite hardware (*): you can program your raster-on-the-fly graphics fast enough without it. (Yes, even the 2600 had one-dimensional sprites which you could reposition or alter between scanlines.)
(*) If something they would have burned into an FPGA should really be called hardware. Firmware, I guess. For programmer purposes it certainly is part of the hardware.
So this is sort of a super Atari 2600 on steroids, if you ignore the easy way out of simply using the static RAM as a framebuffer or using somebody else's prewritten subroutine for doing so. Apply the same freaky scanline by scanline tricks the 2600 programmers did, but accomplish more with them! In that sense, it's a really neat machine, and I wish him luck with this.
Still, I think writing new games for the 2600 appeals to me more. Not that I'll never do it. Hell no. When it comes to assembly, I'm strictly a voyeur and a dilettante.
"Oh look, Geoffrey. The 6507 programmers are sweating as they seek to save a single clock cyle with a cleverly placed jump instruction. Oh my yes, that is quite admirable. I do work up a thirst watching another man work. Another martini? Why thank you..."
Check out the Apostrophe open-source CMS: http://www.apostrophenow.com/
Too bad parent is erroneously marked offtopic... I think it's spot on.
I've actually met André LaMothe. I was 14 or 15 at the time and living in Milpitas. I sent an email to the CompuServe account listed in the back of his book and he not only was kind enough to reply promptly, but explained several of the concepts I was having trouble with to me over the phone in more detail. He gave me a copy of what was at the time his newer book, as well as an older version of Watcom C/C++ that he wasn't using any more.
I was used to always writing to an API (I wrote, as an exercise, my own mini-OS in assembly, but it still used the BIOS for everything) and coding to bare metal was a new thing to me. I knew of the dangers of C macros (multiple evaluation, etc) and so avoided them entirely, but he gave examples of where they could be used to great effect and how to avoid those dangers. I learned about how to implement preemptive multitasking from his books, about effective use of lookup tables, about how knowledge of higher math can be leveraged to write much more efficient code.
His newer books ended up focusing on DirectX, and I lost my interest in game programming when I realized that without a full studio and millions of $ it would be near-impossible to compete with the Quakes and other established models that continually raised the bar. I now program in a business context (amusingly enough, using Lisp macros heavily, which put anything C macros can do to shame) but I'm grateful for everything I learned from his books.
I think this XGameStation is a great idea. I may buy one. Something which can show inexperienced programmers through experience the difference between an effecient algorithm and an inefficient one is an idea I fully support, and this looks like it could end up being lots of fun besides.
TO BUY A NEW CAR WOULD MAKE YOU SEXUALLY ATTRACTIVE.
Graphing Calculators can provide a similar (and much cheaper) tool. Beginners can start off with BASIC programs, and although these require some tweaking to run smoothly, that's part of the fun. (I goofed around with one myself a little while ago). More advanced assembly programmers have achieved some amazing stuff (such as greyscale on a black-and-white screen and a very accurate port of Bubble Bobble. The specs are all there for the asking.
Sure, calculators aren't as cool or powerful as what Lamothe is putting out there, but they do force you to be creative and do a lot with a little. Besides, making something happen on a little box just seems cooler - you tend to take having complex software for granted with PC's.
I produce electronic music and write little games. Have a look.
I work at a game company and the title we're working on is going to be out for PC, xbox, PS2, and GC. The lead guy is doing the PS2 port and we have one dedicated guy doing the GC port. It's all on a single codebase with the makefiles deciding what platform to target. We use C++ with GCC and VectorC for the consoles and Intel's compiler for the PC.
Very little is actually coded in assembly. The main things coded in asm relate to graphics: VU's on PS2, the shitty microcode for GC, and the shaders for xbox. I'd be hard pressed to find any asm on the PC port, mainly because many of the intrinsics are already included in the compiler already.
Honestly... you can pick one up for just over a hundred bucks (even cheaper used), plus a flash cart for about the same or less, and you can develop software for a decent, popular console. It's definitely a more old-school style platform (has a nice, traditional sprite-based VDP) and in order to do anything really fancy, you pretty much have to write some code in ARM assembly (which is, BTW, a joy... the ARM has a *really* nice ISA).
As for documentation, the GBA is only rivaled by the Dreamcast in terms of how well documented it is and how large the homebrew development community is. And much like the DC, you can use the standard GNU toolchain for development. And, of course, you can still run regular ol' games on it, so it doesn't *have* to be just a development platform.
So, why would anyone bother with this thing?