Developing 13 games in a year, porting Wolfenstein in 3 weeks to the SNES, those time frames are insane. Nowadays a meeting can take longer than 3 weeks...
TBF, games back then were exponentially more simple than nowadays. At surface level, a modern 2D pixelart platformer might look like it could run on an NES, but in reality, there's bunch more stuff happening under the hood, a lot more complex scripting and things to keep track of. Bad optimization? Sure, who cares, nobody's playing Shovel Knight on an NES anyways... Or are they?!?! It still is very impressive.
@@jesustyronechrist2330 well, yes and no. When you're doing everything the first time that's a whole different complication. Design-wise and technology-wise.
I would say early Id may be the most inspiring game company due to being only a few charismatic people plus their amount of dedication, productivity & innovation. Definitely check out the recent (5-hour long!) Lex Fridman podcast with John Carmack as well.
I love masters of doom! and Hackers, heroes of the computer revolution. Now Mr.Romero has released his autobiography "Doomguy" and its as good as masters of doom. Its great to read/listen to his own version.
No matter how many time I hear the story I always show up to hear it again. I even have the book! (There's a new one coming out next year apparently!). Rock on John, rock on.
Well that was exceptionally fantastic. Thanks for hosting and publishing that! Thanks to John Romero, too. Even though I'm a gamedev and often watch stuff like this to glean gamedev wisdom, I think my favorite part was hearing about John Carmack's D&D game.
What a joy listening to John. Reflecting on my own life in software through the same eras and of course the awesome nostalgia of playing Doom and Quake. Carrying our whole system (with CRT) around to friends houses to deathmatch.
I think the most influential bug that became a feature was strafe jumping/bunny hopping in Quake, the fact that it is quite essential in some games to this day says a lot.
Nobody ever really seems to talk about Heretic or Hexen anymore, but even though I loved Doom, I preferred Heretic because I was more of a fantasy nerd than action in those days. I readd Lord of the Rings in the 70s and right after that D&D became huge. Then Quake 2 came out and I was a bit older, and into rock music and the music in Quake 2 blew my mind. I used to play in the living room with all the lights off and had my PC plugged into my sound system and it was insane. John & Co, you made this Gen Xer a happy man to this day. Thanks.
17:00 Develop on a system that's superior to your target. I can relate to that. In the early 90's I took over the group developing the software for an industrial machine that used a '286 motherboard. The development machines were PCs built around this same board, because my predecessor wanted the development system to match the target system. It had a detrimental effect on the code, big time. The developers were loathe to change a header, because of the long compilation times. So instead of fixing or updating a core data structure, they would come up with crazy work-arounds that could be done in one CPP file. I came in with my personal '486SX budget machine, and quickly slashed through some of these outstanding issues, rebuilding clean in the time they took to incrementally make. As for needing the same hardware, that applied to the installed cards not the base PC, and targeting the 286 for generated code, and having a "slow clock" or "de-Turbo" mode for testing.
These talks are awesome. Amazing people doing amazing things. Such a great story and always interesting to hear these guys share. Now just to form a small team of multiple. exceptional, passionate, respectful geniuses and give them full ownership of their work. It’s easy!
IIRC, Death Rally, and Tyrian, were both successful 90's shareware/DOS games that started from demoscene groups. There's probably a ton more. There were a lot of Finnish DOS indie games with stellar soundtracks in the 90's.
Brillliant talk! I felt a slight contradiction in that slide about "writing your code for this game" and how all code was written afresh and then later admitting that engine code and elements were reused because they weren't going to get any better. I can imagine that from one perspective the game is new but the engine(s) are always improving.
I haven't fully watched it yet, this is mainly a note to myself to check if John tells what the bug that they had in the game since day one was and how they fixed it. Edit: 30:56 is the timestamp where this is answered. Thanks, me. You're welcome!
I was doing web design in Madison in 1991. Only 5 servers for the web existed at the time and Madison had one. I found one of Wolfenstein shareware on our server.
I always wondered how or why IBM put the scroll registers into the CGA, EGA, and VGA designs. I can remember when people sort of discovered it, and started using it. It was in "trainers" and demos. But why was it even there??? What was the intent? I don't think it was for game programming. And even then, it was just barely even usable. (I sure had a rough time, being a senior in high school, just starting to learn Pascal, ASM, and C/C++.)
So fun side-notes about all of that. The old video adapters had these scroll registers as you mentioned, and if you could configure the memory in such a way, you could draw to one part of video memory while displaying another. Once you were done drawing you could flip the "viewport" register to the other area and start drawing where you were just displaying. This double buffering allowed for flicker-free graphics. The problem was that no standard video mode was setup in a way that allowed for two screens of 8-bit color to fit in video memory at once. You heard Romero mention Michael Abrash. Abrash's discovery was that with a bit of mucking with VGA registers you could kick the cards into a undocumented "unchained mode" allowing for 8-bits of color on a 320x240 display, meaning we could finally get much richer color with double buffering. In addition, there was just enough video memory left around that these two buffers could have a border around them of extra tiles allowing for smooth panning around the screen.
CGA was built on top of HD6845 (extremely common CRT controller chip in 80s), which has scroll registers for obviously scrolling the character display fast, and not having to copy the whole buffer.
No TDD, no *DD, no Agile, no Waterfall, no latest tech buzz IdSoftware delivered 28 games in 5.5 years without version control system and just syncing between them. IdSoftware is mainly responsible for creating the gaming industry That is WOW, today with all the buzz, "latest" software process and it still hard to deliver good software. This talk will save you tone of money on "make a great software" consultants In summary: Sync with your co-workers, especially if you know that your code change can interrupt someone else work
My hypothesis: Most of the "processes" are created by observing the successful companies. But it is hard to observer core behaviors and core principles so what is left are some rituals.
"No prototypes. Just make the game. Polish as you go. Don't depend on polish happening later. Always maintain constantly shippable code." This was not the current age of the game industry.
there are more proccesing calculations happening. pre processing 1. real time data proccesing 2. and post proccesing 3. good luck. curious ones. :p then, single tasking 1.switching-multitasking aka turn based 2. real time aka, simultanious tasking operations. "all at once" 3.
I do think it's funny that iD Software hired someone to do the SNES port of Wolf 3-D & got stiffed on it, then almost the exact same scenario played out when they contracted out Doom 1's networking code! Paid some guy who ran off with the money, they never got the networking code & ended up having to do it themselves.
Damn, I miss the early days of where gaming was born from. I was too young to really experience it. I was a part of the console wars that happened with Atari, Sega, Nintendo, NeoGeo, Turbografix, etc.
Commander Keen 4 was so much fun. I wish someone had asked him if he had a favorite game of all the ones he worked on. Sounded like it might have been Doom.
I have a question about the SNES port of Wolfenstein-3D. You said you had to learn the hardware, convert assets, etc. But was there a practical C compiler that targeted the 65816? You didn't mention having to learn assembly language and work out performance tricks on same.
I got distracted for a moment then came back in the middle of his story about casting spells and talking to demons, and I'd missed the opening where he'd explained he was playing Dungeons and Dragons with his friends. I must say I was extremely confused for a minute there ... haha.
Interesting talk, but Wikipedia tells a completely different, not as rosy story on the Quake development… also, how do you go from multiple games per year to Daikatana? Seems the most interesting part was left untold…
Sorry John but Keen didn't pioneer game engines. Just off the top of my head I can think of Gary Kitchen's Gamemaker (1985), SEUCK (1987), Freescape (as used in Driller, 1987 and several later games like Dark side, Castlemaster etc) and there were several other game engines before Keen arrived.
I played and owned EVERY game made by John Romero and John Carmack. Not one of their games had any bugs. Not one. This is a practice of work ethic that has often fallen to the way side in favor of greed, early access garbage, and lazy people! Every game designer needs to emulate their impeccable work ethic. It is a lie that you can not make games without bugs. A game not mentioned was called Blake Stone, which I thoughourly loved from Apogee. One last thing I will mention is that I have reservations concerning the occult. I would never outwardly engage in any type of magic, witchcraft, or satanism, and that includes D&D.
so, the buffer, a dedicated piece of ram. is limited. and filled with data, from multiple sources say software code and keyboard input and hard drive data. and empty with a clockcycle say X times a second. and then refilled. the data goes to the buffer. say, one by one, and the buffer (or cache) dispose that all together in go, too different things, like music and display.
and then the buffer memory is filled with a higher frequentie or fps then it is emptied. say i.e. 30 hertz per proccesed data stream. per 5 megabyte. per s3cond. and then emptied 1 once a second. buffer under run. ( no use the full buffer memory array. and purge it before it reach its limits. ) buffer limit is by the programmer Pre-designed. or Pre-fixed. in size.that along side the rest of the ram memory usage. thus, " in the meanwhile" or " in the lean time" and alsobhas its own clock freqeuncy. good explaination sir, romero.
Developing 13 games in a year, porting Wolfenstein in 3 weeks to the SNES, those time frames are insane. Nowadays a meeting can take longer than 3 weeks...
TBF, games back then were exponentially more simple than nowadays. At surface level, a modern 2D pixelart platformer might look like it could run on an NES, but in reality, there's bunch more stuff happening under the hood, a lot more complex scripting and things to keep track of. Bad optimization? Sure, who cares, nobody's playing Shovel Knight on an NES anyways... Or are they?!?!
It still is very impressive.
@@jesustyronechrist2330I heard Mario 1 took 10 months
@@jesustyronechrist2330 well, yes and no. When you're doing everything the first time that's a whole different complication. Design-wise and technology-wise.
I am completely addicted to the story of ID. I don’t really know why. It’s just so inspiring. I listen to Masters Of Doom at least once a year.
I would say early Id may be the most inspiring game company due to being only a few charismatic people plus their amount of dedication, productivity & innovation.
Definitely check out the recent (5-hour long!) Lex Fridman podcast with John Carmack as well.
I love masters of doom! and Hackers, heroes of the computer revolution. Now Mr.Romero has released his autobiography "Doomguy" and its as good as masters of doom. Its great to read/listen to his own version.
same...and I don't even like doom all that much (I do like quake though) just the whole 90's dev era radiates crazy inspiring energy
That was the fastest 1.25 hours I've ever spent listening to a lecture. I was engrossed the entire time.
No matter how many time I hear the story I always show up to hear it again. I even have the book! (There's a new one coming out next year apparently!). Rock on John, rock on.
Masters of Doom? That's my favorite book. I listen to it on Audible while I work now-a-days. It inspires me to keep going.
A new book coming out? What's the name?
@@paulk314 DOOM Guy: Life in First Person
Listening to John Romero talk about the history of id software is like listening to the grandpa I never had tell me bedtime war stories!
Just a huge thank you to the person who asked the DnD question.
That answer to the D&D campaign question was so good. Depending on who you are, hearing it could be either terrifying or amazing.
Well that was exceptionally fantastic. Thanks for hosting and publishing that! Thanks to John Romero, too. Even though I'm a gamedev and often watch stuff like this to glean gamedev wisdom, I think my favorite part was hearing about John Carmack's D&D game.
This is the most complete story I've found so far. Excellent talk
Definitely check out the recent (5-hour long!) Lex Fridman podcast with John Carmack as well.
It is always interesting to listen to John's Romero wisdom
What a brilliant talk - honestly i wish it were longer
What a joy listening to John. Reflecting on my own life in software through the same eras and of course the awesome nostalgia of playing Doom and Quake. Carrying our whole system (with CRT) around to friends houses to deathmatch.
I think the most influential bug that became a feature was strafe jumping/bunny hopping in Quake, the fact that it is quite essential in some games to this day says a lot.
I could listen to Romero talk for hours
just like the development style, the presentation was quick and concise, leaving room for other things.
I was in high school for WOLF3D, college for DOOM, and at my first tech job for QUAKE3D. I
Id Soft made the best games I grew up with. Thanks so much for it! 🕹👍😀
I love how he just says that they did some insane, groundbreaking stuff that most of us could only dream of doing, so casually.
I'll never forget the amazement when I first saw Doom on my friends computer as a kid.
Nobody ever really seems to talk about Heretic or Hexen anymore, but even though I loved Doom, I preferred Heretic because I was more of a fantasy nerd than action in those days. I readd Lord of the Rings in the 70s and right after that D&D became huge. Then Quake 2 came out and I was a bit older, and into rock music and the music in Quake 2 blew my mind. I used to play in the living room with all the lights off and had my PC plugged into my sound system and it was insane.
John & Co, you made this Gen Xer a happy man to this day. Thanks.
The legend. Hats off.
Fun DnD story at 44:20
highlight of the talk.
Amazing speech, watched it in one breath with a smile on my face
17:00 Develop on a system that's superior to your target.
I can relate to that. In the early 90's I took over the group developing the software for an industrial machine that used a '286 motherboard. The development machines were PCs built around this same board, because my predecessor wanted the development system to match the target system.
It had a detrimental effect on the code, big time.
The developers were loathe to change a header, because of the long compilation times. So instead of fixing or updating a core data structure, they would come up with crazy work-arounds that could be done in one CPP file.
I came in with my personal '486SX budget machine, and quickly slashed through some of these outstanding issues, rebuilding clean in the time they took to incrementally make.
As for needing the same hardware, that applied to the installed cards not the base PC, and targeting the 286 for generated code, and having a "slow clock" or "de-Turbo" mode for testing.
Can confirm. Develop on good machines (with an OS that doesn't freeze after a bad pointer), and TEST on at least one baseline system.
As a kid who grew up on Doom IN Shreveport its funny to hear about id working off of Lakeshore Dr only a few miles from my childhood home.
So much insight to the art of making games! Thank you, John!
53 and still at it, C++, making useful things.
This was awesome! Thanks to all involved. 🤩
Awesome talk! Every aspiring indie game dev should watch this
12:28 Keep looking at your functions and figure out how you can simplify further.
These talks are awesome. Amazing people doing amazing things. Such a great story and always interesting to hear these guys share.
Now just to form a small team of multiple. exceptional, passionate, respectful geniuses and give them full ownership of their work. It’s easy!
IIRC, Death Rally, and Tyrian, were both successful 90's shareware/DOS games that started from demoscene groups. There's probably a ton more. There were a lot of Finnish DOS indie games with stellar soundtracks in the 90's.
Thanks for this amazing journey through your creative minds
Awesome talk! Thank you so much for talking, filming und uploading!
Incredible person what a great time and place to work they had!
Brillliant talk! I felt a slight contradiction in that slide about "writing your code for this game" and how all code was written afresh and then later admitting that engine code and elements were reused because they weren't going to get any better. I can imagine that from one perspective the game is new but the engine(s) are always improving.
I haven't fully watched it yet, this is mainly a note to myself to check if John tells what the bug that they had in the game since day one was and how they fixed it.
Edit:
30:56 is the timestamp where this is answered. Thanks, me. You're welcome!
Amazing video, superb insight. Moral of the story: Talk to each other!
That was amazing. It never gets boring hearing about that time and what they did.
Reminds me when your young you have so much energy. Such an interesting insight.
Small company. Good friends. Do what you love to do. Embrace the grind. Success follows.
I was doing web design in Madison in 1991. Only 5 servers for the web existed at the time and Madison had one. I found one of Wolfenstein shareware on our server.
this video should be double liked
It is
I always wondered how or why IBM put the scroll registers into the CGA, EGA, and VGA designs. I can remember when people sort of discovered it, and started using it. It was in "trainers" and demos. But why was it even there??? What was the intent? I don't think it was for game programming. And even then, it was just barely even usable. (I sure had a rough time, being a senior in high school, just starting to learn Pascal, ASM, and C/C++.)
So fun side-notes about all of that. The old video adapters had these scroll registers as you mentioned, and if you could configure the memory in such a way, you could draw to one part of video memory while displaying another. Once you were done drawing you could flip the "viewport" register to the other area and start drawing where you were just displaying. This double buffering allowed for flicker-free graphics. The problem was that no standard video mode was setup in a way that allowed for two screens of 8-bit color to fit in video memory at once.
You heard Romero mention Michael Abrash. Abrash's discovery was that with a bit of mucking with VGA registers you could kick the cards into a undocumented "unchained mode" allowing for 8-bits of color on a 320x240 display, meaning we could finally get much richer color with double buffering. In addition, there was just enough video memory left around that these two buffers could have a border around them of extra tiles allowing for smooth panning around the screen.
@@halgari Is that how they got 1024 colros on a 4bit-color IBM with the 8088 MPH demo?ruclips.net/video/yHXx3orN35Y/видео.html
CGA was built on top of HD6845 (extremely common CRT controller chip in 80s), which has scroll registers for obviously scrolling the character display fast, and not having to copy the whole buffer.
is always fun to hear about the early days of game dev
I'm going to have to watch it again...
Friendship is the biggest take-away...
No TDD, no *DD, no Agile, no Waterfall, no latest tech buzz
IdSoftware delivered 28 games in 5.5 years without version control system and just syncing between them. IdSoftware is mainly responsible for creating the gaming industry
That is WOW, today with all the buzz, "latest" software process and it still hard to deliver good software.
This talk will save you tone of money on "make a great software" consultants
In summary:
Sync with your co-workers, especially if you know that your code change can interrupt someone else work
My hypothesis: Most of the "processes" are created by observing the successful companies. But it is hard to observer core behaviors and core principles so what is left are some rituals.
The tighter and smaller the team, the less you need those time sucking processes.
@@neptronix Friends and personality match also helps. Without a good fit motivations of some members become a problem sooner or later.
working on something you are passionate about too
It was actually very agile.
"No prototypes. Just make the game. Polish as you go. Don't depend on polish happening later. Always maintain constantly shippable code." This was not the current age of the game industry.
Awesome, great video!
great talk
The PHP on the screen at 28:40 seems highly out of place... :D
tom hall's press shot is legendary (but also john's)
Great story and great advices!!
there are more proccesing calculations happening. pre processing 1. real time data proccesing 2. and post proccesing 3. good luck. curious ones. :p then, single tasking 1.switching-multitasking aka turn based 2. real time aka, simultanious tasking operations. "all at once" 3.
I do think it's funny that iD Software hired someone to do the SNES port of Wolf 3-D & got stiffed on it, then almost the exact same scenario played out when they contracted out Doom 1's networking code! Paid some guy who ran off with the money, they never got the networking code & ended up having to do it themselves.
Damn, I miss the early days of where gaming was born from. I was too young to really experience it. I was a part of the console wars that happened with Atari, Sega, Nintendo, NeoGeo, Turbografix, etc.
1:12:10 “Don’t you run from me, John!” - Decino 😂😝
@44:23 the Dungeons and Dragons campaign talk
Commander Keen 4 was so much fun. I wish someone had asked him if he had a favorite game of all the ones he worked on. Sounded like it might have been Doom.
20:03 does anyone know what the bug was?
59:10 Wise Mystical Tree?
30:46 for what 11th hour bug Carmack fixed on Doom launch day
John Romero:
Meanwhile
Sandy Peterson: working at id software was like watching a nature documentary about a pack of hyenas devouring a carcass
So awesome👍👍!
I'd pay to read more about that D&D campaign.
Great talk.. thanks so much :)
12:35 just a few hours ago I was still simplifying my code.
I have a question about the SNES port of Wolfenstein-3D.
You said you had to learn the hardware, convert assets, etc.
But was there a practical C compiler that targeted the 65816? You didn't mention having to learn assembly language and work out performance tricks on same.
presumably knowing 6502 from the apple 2 helped a lot
can someone PLEASE name the games at 6:06 ??
I like most of the core principles :-)
Dangerous Dave II my first ever shooter
I got distracted for a moment then came back in the middle of his story about casting spells and talking to demons, and I'd missed the opening where he'd explained he was playing Dungeons and Dragons with his friends. I must say I was extremely confused for a minute there ... haha.
I'm looking forward to Sigil 2
good principles I wish more people saw the wisdom in. common sense not so common in big tech
"We had no source control..." 💀
Where does Zoidphun fit into all this?
55:36 TIL that time == Wolfenstein.
Slordax was the first game I EVER played
Interesting talk, but Wikipedia tells a completely different, not as rosy story on the Quake development… also, how do you go from multiple games per year to Daikatana? Seems the most interesting part was left untold…
17:33 *nix FTW
Makes you want to start hacking on some games. I remember being in like kindergarden and playing Dangerous Dave off a floppy disc.
33:33: Sounds like a constant continuous hackathon.
Lovely interview with a vampire.
always fun to listne to John
Some of the smartest developers even today all together back then. That’s more than just luck
It is time for another "Time Warp" podcast John. Just saying.
So actually, he doesn't say write the entire game from scratch.
Sorry John but Keen didn't pioneer game engines. Just off the top of my head I can think of Gary Kitchen's Gamemaker (1985), SEUCK (1987), Freescape (as used in Driller, 1987 and several later games like Dark side, Castlemaster etc) and there were several other game engines before Keen arrived.
WTF, We still use hardware breakpoints in games dev now. On all platforms. On some its called a data breakpoint though.
I played and owned EVERY game made by John Romero and John Carmack. Not one of their games had any bugs. Not one. This is a practice of work ethic that has often fallen to the way side in favor of greed, early access garbage, and lazy people! Every game designer needs to emulate their impeccable work ethic. It is a lie that you can not make games without bugs. A game not mentioned was called Blake Stone, which I thoughourly loved from Apogee. One last thing I will mention is that I have reservations concerning the occult. I would never outwardly engage in any type of magic, witchcraft, or satanism, and that includes D&D.
so, the buffer, a dedicated piece of ram. is limited. and filled with data, from multiple sources say software code and keyboard input and hard drive data. and empty with a clockcycle say X times a second. and then refilled. the data goes to the buffer. say, one by one, and the buffer (or cache) dispose that all together in go, too different things, like music and display.
and then the buffer memory is filled with a higher frequentie or fps then it is emptied. say i.e. 30 hertz per proccesed data stream. per 5 megabyte. per s3cond. and then emptied 1 once a second. buffer under run. ( no use the full buffer memory array. and purge it before it reach its limits. ) buffer limit is by the programmer Pre-designed. or Pre-fixed. in size.that along side the rest of the ram memory usage. thus, " in the meanwhile" or " in the lean time" and alsobhas its own clock freqeuncy. good explaination sir, romero.