Rendered at 18:52:43 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
mromanuk 2 days ago [-]
I know what is AGI, I read the queues Adventures, Sierra, should be obvious where this lead. But my mind went to "oh, thanks to LLM coding agents, they massively ported every Sierra Adventure to C64".
pimlottc 2 days ago [-]
*cues
JoshTriplett 3 days ago [-]
The only good AGI.
Impressive fitting into the machine. I wonder if SCI would be possible?
bzzzt 2 days ago [-]
Only with some serious downgrading of the assets. I remember playing SCI games on an 8MHz XT class PC and everything ran dog slow. Some SCI games had optional animations that didn't play on a slow machine.
aidenn0 2 days ago [-]
AGI was an impressive game engine. It's just a shame it was wasted on so many bad games.
JoshTriplett 2 days ago [-]
Some of them weren't wildly worse than the era. I think many old games substituted punishing difficulty for depth, so that they'd have more staying power. That was true of some of the early NES games as well.
aidenn0 20 hours ago [-]
I agree that they weren't terrible for the era, but they weren't great either.
I'll pick on SQ1, since that's the highlight of TFA. Obtaining the Jetpack is not just punishingly difficult, even once you get to the point where you know you need a jetpack there are zero hints where you might obtain one; the only solution is to replay the game and try random things differently than the first time you played through.
JoshTriplett 2 hours ago [-]
Yeah, agreed. Many games did the same, but that didn't make it good.
On the "cruelty scale", that'd be a "Cruel", since it has a "walking dead" state of "you did it wrong a while back and have to start over":
Early King's Quest games were in that same category of "Cruel", where an incorrect choice is not immediately obvious as a mistake.
I think King's Quest V was closer to "Nasty" (because it can catch you off guard with things you fail, but it'll be obvious that you did something wrong). But there are a couple of exceptions that keep it in "Cruel", though: you can eat the wrong item when you're starving, and in some cases if you miss picking up an item you can't go back for it, so if you didn't see it at all then you have to go back to an earlier save. That one definitely required keeping a full array of save files. (Very satisfying to complete, though.)
Even King's Quest VI, which is much more forgiving, has some cases of "you need the right item before proceeding into this section you won't be able to leave (for a while)". But it goes out of its way to advertise those "points of no return", so I'd rate it firmly at "Nasty" rather than "Cruel", with the asterisk that there are ~3 notable points where you might have to go back by more than just your most recent action, but it'll be obvious you want save points at those points to go back to if you find you can't progress inside. (Before the catacombs, before the land of the dead, before entering the castle.) The entry to the catacombs even offers you the chance to go prepare further, if and only if you don't have everything you need.
aidenn0 2 hours ago [-]
The cruelty scale is somewhat orthogonal to difficulty, and misses another axis; you can have punishingly difficult puzzles seem to have logic, and punishingly difficult puzzles that do not.
Ending up in "walking dead" with no hints that you are in that state means "Cruel" but it's different depending on where things go next. If the guy buying your skimmer in SQ1 had a visible jetpack, then once you know you need a jetpack later in the game, you have a hint as to where to go back to. This would make it just a slightly more cruel form of "trial by death" (which was kind of expected in a lot of early adventure games).
w.r.t KQ5, not saving the mouse from the cat is something that will get you if you haven't played enough Sierra games to know that missing a single random event with a short timer is likely enough to render the game unwinnable; when I played it I just said "damn, I think I need to reload" when the mouse died. IIRC there are no hints in the cellar that saving the mouse would have helped you, but I could be wrong.
JoshTriplett 1 hours ago [-]
> The cruelty scale is somewhat orthogonal to difficulty, and misses another axis; you can have punishingly difficult puzzles seem to have logic, and punishingly difficult puzzles that do not.
I'd agree with that. I dislike "read the author's mind" puzzles.
> w.r.t KQ5, not saving the mouse from the cat is something that will get you if you haven't played enough Sierra games to know that missing a single random event with a short timer is likely enough to render the game unwinnable; when I played it I just said "damn, I think I need to reload" when the mouse died.
I would classify the mouse as "Nasty", yeah; you know immediately that you missed something.
But, by contrast, you can eat the custard pie rather than the leg of lamb and be in a walking-dead state (and it's not even obvious when you reach the yeti, you don't go "if only I still had a custard pie"); that one is definitely "Cruel".
There are many cases of Cruel that are simply "fail to pick up an item in a place you can't return to" cases (which are more forgivable, I think). But even if you ignore those, there are many other cases of Cruel walking-dead states.
Getting seen by the cat in the castle (not completely obvious you've done something irrevocable) will result in you getting caught later. And the most annoying, I think: if you are too good at the game's moon logic, you can use the peas on the monster the first time it tries to capture you, never get caught, and thus fail to obtain the required item (cheese) from the cell. That one is firmly "you think you solved the puzzle and can proceed" walking-dead.
(Also, the fact that Mordack can just randomly show up in the castle and kill you is just gratuitously obnoxious. You can do everything right, and still fail for reasons of complete randomness that you can't avert.)
vunderba 2 hours ago [-]
KQ4 suffers many of the same issues.
1. Reach location A.
2. You need to "get equipped with Item X" here, though it's for a future puzzle you haven't even encountered yet, so you don't even realize you need said item.
3. Reach location B with the puzzle.
4. Realize you don't have the item required.
5. Location A is permanently inaccessible, so you can't even go back and look for it.
The fusion of "64-bit architectures" and "64K total RAM machine" brings a bit of a smile to my face with projects with names like this.
a1o 3 days ago [-]
Wait, how? What is the background technology? I need more explanation. Is this a tool to recompile these games?
wzdd 3 days ago [-]
AGI is here and its first task is obviously completing the C64’s game library.
Not really. AGI is a game interpreter, like LucasArts’ SCUMM or Infocom’s z machine. Games are shipped as portable code and a machine specific AGI interpreter runs them. Looks like they’re preprocessing the graphics first, though. https://en.wikipedia.org/wiki/Adventure_Game_Interpreter
skissane 3 days ago [-]
Sierra released most of their AGI games on the Apple IIc/IIe. Given C64 is a roughly equivalently specced machine, there is no reason in principle why they couldn't have ported AGI to C64 as well – it just, for whatever reason, never happened. (One AGI game, "Donald Duck's Playground", was originally developed as a native C64 game, then AGI was used to port it to other platforms.)
Also, given different AGI platforms had different graphics capabilities, I expect even the original Sierra AGI games had somewhat different graphics assets on different platforms – so whatever preprocessing of graphics they are doing to make them C64-friendly, Sierra quite possibly would have done something like that anyway had they made a C64 port back in the day.
TMWNN 3 days ago [-]
> Given C64 is a roughly equivalently specced machine, there is no reason in principle why they couldn't have ported AGI to C64 as well – it just, for whatever reason, never happened.
If I recall correctly, the art is done with a form of early vector drawing so that they can use one set of assets on all platforms. They're designed to 16 colors but get downsampled for 4-color CGA mode or monochrome mode.
glhaynes 2 days ago [-]
I've always imagined it to be first and foremost about storage savings. Although, come to think of it, an RLE-encoded bitmap might be comparable since we're talking about images with large areas of so few colors. Might be interesting to do an empirical test.
But, yes, it's all a bunch of vector draw/fill commands. The first few AGI games showed that process happening; and, as a little kid having one of their first encounters with a computer, I found it fascinating to watch. Later games drew it in an offscreen buffer and showed it all at once, which looked more "professional" but took away some magic.
EDIT: I had an agent make a guess and it came up with reasonable looking numbers that said an RLE bitmap would be a small handful of times larger. But it also interestingly pointed out that the vector commands encoded priority (i.e., what you can walk behind/through) whereas that'd have to be a second data set for a bitmap approach.
WorldMaker 2 days ago [-]
The vector art also included some complex dithering patterns that balloon RLE-encoded bitmaps. Dithered drawing is actually one of the worst cases for simple, pure RLE. For dithering compatible bitmap encoding you would want other more complex compression schemes than simple RLE. Some of the palette-based compression techniques of "modern" GIF were reused from such early predecessors, even though there's a belief today that RLE was most of what games at the time used. (It's useful to point out that early AOL was also heavily dithered, especially in monochrome versions. AOL also spanned the gap between dithered, heavily palette-based graphics to "early modern graphics".)
kolanos 2 days ago [-]
Yes, this is correct. I have a distinct memory of playing Space Quest (forget if I or II) on a CGA monitor and then seeing it on my friend's EGA monitor and being blown away.
noufalibrahim 3 days ago [-]
I think the first instance of games being separated from the "engine" so that you could make the game and it would run on all platforms that the engine ran out without porting was Infocom's Z-machine. http://maher.filfre.net/if-book/if-4.htm
There was an interesting article which I read a long time. It's linked to from the post above called "How to Fit a Large Program into a Small Machine" published in 1980.
AGI, SCI, Scumm etc. were all larger and more capable versions of this.
Z-Machine is also useful to point out because it's one of the clearest lines to draw from it to the modern JVM and .NET CLR. Whereas AGI/SCI/SCUMM all have very game specific virtual machines, Infocom's was Lisp influenced and even built it with business/enterprise apps in mind, though Infocom's one attempt at doing exactly that with their database app Cornerstone was a market flop. (It's heavily enhanced fork of the Z-Machine was also only released for a fraction of the machines the game Z-Machine supported.)
I still think it's an interesting question if Infocom had not tried to keep the Z-Machine proprietary "secret sauce" if it would have been a larger platform of general computing for much longer and broader than Infocom's own lifespan and primarily games focus. If Infocom had taken a Sun-like stance on the Z-Machine being a JVM ahead of its time and building a broader compatibility ecosystem "write once, run everywhere" might have been an earlier part of computing.
vkaku 2 days ago [-]
We finally achieved AGI outside these AI companies on portable home computers
1970-01-01 2 days ago [-]
>Made for a 1 MHz machine
That's impressive work. It's fun to think how today's CPU rounding errors can run hit adventure games. Your smartwatch uses more cycles to render the time.
Impressive fitting into the machine. I wonder if SCI would be possible?
I'll pick on SQ1, since that's the highlight of TFA. Obtaining the Jetpack is not just punishingly difficult, even once you get to the point where you know you need a jetpack there are zero hints where you might obtain one; the only solution is to replay the game and try random things differently than the first time you played through.
On the "cruelty scale", that'd be a "Cruel", since it has a "walking dead" state of "you did it wrong a while back and have to start over":
https://eblong.com/zarf/essays/cruelty.html (and https://eblong.com/zarf/essays/cruelty-revisited.html )
Early King's Quest games were in that same category of "Cruel", where an incorrect choice is not immediately obvious as a mistake.
I think King's Quest V was closer to "Nasty" (because it can catch you off guard with things you fail, but it'll be obvious that you did something wrong). But there are a couple of exceptions that keep it in "Cruel", though: you can eat the wrong item when you're starving, and in some cases if you miss picking up an item you can't go back for it, so if you didn't see it at all then you have to go back to an earlier save. That one definitely required keeping a full array of save files. (Very satisfying to complete, though.)
Even King's Quest VI, which is much more forgiving, has some cases of "you need the right item before proceeding into this section you won't be able to leave (for a while)". But it goes out of its way to advertise those "points of no return", so I'd rate it firmly at "Nasty" rather than "Cruel", with the asterisk that there are ~3 notable points where you might have to go back by more than just your most recent action, but it'll be obvious you want save points at those points to go back to if you find you can't progress inside. (Before the catacombs, before the land of the dead, before entering the castle.) The entry to the catacombs even offers you the chance to go prepare further, if and only if you don't have everything you need.
Ending up in "walking dead" with no hints that you are in that state means "Cruel" but it's different depending on where things go next. If the guy buying your skimmer in SQ1 had a visible jetpack, then once you know you need a jetpack later in the game, you have a hint as to where to go back to. This would make it just a slightly more cruel form of "trial by death" (which was kind of expected in a lot of early adventure games).
w.r.t KQ5, not saving the mouse from the cat is something that will get you if you haven't played enough Sierra games to know that missing a single random event with a short timer is likely enough to render the game unwinnable; when I played it I just said "damn, I think I need to reload" when the mouse died. IIRC there are no hints in the cellar that saving the mouse would have helped you, but I could be wrong.
I'd agree with that. I dislike "read the author's mind" puzzles.
> w.r.t KQ5, not saving the mouse from the cat is something that will get you if you haven't played enough Sierra games to know that missing a single random event with a short timer is likely enough to render the game unwinnable; when I played it I just said "damn, I think I need to reload" when the mouse died.
I would classify the mouse as "Nasty", yeah; you know immediately that you missed something.
But, by contrast, you can eat the custard pie rather than the leg of lamb and be in a walking-dead state (and it's not even obvious when you reach the yeti, you don't go "if only I still had a custard pie"); that one is definitely "Cruel".
There are many cases of Cruel that are simply "fail to pick up an item in a place you can't return to" cases (which are more forgivable, I think). But even if you ignore those, there are many other cases of Cruel walking-dead states.
Getting seen by the cat in the castle (not completely obvious you've done something irrevocable) will result in you getting caught later. And the most annoying, I think: if you are too good at the game's moon logic, you can use the peas on the monster the first time it tries to capture you, never get caught, and thus fail to obtain the required item (cheese) from the cell. That one is firmly "you think you solved the puzzle and can proceed" walking-dead.
(Also, the fact that Mordack can just randomly show up in the castle and kill you is just gratuitously obnoxious. You can do everything right, and still fail for reasons of complete randomness that you can't avert.)
1. Reach location A.
2. You need to "get equipped with Item X" here, though it's for a future puzzle you haven't even encountered yet, so you don't even realize you need said item.
3. Reach location B with the puzzle.
4. Realize you don't have the item required.
5. Location A is permanently inaccessible, so you can't even go back and look for it.
Not really. AGI is a game interpreter, like LucasArts’ SCUMM or Infocom’s z machine. Games are shipped as portable code and a machine specific AGI interpreter runs them. Looks like they’re preprocessing the graphics first, though. https://en.wikipedia.org/wiki/Adventure_Game_Interpreter
Also, given different AGI platforms had different graphics capabilities, I expect even the original Sierra AGI games had somewhat different graphics assets on different platforms – so whatever preprocessing of graphics they are doing to make them C64-friendly, Sierra quite possibly would have done something like that anyway had they made a C64 port back in the day.
Agreed about the issues not being (completely) technical. <https://np.reddit.com/r/c64/comments/1r8hgwk/kings_quest_i/o...>
But, yes, it's all a bunch of vector draw/fill commands. The first few AGI games showed that process happening; and, as a little kid having one of their first encounters with a computer, I found it fascinating to watch. Later games drew it in an offscreen buffer and showed it all at once, which looked more "professional" but took away some magic.
Here's a video showing the drawing process: https://www.youtube.com/watch?v=v5DSFvWrNWA. Magic!
EDIT: I had an agent make a guess and it came up with reasonable looking numbers that said an RLE bitmap would be a small handful of times larger. But it also interestingly pointed out that the vector commands encoded priority (i.e., what you can walk behind/through) whereas that'd have to be a second data set for a bitmap approach.
There was an interesting article which I read a long time. It's linked to from the post above called "How to Fit a Large Program into a Small Machine" published in 1980.
AGI, SCI, Scumm etc. were all larger and more capable versions of this.
https://github.com/johanberntsson/PunyInform
Tristam Island as an example:
https://github.com/hlabrand/tristam-island
https://github.com/johanberntsson/ozmoo
https://ozmoo.online/
I still think it's an interesting question if Infocom had not tried to keep the Z-Machine proprietary "secret sauce" if it would have been a larger platform of general computing for much longer and broader than Infocom's own lifespan and primarily games focus. If Infocom had taken a Sun-like stance on the Z-Machine being a JVM ahead of its time and building a broader compatibility ecosystem "write once, run everywhere" might have been an earlier part of computing.
That's impressive work. It's fun to think how today's CPU rounding errors can run hit adventure games. Your smartwatch uses more cycles to render the time.