Savage Island Part 2: ROBOPIRATE   5 comments

Okay, things got good.

They got good enough that I’m going to give fair warning about plot spoilers — if you’re genuinely wanting to play this (once you’re past the breath control stuff it gets better) you might want to veer away now. (You’ll need to have played the first game or at least read my own playthrough from last year. There’s an online BBC Micro version that runs well.)

Also, before I dig in, I realize I neglected in my first post to mention that Scott Adams has a co-author credit on this one: Russ Wetmore. A year later Mr. Wetmore went on to write the Frogger clone Preppie! for Atari computers. He’s still active to this day in software, just not in games.

I had previously been stuck in a small area with a PSYCHOTRANSFIGURATION device and a Neanderthal. The device is in a “red metal room” and it is next to another empty “red metal room”. I had been thinking (given the game’s description of the Neanderthal as a “non-transfigured controller”) that I needed to lure the Neanderthal over to the empty room and then start the device. The Neanderthal sometimes follows and sometimes lags behind (so the idea was technically sound) but unfortunately, I could never get them to step into the right room.

I eventually resorted to a hint (early, I know, but HYPERVENTILATE was enough to make me wonder if the rest of the game would be equally hard) and found out that the flower that smelled like a sedative could be smashed.


I feel
I am dead.

(I would normally use BREAK but that’s the BREATHE verb — the game understands the first four letters of each word as opposed to three as in previous Scott Adams games, but that’s still enough to cause overlapping. CRUSH works too.)

Fortunately, I had just undergone breath control boot camp and knew what to do next: HOLD BREATH (which you can do before releasing the Neanderthal) and then SMASH or CRUSH the flower while he’s in the room.

He’s then sedated, although trying to TAKE NEAN leads to “How? too large”.

I figured this out almost immediately after so I suppose it’s better at emphasizing the “weight” of the action over just GET, but I could still see someone getting unnecessarily stuck here by visualizing the PULL and GET actions as equivalent.

I deposited the Neanderthal in the empty room as I was theorizing, went to push the button, and got a different reaction than before.

metallic voice whispers in my mind:
“Vocalize password Please”

I confess I had to reach for another hint. I’m slightly embarrassed by this one; I knew from the previous game I was already carrying over the bandanna, so I had the notion of carrying physical objects from Part 1, but it didn’t occur to me I might also need to carry information from Part 1. Here’s the relevant screenshot from that game:

The password here (FREE) was used in the previous game, but it also is used here. While I’m pulling out old screenshots, let me yank out one more which I’ll be referring back to.

Back to Part 2. Applying the password:

The “Rayshield” mentioned is the bandanna carried from Part 1; I was wearing it so I solved the puzzle by default, and found out what it was doing when I looked up the password puzzle.

It took me a few beats to realize PSYCHOTRANSFIGURATION was referring to brain transfer. You are now the Neanderthal.

And with one stroke, Adams/Wetmore pulled off what is likely the first change in player character for an adventure game. There’s still lots of 1981 to go, of course, but still, this was a jaw-dropping moment for me.

The Neanderthal can read the alien script on the dials. I assume the aliens genetically engineered the knowledge in.

I can
read it!
“Pull lever to reverse Transfiguration Process”

The Neanderthal also has more lung capacity than the original character, and so can last in vacuum longer and make it all the way to the blinking-red-light force field.

There’s two “me”s going on now — the Me as the now-unconscious start character, and the “me” being referred to here.

The OPEN EYES needed when waking up as the Neanderthal trained me in knowing I could CLOSE EYES as well.


I am in a metal room. Visible items:

Force field. Control console. Treadmill. Large medicine ball. Freestanding punching bag.

I haven’t played much with the devices here yet. You can GO TREADMILL which leads to a “room” that I can’t get out of (I’m guessing parser trouble). PUNCH BAG busts the punching bag and leaves scraps of cloth; I’m going to try bringing it back to my original body and see what happens with it when I switch minds back and punch it there.

The rooms nearby (see map above) include a main control room…

…and a cabin with a “small plastic wafer” which was sufficient for me to realize the plot.

Let me summarize:

Aliens came to Earth to seed humanoids. In order to seed them, they needed to remove the dinosaurs first. A mutiny happened and the mission was never completed. Without the Neanderthal seed being placed on Earth, humanoids will never develop. Our hero was thrown back in time to the alien ship. If the alien ship is destroyed, there is a time paradox as a human destroys the method that humans exist in the first place.

(You can destroy the ship early on by picking up a “plastic block” which turns into a “GLOWING plastic block” and then: “EXPLOSION! Time paradox shatters reality”.)

This means our quest is to kill the dinosaurs, seed the first humanoids, and save reality.

I’m hoping I get to use the command >DROP METEOR later.

Also note the picture with “me” in on the cave wall from Part 1, or even the “Large stone head” from the very first room of Part 1:

Our hero has already been memorialized for deeds that they have yet to do!

Posted March 11, 2020 by Jason Dyer in Interactive Fiction

Tagged with

5 responses to “Savage Island Part 2: ROBOPIRATE

Subscribe to comments with RSS.

  1. This is properly clever stuff.

    I wonder what Jimmy “Filfre” Maher would make of this one: he’s argued on occasion (e.g. comments to <a href="; this post) that the limited parser and world model of Scott Adams games and similar made it impossible to design complex and satisfying puzzles, Savage Island, part 2 feels like it might be a disproof by counter-example.

    • 1. Is it possible to create good-but-tough puzzles with a two-word parser?

      Jimmy mentions the Phoenix games in the thread you pointed to; I’d say it does need a “good parser” in that it has to have some responsiveness to incorrect and/or unrecognized input. It’s possible to have a multi-word parser that’s still bad in this sense (Haunt had a couple parts like this; Warp also had some moments where I had to guess-the-phrase).

      2. Is it possible to create good-but-tough puzzles with a bad world model?

      I’ve argued this was the main issue with early adventure writers who had ambitions for more difficulty, like in Devil’s Palace and Poseidon Adventure. Without a good model, you’re stuck with apply-verb-to-right-spot and apply-object to right spot, and making those difficult requires a.) hiding the verbs or b.) hiding the spots or c.) hiding the objects, all which are limited.

      3. Does the Scott Adams system allow both a decent parser and a decent world model?

      The parser is clearly a cut above the “I CANT” system from Planet of Doom. I feel like it’s hanging barely on the fringes and there are certain circumstances where more responses would be helpful, but the 16K limit was getting pushed to the max.

      As far as world model goes, yes, clearly here there are very interesting things going on, so it’s *capable* of some interesting world modeling. I wouldn’t underestimate ease-of-use as a consideration in all this, though; while the use of daemons goes all the way back to Adventureland (see the bees, for instance) having the complexity of, say, the Spellbreaker spell list might be technically possible but require serious coding chops. Back when I was writing about Arnstein’s Haunted House I observed how it was a technical marvel of sorts with 4K being used on each side; I had a commentator claim it’d be easy to develop the game because they’d just “develop in C and cross-compile”, which is _technically_ doable but the general 1979 knowledge for such (and availability of C past UNIX) made that prohibitive. So I’d be careful with just stating “there is the possibility of coding this complicated thing” as being equivalent to meaning technical hurdles were not a reason for some of the badness of old puzzles.

      I incidentally don’t believe Jimmy played the Savage Island games, I remember some comment of his more or less implying he was waiting for me to do it.

      • Here we come to the probem of what we mean by “parser”.

        I’ve always understood it to refer to the technical capabilities of (on the one hand) just recognising two-word commands of the GET LAMP type, via slight cosmetic improvements such ignoring articles to understand GET THE LAMP, though into the ability to understand genuinely richer command — notably those involving more than one object, as in TIE ROPE TO RAIL.

        But as I’ve read blogs like yours, Jimmy’s and Nathan P. Mahney’s CRPG Adventures, I have grown increasingly reconciled the use of “parser” to mean something more like “the way the game responds to various inputs” — so, for example, recognising PUSH and a synonym for MOVE is seen as a parser feature.

        Anyway — while Scott Adams games (in the broad sense of all games written using his system or one of equivalent power) are obviously limited as to their parser in the first sense (two words only), there is no reason why within those technical constraints a developer could not anticipate various incorrect or nearly-correct inputs and respond in a useful way. What’s more surprising is that more complex actions don’t seem to be a big problem either — as in “THROW AXE / What should I throw it at? / AT SHORE”. Admittedly, implementing something like Enchanter‘s DRAW LINE FROM D TO F this was would be a pain — but it would not be a fundamental roadblock.

        What probably would prevent you from implementing Enchanter in the Scott Adams system would be the inability to represent and manipulate state such as the set of line currently drawn between the various lettered points.

        The intriguing thing about Savage Island, part 2 (based only on your blog-posts; I’ve not played it) is that it seems to be implementing a rich world model — not as rich as Enchanter, but notably richer than, say, Adventureland. Despite being written in the same system.

      • Accepting “THE” and all that are nice but not strictly relevant to if you can have complicated puzzles — clearly a lot of the Euro 8-bit market managed.

        On the other extreme, having a parser that *only* understands the “correct” answer to a problem makes it so requiring anything even slightly unexpected from the user causes the difficulty to shoot up to 11. (This causes the puzzles being either super-easy or super-duper-hard with no in-between problem.)

        Savage Island Part 2 presents an interesting thought experiment with its parser (which I’ll likely save for my last post, because I want to know if more things come up) but generally: suppose you really wanted a puzzle involving controlling your breath. How would you make it scrupulously fair? The game as is already understands a decent amount and having more than two words doesn’t really help.

  2. Pingback: Savage Island Part 2: Finished! | Renga in Blue

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: