Sign in with
Sign up | Sign in
Your question

REVIEW: Sentinel: Descendants in Time

Tags:
  • PC gaming
  • Games
  • Video Games
Last response: in Video Games
Share
Anonymous
February 5, 2005 7:45:09 AM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

REVIEW: Sentinel: Descendants in Time

(Review copyright 2005, Andrew Plotkin <erkyrath@eblong.com>)

Graphics: good
Atmosphere: pretty good
Story: good
Writing and dialogue: not bad
Puzzles: very good
Difficulty: not very hard
Forgiveness rating: you cannot die or make a fatal mistake.

I open the _Sentinel_ manual and the second thing I see (after the cursor
explanation, same as every other adventure cursor in history) is "Story by
Terry Dowling -- based on his short story "'The Dormeuse and the
Ichneumon'". Huh, I say. Dowling is not well-known in the US; he's an
Australian science fiction writer, who typically deals in far-future human
civilizations. I have a couple of collections of his stories, but they don't
include this one.

Since most adventure games are "based on a thin idea which the designer came
up with, not that he's a writer or anything, it's science fiction, how hard
can it be?" I was prepared for something interesting.

_Sentinel_ is pretty interesting. I don't think it's a great adventure-game
translation of a short story. The story is a bit murky and hard to get into.
But then, I'm judging it by short-story standards, not adventure standards.
Which means _Sentinel_ has crossed some kind of watershed line: I'm playing
it *as* literate science fiction, written by an actual author, and how many
games can provoke that?

The scenario (just to describe the game's opening) has you entering Tomb 35,
one of the famous tombs of the lost Tastan civilization. You've been into
Tastan tombs before, but 35 is apparently one of the most dangerous. You're
only going in because some thug has your little sister hostage. He wants
tomb treasure; you're going to have to get past the Tastan defense program,
the Dormeuse ("Sleeper"), to find some.

So here's why it doesn't work: not enough text. (I bet you knew I'd say
that.)

There's a lot of text. The designers are doing the right thing: the
storyline is presented through a series of dialogue snippets, scattered
thickly throughout the game. Some of them are voice-overs; some are full
cut-scenes, in which you can see Dormeuse moving around and interacting with
scenery. None of the snippets are too long, and the voice acting is quite
good. (Okay, Dormeuse is quite good. You are played by an annoying whiner
who never manages to get his head out of the script. But Dormeuse is the
interesting one.)

Despite this, I never quite had a sense of what was going on. Or rather, I
never had a sense of how the story *began*. The opening monologue is highly
compressed. Yes, it presents everything the story needs. When I finished the
game, I went back and realized that the opening provides foreshadowing as
well. It's a very well-done bit of prose. But it's still too squished. The
background material needs to be reinforced a couple of times, early in the
game, or the player just winds up feeling lost.

Note that I'm not talking about the story which emerges through the course
of the game. That develops over many dialogue interludes; the pacing works.
I'm talking about the protagonist's *context* -- his initial state of mind,
within which you must begin for the story to hold together.

For example: the game is obviously going to involve learning about the
ancient Tastans. Just by showing a tomb, and saying that you're going to
explore it, the designers set up that much. What they *don't* set up
(clearly enough) is how much you know *initially* -- what you *expect* the
tombs to contain. What kind of ancient tombs are guarded by a holographic
defense program? Are artificial intelligences normal where you come from?
Are your people even more advanced than the Tastans? Have you grown up
hearing stories about the Tastan sentinels, or did you just learn about this
one?

Again, it's not that the designers *omit* this stuff. The first time you
meet the defense program, you're walking into the tomb entrance calling out
"Dormeuse! Dormeuse!" So you can deduce that the protagonist *does* expect
to encounter her.

But that goes by very quickly. Everything does. In a printed story, every
line of description and narrative would convey tiny clues about the the
protagonist, his background, his expectations and understanding of the
world. It's the basic trick of science fiction. In _Sentinel_, this material
is compressed into the opening and the first few pieces of dialogue, and it
just doesn't quite fit.

Enough of that disgression. How does the game play out? Well, a Tastan tomb
apparently contains a set of virtual worlds, which represent places the
owner loved. Unsurprisingly, these worlds are full of puzzles.

(The game takes pains to stress that they're only *snapshots* of real
worlds. This, I assume, is meant to explain the Great Adventure Convention
-- that any real place would have so many perfectly-designed and arranged
puzzles. Personally I recommend that designers leave the subject alone. The
more you point, the more irritating the question becomes -- no matter how
good your explanation. Like a Star Trek episode trying to seriously explain
the Klingon forehead thing. Which, I hear, they're about to do. Sigh.)

The puzzles are a weird bunch. They are uniformly, and unashamedly, abstract
puzzles. *All* of them. Pattern-matching, pattern analysis, combination
locks; colors and symbols and sounds. (Many audio puzzles -- you don't need
perfect pitch, but you'd better be good at distinguishing sounds.)

You will find *no* puzzles based on physical reality. No heating things with
fire; no chilling with ice; no objects pushing or bumping or leaning on
other objects. No levers. Certainly no chemical or biological properties.
Not even any pipes or wires to trace from location A to location B. A couple
of the puzzles touch on the geometry of the locations, but only in the most
abstract way. There's a floating bridge at one point, whose height varies
with the height of the water; but you could just as well think of it as a
three-position switch.

In short, _Sentinel_ has scenery and puzzles, and the two have practically
nothing to do with each other. The game worlds are pretty, but you don't
have to *inhabit* them to understand the puzzles. They're just a tray to
serve forth the symbolic information you need.

The puzzles don't relate to the storyline, either. They're built into the
domains of Tomb 35, and at the end of each domain you get a crystal which
advances the game state. But there's no thematic relation between the
puzzles and the domains (or the Dormeuse, or the plot). Your actions do not
tie into the storytelling. Not until the very last puzzle, and even there
the thematic relation is optional -- you can simply solve it as a pattern
puzzle.

I'd compare it to _Jewels of the Oracle_, except that _Sentinel_ does *have*
a story.

Also, now that I think about it, _Jewels_ had some very complex puzzles --
ornate slider variations which required multiple levels of manipulation. The
puzzles in _Sentinel_ never get that fiendish. Some of them are
*complicated*, in the sense of requiring you to discover and organize a lot
of information; but once you have it all written down, the puzzle mechanism
itself is usually simplistic.

This is not exactly a complaint. I like abstract puzzles. I enjoyed
_Sentinel_ as a puzzle collection -- with some high spots and low spots, of
course. I just prefer games in which the challenges are inextricably tied to
the world. The pattern-matching is not the reason I keep going back to the
adventure wellhead.

(The worst puzzle in _Sentinel_ is a blinky-lights affair in which
thirty-two switches control twenty lights. You have to pick the four correct
switches. As far as I can tell, there's *no* way to solve this except to
write down *all 32* switch patterns, and *then* start combining them by
brute force. I wound up writing a program to solve it for me, which at least
saved the brute-force step.)

_Sentinel_ has a fully 3D environment. This is a recent trick which I expect
to see more of in adventure games, and _Sentinel_ does it very well,
although not brilliantly. The necessary comparison is to _Uru_, and -- well
-- _Uru_ looks better. _Sentinel_ has a lot of excellent scenery, and
there's plenty of model detail. But the designers rarely added that one
extra layer of dirt, gloss, wear, or glow which made _Uru_ so stunning. The
surfaces look just a bit flat and repetitive.

Navigation is no problem; you run around in the usual first-person 3D
manner. Mouse to turn, arrow keys to move or sidestep, left-click to trigger
whatever's directly in front of you. (The cursor is locked to the center of
the screen.) My only problem was that the mouse was very jittery; I had to
turn the sensitivity all the way down. Oh, and there is a jump key -- but
it's a tiny hop, not any part of the gameplay. I think jumping is only there
in case you get wedged against a floor polygon in a funny way. (Which you
can probably also solve by backing up and trying again from a different
angle.)

One odd interface addition is a "directional hint" -- a translucent arrow
which floats near your cursor and points at the nearest interactable
hotspot. (Or several arrows, if you're near several machines.) I'm not sure
I like it. It's *effective*, in that it's usually easy to follow it to a
nearby clue or puzzle. But it's also somewhat distracting. And if the
"nearest" hotspot is on the other side of a wall, it can become
frustratingly useless -- you're on your own as far as *reaching* the item,
if it's even reachable at that point in the game.

I'm not necessarily opposed to out-of-game interface mechanisms like this.
But I worry that they can become substitutes for exploring, rather than
helpful additions. The directional arrows in _Sentinel_ rarely told me
anything useful, because the game design was generally good enough to make
important objects stand out. And the arrows may have added to my sense of
disconnection with the game world. On the other hand, I didn't *need* to
feel connected to the game world, since _Sentinel_ is all about the puzzles.
So maybe I'm playing the game correctly, and just unloading my discomfort on
the directional arrows.

*Overall:* A mixed report. _Sentinel_ has an interesting story, which is
imperfectly presented. It has a lot of enjoyable puzzles, but it doesn't
even try to integrate them into the story or the world. Exploring it is
cool, but the scenery is just background art, and it didn't knock my socks
very far off. _Sentinel_ is very well done for what it is: a pretty puzzle
collection with a bonus story included. You should play it if that's what
you like. But I wish it had been more.

(This review, and my reviews of other adventure games, are at
http://www.eblong.com/zarf/gamerev/index.html)

--Z


"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
I'm still thinking about what to put in this space.

More about : review sentinel descendants time

Anonymous
February 5, 2005 10:07:18 AM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Andrew Plotkin wrote:
>
> (The game takes pains to stress that they're only *snapshots* of real
> worlds. This, I assume, is meant to explain the Great Adventure Convention
> -- that any real place would have so many perfectly-designed and arranged
> puzzles. Personally I recommend that designers leave the subject alone. The
> more you point, the more irritating the question becomes -- no matter how
> good your explanation. Like a Star Trek episode trying to seriously explain
> the Klingon forehead thing. Which, I hear, they're about to do. Sigh.)

Really?
It better be good.


> (The worst puzzle in _Sentinel_ is a blinky-lights affair in which
> thirty-two switches control twenty lights. You have to pick the four correct
> switches. As far as I can tell, there's *no* way to solve this except to
> write down *all 32* switch patterns, and *then* start combining them by
> brute force. I wound up writing a program to solve it for me, which at least
> saved the brute-force step.)

I also had trouble with the one in Sanselard - the snow
world with all the windmill things - where you had to
figure out how to match up the weird sounds. I'd lose track
of where I was because everything looked so similar.
I couldn't figure out how to map the place.

Thanks for the review.
Anonymous
February 5, 2005 7:28:59 PM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Here, Jenny100 <nospam@nospam.com> wrote:
>
> > (The worst puzzle in _Sentinel_ is a blinky-lights affair in which
> > thirty-two switches control twenty lights. You have to pick the four correct
> > switches. As far as I can tell, there's *no* way to solve this except to
> > write down *all 32* switch patterns, and *then* start combining them by
> > brute force. I wound up writing a program to solve it for me, which at least
> > saved the brute-force step.)
>
> I also had trouble with the one in Sanselard - the snow
> world with all the windmill things - where you had to
> figure out how to match up the weird sounds. I'd lose track
> of where I was because everything looked so similar.
> I couldn't figure out how to map the place.

I never mapped it, but I was able to get to any spot by going
clockwise around the outside edge until I was there. Except for the
two-level center island, but since that was central, I could stand
anywhere and look around for an upper or lower bridge leading inwards.

I frequently lost track of where I was, but I just circled around
the edge until I found the teleporter.

--Z

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
I'm still thinking about what to put in this space.
Related resources
Anonymous
February 5, 2005 8:59:11 PM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Andrew Plotkin <erkyrath@eblong.com> wrote in
news:cu1j0l$8oj$2@reader2.panix.com:

> (The worst puzzle in _Sentinel_ is a blinky-lights affair in which
> thirty-two switches control twenty lights. You have to pick the four
> correct switches. As far as I can tell, there's *no* way to solve this
> except to write down *all 32* switch patterns, and *then* start
> combining them by brute force. I wound up writing a program to solve
> it for me, which at least saved the brute-force step.)

It's nice to know that I wasn't the only one to solve this puzzle by
writing a program. I only entered the data and performed the brute-force
search on the left half of the pattern pairs; it was enough to narrow the
solution down to two possibilities.

--
Murray Peterson
Email: murray.spamtrap1@shaw.ca
URL: http://members.shaw.ca/murraypeterson/
Anonymous
February 6, 2005 3:54:03 AM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Andrew Plotkin wrote:

> (The worst puzzle in _Sentinel_ is a blinky-lights affair in which
> thirty-two switches control twenty lights. You have to pick the four correct
> switches. As far as I can tell, there's *no* way to solve this except to
> write down *all 32* switch patterns, and *then* start combining them by
> brute force. I wound up writing a program to solve it for me, which at least
> saved the brute-force step.)

Well there are some methods for pruning the decision tree so it is not
completely brute force. Still you do need to determine the patterns for
each possible combination which is too close to brute force for me.

Here is some pruning logic. If you add the lights on each bridge
abutment, 4 is overload and 3 is a working bridge. So, obviously you
need each abutment to have three lights (sums to blue) on. That being
said, work backwards.

For any three switch panels, there must be at least 2 lights on for
every position. For any two panels, there must be at least 1 light on
for every position. That means you can eliminate some combinations with
just two switch panels. That will rapidly prune down the total number of
patterns to test.

Rather than write a program to test results, I entered the data into a
spread sheet. The totals were calculated using IF, INDEX, and addition
equations.
Anonymous
February 6, 2005 7:56:36 AM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Here, Robert Gault <robert.gault@worldnet.att.net> wrote:
> Andrew Plotkin wrote:
>
> > (The worst puzzle in _Sentinel_ is a blinky-lights affair in which
> > thirty-two switches control twenty lights. You have to pick the four correct
> > switches. As far as I can tell, there's *no* way to solve this except to
> > write down *all 32* switch patterns, and *then* start combining them by
> > brute force. I wound up writing a program to solve it for me, which at least
> > saved the brute-force step.)
>
> Well there are some methods for pruning the decision tree so it is not
> completely brute force. Still you do need to determine the patterns for
> each possible combination which is too close to brute force for me.
>
> Here is some pruning logic. If you add the lights on each bridge
> abutment, 4 is overload and 3 is a working bridge. So, obviously you
> need each abutment to have three lights (sums to blue) on. That being
> said, work backwards.

It's easier to restate it this way: every light has to be *off* from
exactly one switch. If you consider a switch to touch five lights
("off"), then it's a jigsaw puzzle: find the four patterns that fit
together with no overlaps or gaps.

But you still have to copy down all 32 patterns to solve it. And then
start looking at combinations.

--Z

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
I'm still thinking about what to put in this space.
Anonymous
February 6, 2005 7:21:23 PM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Andrew Plotkin wrote:

> But you still have to copy down all 32 patterns to solve it. And then
> start looking at combinations.
>

True. But to provide more detail on my approach, start with the left two
panels and try all switch combinations (64), eliminate all results that
have any light off at any position, test the remaining (11) with all
combinations of the next panel to the right (88), discard all results
with any red light (sums to 1), and finally test remaining (4) with the
right most panel for the correct result (takes from 4 to 32 tries).
Required tests could be as little as 156 out a total of 4096.

If your jigsaw approach is faster, please explain in detail.
Anonymous
February 6, 2005 8:11:08 PM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Here, Robert Gault <robert.gault@worldnet.att.net> wrote:
> Andrew Plotkin wrote:
>
> > But you still have to copy down all 32 patterns to solve it. And then
> > start looking at combinations.
> >
>
> True. But to provide more detail on my approach, start with the left two
> panels and try all switch combinations (64), eliminate all results that
> have any light off at any position, test the remaining (11) with all
> combinations of the next panel to the right (88), discard all results
> with any red light (sums to 1), and finally test remaining (4) with the
> right most panel for the correct result (takes from 4 to 32 tries).
> Required tests could be as little as 156 out a total of 4096.
>
> If your jigsaw approach is faster, please explain in detail.

It's not easily comparable, because I didn't try any of the
combinations with the actual switches. I flipped each one by itself,
and wrote down the five (off) lights (as letters A to T). That took
a while.

Then I tried paring down the possibilities by looking at lights,
rather than switches. How many switches hit A? (Five.) If the first of
those is right, how many switches hit B without overlapping the first?
(Two.) What about C? (No possibilities left for either of the two B
options. Back up to stage 1...)

Then I realized this would still be a lot of trials, even though I was
doing them on paper, which is quicker than on the screen. I don't know
if it would be more or less than your 156, but I wasn't up for it.
That's when I wrote a simple Python script. I didn't try to make the
script efficient -- it ran through all 8^4 combinations -- because
there was no reason to make it efficient.

--Z

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
I'm still thinking about what to put in this space.
Anonymous
February 6, 2005 8:30:31 PM

Archived from groups: comp.sys.ibm.pc.games.adventure (More info?)

Here, Robert Gault <robert.gault@worldnet.att.net> wrote:
> Andrew Plotkin wrote:
>
> > But you still have to copy down all 32 patterns to solve it. And then
> > start looking at combinations.
> >
>
> True. But to provide more detail on my approach, start with the left two
> panels and try all switch combinations (64), eliminate all results that
> have any light off at any position, test the remaining (11) with all
> combinations of the next panel to the right (88), discard all results
> with any red light (sums to 1), and finally test remaining (4) with the
> right most panel for the correct result (takes from 4 to 32 tries).
> Required tests could be as little as 156 out a total of 4096.

Let me add, I like your approach. More testing but less writing stuff
down. I probably would have done that if I hadn't started thinking
about the problem from the opposite end.

--Z

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
I'm still thinking about what to put in this space.
!