Proposal: Don't make players SAVE their game

1 view
Skip to first unread message

Tom Spidell

unread,
Mar 28, 1999, 3:00:00 AM3/28/99
to
In explaining interactive fiction to people, one thing that seems to hang
people up is the need to remember to save their game manually, especially
when trying something dangerous. Of course, many Infocom-style games allows
you to resume after "dying", with reduced points, but I never found that
satisfactory, as you knew you then couldn't complete the game with all of
the points.

In comparing interactive fiction to regular fiction, it occurs to me that it
would be nice to be able to resume from any point in the past, just like you
can go back to any part of a book at any time. Of course, this needs to be
constrained somewhat, because as soon as the player starts going forward
again, he/she is creating a "parallel story" to the one they experienced the
first time. Managing this tree of stories would probably get out of hand.
However, if the rule is that once you start going forward your "parallel
stories" are lost, it might be understandable.

Anyway, this is just an idea I'm throwing out to the IF community - let me
play without worrying about SAVEing at all. If I get killed, let me somehow
scroll back, and resume from a point that I choose in the story. Yes, it's
a fair amount of state information to keep after every move, but this seems
doable with modern machines.

Does anyone know if this sort of thing has been attempted?

Tom Spidell

Pezman

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to
Tom Spidell wrote:
[snip]

> In comparing interactive fiction to regular fiction, it occurs to me that it
> would be nice to be able to resume from any point in the past, just like you
> can go back to any part of a book at any time.

It would be tricky, but I can think of three options here.
1) Have an automatically saved game after _every_ turn (uses lots of
disk space),

2) Removing the option for random number generation, and saving the list
of commands (so that one series of commands can result in only one game
position) (Makes places like the Bed quilt Room (?) in Adventure
impossible). This would necessitate stepping through all the previous
moves (possibly time consuming).

3) Having a set random number generation algorithm (if saved games are
to be transferable between interpreters), and saving the initial seed
with the current saved state; again saving the list of previous
commands. Again all previous moves would have to be reperformed by the
interpreter.


--
bits and peace

Pezman aka Nicholas Daley
<mailto:link_...@geocities.com>

Jonadab the Unsightly One

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to
"Tom Spidell" <spi...@excite.com> wrote:

> Anyway, this is just an idea I'm throwing out to the IF community - let me
> play without worrying about SAVEing at all. If I get killed, let me somehow
> scroll back, and resume from a point that I choose in the story. Yes, it's
> a fair amount of state information to keep after every move, but this seems
> doable with modern machines.
>
> Does anyone know if this sort of thing has been attempted?

Frotz's multi-undo capability is a step toward this. It seems to be
limited to a few moves, however (possibly depending on your machine).


I also have implemented a thing called an AutoSave feature for
Inform, which can be set by the user once to certain options
(automatically save every N moves, remind me every N moves that I
should save, automatically save whenever my score goes up, remind me
whenever my score goes up that I should save, ...) and then carries
them out indefinitely (unless the user resets the options or turns
the feature off). It isn't quite the same, but it does reduce the
pressure on the user to *remember* to save.

(I haven't released the AutoSave feature formally, but if anyone's
interested I can send you my code; or you can code it yourself. It
isn't all that hard...)


If zerospam.com bounces substitute bright.net

-- Jonadab the Unsightly One

ne...@norwich.edu

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to
In article <7dn312$9oh$1...@birch.prod.itd.earthlink.net>,

"Tom Spidell" <spi...@excite.com> wrote:
> In explaining interactive fiction to people, one thing that seems to hang
> people up is the need to remember to save their game manually, especially
> when trying something dangerous. Of course, many Infocom-style games allows
> you to resume after "dying", with reduced points, but I never found that
> satisfactory, as you knew you then couldn't complete the game with all of
> the points.

One more ancient text adventure trope that will hopefully be stamped out.
However, it isn't unique to text adventures. Almost all adventure games use
the technique you describe above (but see below).

> In comparing interactive fiction to regular fiction, it occurs to me that it
> would be nice to be able to resume from any point in the past, just like you

> can go back to any part of a book at any time. Of course, this needs to be
> constrained somewhat, because as soon as the player starts going forward
> again, he/she is creating a "parallel story" to the one they experienced the
> first time. Managing this tree of stories would probably get out of hand.
> However, if the rule is that once you start going forward your "parallel
> stories" are lost, it might be understandable.
>

> Anyway, this is just an idea I'm throwing out to the IF community - let me
> play without worrying about SAVEing at all. If I get killed, let me somehow
> scroll back, and resume from a point that I choose in the story. Yes, it's
> a fair amount of state information to keep after every move, but this seems
> doable with modern machines.
>
> Does anyone know if this sort of thing has been attempted?

Yes. "The Last Express" used an auto-save system in the shape of egg -- with a
clock on it. You could rewind the clock to any time you wanted and play from
that point. When you did that, you had 60 seconds to fool around before the
parallel future you came from was eradicated.

See Zarf's review for more information:

http://www.eblong.com/zarf/gamerev/lexpress.html

I liked the system. It worked well. They also included 3 eggs, so you could
have up to 3 totally different and mutually exclusive parallel universes.

Neil Cerutti
ne...@norwich.edu

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own

Andrew Plotkin

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to
ne...@norwich.edu wrote:
> In article <7dn312$9oh$1...@birch.prod.itd.earthlink.net>,
> "Tom Spidell" <spi...@excite.com> wrote:
> > Of course, many Infocom-style games allows
> > you to resume after "dying", with reduced points, but I never found that
> > satisfactory, as you knew you then couldn't complete the game with all of
> > the points.

> One more ancient text adventure trope that will hopefully be stamped out.

I think it pretty much has been stamped out, or rather allowed to expire.
When was the last time you say it?

> > [no need for saving]


> > Does anyone know if this sort of thing has been attempted?

> Yes. "The Last Express" used an auto-save system in the shape of egg -- with a
> clock on it. You could rewind the clock to any time you wanted and play from
> that point. When you did that, you had 60 seconds to fool around before the
> parallel future you came from was eradicated.

> See Zarf's review for more information:

> http://www.eblong.com/zarf/gamerev/lexpress.html

> I liked the system. It worked well. They also included 3 eggs, so you could
> have up to 3 totally different and mutually exclusive parallel universes.

But, as I wrote in the review, the system has other limitations. You
*can't* do everything that's possible with a manual save system. You can't
keep hold of several branched states of the *same* universe-tree.

--Z

--

"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the
borogoves..."

ne...@norwich.edu

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to
In article <erkyrathF...@netcom.com>,

erky...@netcom.com (Andrew Plotkin) wrote:
> ne...@norwich.edu wrote:
> > In article <7dn312$9oh$1...@birch.prod.itd.earthlink.net>,
> > "Tom Spidell" <spi...@excite.com> wrote:
> > > Of course, many Infocom-style games allows
> > > you to resume after "dying", with reduced points, but I never found that
> > > satisfactory, as you knew you then couldn't complete the game with all of
> > > the points.
>
> > One more ancient text adventure trope that will hopefully be stamped out.
>
> I think it pretty much has been stamped out, or rather allowed to expire.
> When was the last time you say it?

I placed the above comment carelessly. I meant it in response to the idea that
newbie's to IF must learn to save their game effectivly. I didn't intend to
comment on the practice of resurrecting dead characters and inflicting a minor
penalty.

> > > [no need for saving]
> > > Does anyone know if this sort of thing has been attempted?
>
> > Yes. "The Last Express" used an auto-save system in the shape of egg --
with a
> > clock on it. You could rewind the clock to any time you wanted and play from
> > that point. When you did that, you had 60 seconds to fool around before the
> > parallel future you came from was eradicated.
>
> > See Zarf's review for more information:
>
> > http://www.eblong.com/zarf/gamerev/lexpress.html
>
> > I liked the system. It worked well. They also included 3 eggs, so you could
> > have up to 3 totally different and mutually exclusive parallel universes.
>
> But, as I wrote in the review, the system has other limitations. You
> *can't* do everything that's possible with a manual save system. You can't
> keep hold of several branched states of the *same* universe-tree.

Thanks. I remembered you had some interesting comments about it, but forgot
what the comments were.

Jacob Munkhammar

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to
In article <36FF1896.9E3109A6@-.->, Pezman <-@-.-> wrote:

> Tom Spidell wrote:
> [snip]


> > In comparing interactive fiction to regular fiction, it occurs to me that it
> > would be nice to be able to resume from any point in the past, just like you
> > can go back to any part of a book at any time.
>

> It would be tricky, but I can think of three options here.
> 1) Have an automatically saved game after _every_ turn (uses lots of
> disk space),
>
> 2) Removing the option for random number generation, and saving the list
> of commands (so that one series of commands can result in only one game
> position) (Makes places like the Bed quilt Room (?) in Adventure
> impossible). This would necessitate stepping through all the previous
> moves (possibly time consuming).
>
> 3) Having a set random number generation algorithm (if saved games are
> to be transferable between interpreters), and saving the initial seed
> with the current saved state; again saving the list of previous
> commands. Again all previous moves would have to be reperformed by the
> interpreter.

4) Storing the *change* to the world state for every turn. This requires
some kind of internal "script" that tells *changes*, like "object
this-or-that is moved from location X to location Y", "variable something
changes from A to B". Note the form, allowing a trace both forward and
backwards.
This would allow for infinite multiple undos, as well as re-dos.
(Techninacally, a tree of these changes would be possible, but is that
really interesting?)
If, then, normal "saves" are stored in this chain of changes as
"snapshots", one could jump to a snapshot of choice and redo or undo from
there.
Difficult thing, though, is the user interface; how is the result of the
un- and re-dos to be reported, when changes to the state is traced, not
actual events?

/Jacob

--
Hemma hos: http://www.stud.ntnu.no/~jacob/
Textäventyr: http://www.stud.ntnu.no/~jacob/SAK/
Bugatti: http://www.stud.ntnu.no/~jacob/BILsidor/Bugatti/
Maskiner: PowerMac 4400, Mac LC, Mac Plus, Sinclair Spectrum

Bastiaan Zapf

unread,
Mar 29, 1999, 3:00:00 AM3/29/99
to

Tom Spidell schrieb in Nachricht
<7dn312$9oh$1...@birch.prod.itd.earthlink.net>...

>Does anyone know if this sort of thing has been attempted?
Well, the Final Fantasy RPG series has "save points" - you can save only
standing on one or when you wandered on the world map. Save points were only
where you had to take an important decision or where risks were waiting.

Maybe a similar thing would be good for IF.

"You open the door. You step into a room with three doors in the north. The
leftmost door looks old and worn, while the other two look better.

!return point 7!

Which way do you want to take?"

Or something like that.

So you only would have to save the game states when you pass such a point.

Basti

Knight37

unread,
Mar 30, 1999, 3:00:00 AM3/30/99
to

Bastiaan Zapf <B.Z...@t-online.de> wrote in message

>Maybe a similar thing would be good for IF.
>
>"You open the door. You step into a room with three doors in the north. The
>leftmost door looks old and worn, while the other two look better.
>
>!return point 7!
>
>Which way do you want to take?"
>
>Or something like that.
>
>So you only would have to save the game states when you pass such a point.

This smells fishily like "save points" found in console games.
I hate those. As long as an author still allowed the conventional
"save anywhere" ability as is practiced currently in IF, I would
not mind having these "save points" as an addition, for times
when I forgot to save myself. But having ONLY "save points" will
truely suck.

knight37

Iain Merrick

unread,
Mar 30, 1999, 3:00:00 AM3/30/99
to
Anson Turner wrote:

[...]
> Personally, I'm hoping that _Photopia_ will usher in a golden age when no
> one will ever need more than one saved state.

I love "Photopia", but I don't think it's the best example of a
no-saves-required game, since it's linear and almost puzzle-free. It's
possible to do the same thing within the framework of sprawling
non-linear puzzle-based game, too: Mike Roberts' _Perdition's Flames_ is
a good example.

And there's still room for a bit of cruelty, I'd have thought; one thing
I like about games like _Jigsaw_ is the way you can admire your nice
collection of saved game files after you finish.

> Points, of course, are completely passé.

Nothing is completely passe. Try "Small World" or "Zero Sum Game".

Uh, I do actually agree with you, really. But let's not be too quick to
throw out unpopular game features: with a novel twist, even maze puzzles
can still work -- as "Photopia" demonstrates.

--
Iain Merrick

Christopher Nebel

unread,
Mar 30, 1999, 3:00:00 AM3/30/99
to
In article <7dn312$9oh$1...@birch.prod.itd.earthlink.net>, "Tom Spidell"
<spi...@excite.com> wrote:

>In explaining interactive fiction to people, one thing that seems to hang
>people up is the need to remember to save their game manually, especially

>when trying something dangerous. Of course, many Infocom-style games allows


>you to resume after "dying", with reduced points, but I never found that
>satisfactory, as you knew you then couldn't complete the game with all of
>the points.
>

>In comparing interactive fiction to regular fiction, it occurs to me that it
>would be nice to be able to resume from any point in the past, just like you

>can go back to any part of a book at any time. Of course, this needs to be
>constrained somewhat, because as soon as the player starts going forward
>again, he/she is creating a "parallel story" to the one they experienced the
>first time. Managing this tree of stories would probably get out of hand.
>However, if the rule is that once you start going forward your "parallel
>stories" are lost, it might be understandable.

Infinite undo/redo would do the trick. Incidentally, the way applications
(e.g. Photoshop) handle it is that you can undo and redo along your path
as much as you want, but as soon as you do anything new, everything that's
in the "future" is wiped out.

TADS has multiple undo, and it works great as long as the thing you want
undone is only four or five turns in the past, which is usually the case
for things that get you killed. More than that, and it gets very
confusing to figure out exactly where it is you are along your history,
since you don't get any feedback aside from your location changing.
(Also, there's no redo!)

If you want to provide easy re-reading, then that's a moderately tough
problem. There was a discussion about it here perhaps a couple of months
ago. Depending on how "branchy" the game is, there may be cheap
solutions, such as providing several saved games at key points in the
game. (Consider: say there are four tasks you have to accomplish to
finish the game. If they have to be done in order, then there are only
four key points. If they can be done in any order, then there are
fifteen!)

If all you want is to keep people from going "^$%^@! I'm hosed, and I
never saved my game! Now I have to replay _everything_!", then that's
much easier. You can just throw up a prompt at points suggesting that now
would be a good time to save, and would the player like to do so? The
Legend Lives! and Firebird both use this; the code is in WorldClass (see
warnsave).


-Christopher Nebel

Tom Spidell

unread,
Mar 30, 1999, 3:00:00 AM3/30/99
to
Jacob Munkhammar wrote in message ...

>In article <36FF1896.9E3109A6@-.->, Pezman <-@-.-> wrote:
>
>> Tom Spidell wrote:
>> [snip]
>> > In comparing interactive fiction to regular fiction, it occurs to me
that it
>> > would be nice to be able to resume from any point in the past, just
like you
>> > can go back to any part of a book at any time.

>> [snip]


>
>4) Storing the *change* to the world state for every turn. This requires
>some kind of internal "script" that tells *changes*, like "object
>this-or-that is moved from location X to location Y", "variable something
>changes from A to B". Note the form, allowing a trace both forward and
>backwards.
>This would allow for infinite multiple undos, as well as re-dos.
>(Techninacally, a tree of these changes would be possible, but is that
>really interesting?)
>If, then, normal "saves" are stored in this chain of changes as
>"snapshots", one could jump to a snapshot of choice and redo or undo from
>there.
>Difficult thing, though, is the user interface; how is the result of the
>un- and re-dos to be reported, when changes to the state is traced, not
>actual events?

Deltas are certainly the most efficient way to store the states, but
trickier to implement.

Regarding the UI, even if the "no SAVEs needed" is implemented using deltas,
as you describe, I don't think this needs to be exposed to the user. No
matter how it's implemented, I envision the UI as allowing me to scroll back
through all of my moves, showing me at each step my location, a description
of my location (LOOK), my points, and my inventory (and perhaps my previous
command). I'd expect the UI to update these things in place, so you could
quickly scan and scroll back to an appropriate point to resume a game.

Tom Spidell

Jacob Munkhammar

unread,
Mar 31, 1999, 3:00:00 AM3/31/99
to
In article <7dscis$je7$1...@ash.prod.itd.earthlink.net>, "Tom Spidell"
<spi...@excite.com> wrote:

> Deltas are certainly the most efficient way to store the states, but
> trickier to implement.
>
> Regarding the UI, even if the "no SAVEs needed" is implemented using deltas,
> as you describe, I don't think this needs to be exposed to the user. No
> matter how it's implemented, I envision the UI as allowing me to scroll back
> through all of my moves, showing me at each step my location, a description
> of my location (LOOK), my points, and my inventory (and perhaps my previous
> command). I'd expect the UI to update these things in place, so you could
> quickly scan and scroll back to an appropriate point to resume a game.
>

I realized this a few hours after writing my post. This then would, of
course, mean that all text output must also be saved in parallell to the
changes, or deltas. Which, of course, is no problem - except perhaps on
small systems, like palmtops.

Iain Merrick

unread,
Mar 31, 1999, 3:00:00 AM3/31/99
to
Anson Turner wrote:

> * Minor _Photopia_ spoiler(s) ahead *

Extremely minor spoilers. But have some space anyway.


> In article <3700C8...@cs.york.ac.uk>, Iain Merrick <i...@cs.york.ac.uk> wrote:


>
> :Anson Turner wrote:
> :
> :[...]
> :> Personally, I'm hoping that _Photopia_ will usher in a golden age when no
> :> one will ever need more than one saved state.
> :
> :I love "Photopia", but I don't think it's the best example of a
> :no-saves-required game, since it's linear and almost puzzle-free. It's
> :possible to do the same thing within the framework of sprawling
> :non-linear puzzle-based game, too: Mike Roberts' _Perdition's Flames_ is
> :a good example.
>

> I don't quite understand this. Granted, no-saves-required does not demand
> linearity or puzzlelessness, but if you love _Photopia_, why do you think
> a non-linear, bepuzzled game would necessarily be a better example?

Well, if your game has no plot branches and no hard puzzles, you
practically get the no-saves-required property for free. You might even
get it accidentally.

Whereas giving a non-linear, bepuzzled [1] the no-saves-required
property is pretty difficult; this constitutes a major addition to the
game. It seems to me that this sort of game is a better example, since
you can easily imagine what the game would be like _without_ the
no-saves-required property.

(Hmmm, 'no-saves-required' should really be 'no-restore-required' or
'no-backtracking-required', but you get the idea.)

[1] Good adjective. I like it.

[...]
> As far as cruelty, I couldn't name a crueler work of IF than _Photopia_.

I was thinking of cruelty in the Zarfian sense, which basically covers
the same ground as 'no backtracking required'. The full definition will
be floating around dejanews somewhere, but basically in a cruel game you
can do something which makes the game unwinnable, and not realise you've
done so until much later on.

Moving from cruel to merciful, games might: give you fair warning about
dangerous actions; give you immediate feedback so that you can UNDO
them; or even prevent you from doing dangerous actions in the first
place.

--
Iain Merrick

David Glasser

unread,
Mar 31, 1999, 3:00:00 AM3/31/99
to
Anson Turner <anson@DELETE_THISpobox.com> wrote:

> Personally, I'm hoping that _Photopia_ will usher in a golden age when no
> one will ever need more than one saved state.

<photopia spoilers: if you haven't played it, do please>

Perhaps to complete the program you didn't need to have dozens of saves,
but are you trying to say that you completely understood the entire game
with only one scan through and no going back?

--
David Glasser: gla...@uscom.com | http://onramp.uscom.com/~glasser/
"It's good to explore the G.U.E. caves / It's good to explore the G.U.E.
caves / You can count all the leaves / You can KILL TROLL WITH SWORD /
You'll get stuck but you won't be bored"-Joe.Mason, rec.arts.int-fiction

Ryan N. Freebern

unread,
Apr 1, 1999, 3:00:00 AM4/1/99
to
anson@DELETE_THISpobox.com (Anson Turner) wrote:
>As far as cruelty, I couldn't name a crueler work of IF than _Photopia_.

Heh. The night I played Photopia, I went through it three times in a
row and then got on ifMUD and yelled at Adam because the game made me
very sad and angry. Then I knew how powerful it was.

Also, I had the hots for Alley.

-r


Jonadab the Unsightly One

unread,
Apr 1, 1999, 3:00:00 AM4/1/99
to
ja...@stud.ntnu.no (Jacob Munkhammar) wrote:

> I realized this a few hours after writing my post. This then would, of
> course, mean that all text output must also be saved in parallell to the
> changes, or deltas. Which, of course, is no problem - except perhaps on
> small systems, like palmtops.

Or for really large games.

Don't look at me like that. It's *theoretically* possible to have a
game that generates more text than can be saved on a 4GB hard
drive... I didn't say it was *likely*...

Reply all
Reply to author
Forward
0 new messages