Interactive Fiction Online?

14 views
Skip to first unread message

Gary Leighton

unread,
Jun 13, 2005, 5:46:03 AM6/13/05
to
It seems that most (all?) of the IF developed by this community is
offline material.

I have several related questions...

Are there any IF works that can be played on-line? Are they popular
among the community? (I'm not talking about MUDS, I'm more interested
in single player IF at the moment).

I have been looking at the new, online version of "The Hitchhiker's
Guide to the Galaxy" on the BBC's website. Did anyone enjoy playing
this? Opinions?

Are there any development tools available for writing online IF? Is
there a version of Inform/TADS etc that can be used online? What kind
of technologies are best for developing online IF?

I noticed Emily mention a PHP version of Zork and that she had been
looking at the logs. This sounds like it would be a very useful feature
for an IF system. My opinion is that this might be the strongest reason
why someone might want to develop online IF. i.e. it would be very
useful to get direct feedback on your game by analysing other people's
sessions. Any opinions on this?

It would be cool if there were a web-site somewhere that allowed users
to play any of the games from the IF-archive online, without the need
for downloading a seperate interpreter or the game data files. Is there
something like this already?

Assuming that a good online interpreter exists, would you prefer to
play online or offline. Why?

Appologies if this has all been discussed before now.

Gary

Vivienne Dunstan

unread,
Jun 13, 2005, 5:58:02 AM6/13/05
to
Gary Leighton <leight...@hotmail.com> wrote:

> I have been looking at the new, online version of "The Hitchhiker's
> Guide to the Galaxy" on the BBC's website. Did anyone enjoy playing
> this? Opinions?

I enjoyed it a lot (and I played the original version too many years
ago), but found it inconvenient to play since I had to be online. I'm
not always online when at a computer and prefer games that I can play
offline. I wouldn't have minded downloading a copy and playing it
offline that way, but that sort of defeats the whole online thing!

And that leads to ...

> Assuming that a good online interpreter exists, would you prefer to
> play online or offline. Why?

Offline because it's more convenient for me. I have broadband now but
offline is still generally more convenient. I also like the facilities
that offline game interpreters provide, like good reliable saving
facilities, abbreviated commands like x for examine (typing words like
that in full that was a pain with the online HHGTTG game!), etc.

Being offline is pretty important for me because it means I can play the
game any time, any place. I even play IF on a palmtop (with a keyboard)
on long boring train journeys. Online only wouldn't be much use there.
Admittedly for some people they'd argue that any time, any place is well
suited to online playing, but not me. Offline is default, online not :)

I have played a lot of MUDs in the past and would definitely be happy to
play those online. But they're not very convenient when I'm not online!

Hope this helps.

Viv Dunstan

JohnnyMrNinja

unread,
Jun 13, 2005, 6:23:51 AM6/13/05
to

Gary Leighton wrote:

> It would be cool if there were a web-site somewhere that allowed users
> to play any of the games from the IF-archive online, without the need
> for downloading a seperate interpreter or the game data files. Is there
> something like this already?

There are *many* sites that have web-based Z-terps, the best (terp) by
far is at Ifiction.org. The player's end of the game is entirely HTML,
so "Back" is undo, and to save you just save the webpage. Other terps
(that I know of) download the file to your computer to play.

> Assuming that a good online interpreter exists, would you prefer to
> play online or offline. Why?

Personally, offline. Besides the fact that I have dial-up, the biggest
plus to online play (besides multiplayer or dynamic content, neither of
which you're talking about) is that you can play anywhere, but there's
no way to access saves from another computer (again, that I know of).
1) I save often, 2) I rarely play computer games away from home.

Maybe this is absurd, but what if an online terp saved your game
through email? And maybe you could email it back? This probably
wouldn't be easy with a QUETZAL save, but what about IFiction's
HTML-based terp?

Can I ask what the appeal is to a non-newcomer to play IF games online?

Arnel Legaspi

unread,
Jun 13, 2005, 6:43:49 AM6/13/05
to
JohnnyMrNinja wrote:
> > Assuming that a good online interpreter exists, would you prefer to
> > play online or offline. Why?
>
> Personally, offline. Besides the fact that I have dial-up, the biggest
> plus to online play (besides multiplayer or dynamic content, neither of
> which you're talking about) is that you can play anywhere, but there's
> no way to access saves from another computer (again, that I know of).
> 1) I save often, 2) I rarely play computer games away from home.

Ditto.

Current online terps don't have the ability to save/restore. I think
they're there mostly to display the finished product, not to have the
game be played online.

I'm also on dial-up, and since I consider IF more of a hobby, I'm not
really willing to spend Internet time for that. (Not sure if I'd react
the same if I were on broadband, though :))

--Arnel

Jan Thorsby

unread,
Jun 13, 2005, 6:44:43 AM6/13/05
to

"Gary Leighton" <leight...@hotmail.com> skrev i melding
news:1118655963.5...@z14g2000cwz.googlegroups.com...

> It seems that most (all?) of the IF developed by this community is
> offline material.
>
> I have several related questions...
>
> Are there any IF works that can be played on-line?

The IF wiki mentions a few pages:
http://www.ifwiki.org/index.php/Websites_for_downloading_or_playing_IF
If anybody know of any other pages where one can download or play IF, add
them to the wiki.


Jason Lai

unread,
Jun 13, 2005, 7:11:29 AM6/13/05
to
JohnnyMrNinja wrote:
> There are *many* sites that have web-based Z-terps, the best (terp) by
> far is at Ifiction.org. The player's end of the game is entirely HTML,
> so "Back" is undo, and to save you just save the webpage. Other terps
> (that I know of) download the file to your computer to play.

There's also some Java applet interpreters like ZPLet, but they suffer
from the same lack-of-saving issue.

>>Assuming that a good online interpreter exists, would you prefer to
>>play online or offline. Why?
>
> Personally, offline. Besides the fact that I have dial-up, the biggest
> plus to online play (besides multiplayer or dynamic content, neither of
> which you're talking about) is that you can play anywhere, but there's
> no way to access saves from another computer (again, that I know of).
> 1) I save often, 2) I rarely play computer games away from home.

I'm not aware of any online web interpreters that let you do this, but
there's no fundamental limitation to saving on the server. Disk space is
cheap (at least for dedicated machines) and it'd save a several kb of
roundtrip bandwidth.

Aside from web sites, there's some nifty instant messaging bots (on AIM
primarily) that let you play Z-code games, and save. One upshot is that
since you're in a continuous session, it doesn't have to keep
saving/loading the game every time, and it doesn't load a new page. This
could be done on a website too but it's a lot trickier. Part of the
problem being that the terps aren't designed for stateless play.

>
> Can I ask what the appeal is to a non-newcomer to play IF games online?
>

I think it could be useful for non-standard IF interpreters. It's
troublesome to go get Hugo or ADRIFT terps to play one of the small
number of games for those platforms. Especially if you write your own IF
app without a popular interpreter, or for those ancient ones written in DOS.

Perhaps something like http://www.rinkworks.com/adventure/

It does also open up the possibility of dynamic content, as you said, or
perhaps something slightly outside of IF like the defunct Majestic -- a
mystery thriller that would send you IMs, emails, pre-recorded phone
calls, etc. Maybe not that extreme, but it could be a game where you
look for clues around the website and use that knowledge in the game.

You could also create a more sophisticated form of HTML TADS and take
full advantage of all the powerful layout and interactivity stuff you
can do in browsers. And then to do all this fancy stuff, you don't need
to write a complicated cross-platform rendering engine because everyone
already has a browser. They don't even have to bother to download anything.

- Jason

Michael Coyne

unread,
Jun 13, 2005, 5:16:12 PM6/13/05
to
On Mon, 13 Jun 2005 02:46:03 -0700, Gary Leighton said to the parser:

> It seems that most (all?) of the IF developed by this community is offline
> material.
>
> I have several related questions...
>
> Are there any IF works that can be played on-line? Are they popular among
> the community? (I'm not talking about MUDS, I'm more interested in single
> player IF at the moment).

I've got a playable version of Risorgimento Represso here:
http://www.mts.net/~coyne/zplet/risorg.html

It suffers from the usual zplet problems with saving and restoring.

The intention is mainly to provide people with a chance to sample it
before going through the hassle of downloading and installing an
interpreter--yes, not a huge hassle, but we're talking about friends and
family who have no idea what IF really is... "You mean those old games,
where you just type?"


Michael

ems...@mindspring.com

unread,
Jun 13, 2005, 6:20:28 PM6/13/05
to
Gary Leighton wrote:
> I noticed Emily mention a PHP version of Zork and that she had been
> looking at the logs.

I think it has been taken down now, but for a while there was an online
Zork which also featured the ability, not to view all the logs, but to
look at a random sampling of commands other players had recently
issued. I played with it for a while just to see whether any patterns
emerged, especially any commands that seemed sensible from a newbie
perspective but wouldn't be accepted by the average IF parser. So that
was interesting, though obviously one could do a lot more if one had
access to a full log file, and some way of sifting for the commands
rejected by the interpreter.

I don't think that the people (person?) behind PHP Zork was doing any
such thing, though. I've occasionally thought about trying more or less
what you describe*, but it would take me a long time to teach myself
what I'd need to know in order to build it, so other projects keep
taking priority...

* Actually, the full form of my idea was to try making what would
basically be a chatbot interface to an IF interpreter. It would let
people type, log the commands entered, use various filters to convert
invalid forms of command into IFese, pass the fixed commands to the
interpreter, and report the results to the player. Gradual refinements
(so the theory went) would make it receptive to a much broader range of
input than the average IF parser, by first applying a sort of AIML-like
methodology for extracting the real content from the dross.

This may be completely cracked, but it struck me as an interesting idea
to try. I suspect that the invalid commands would break down into
several categories:

1) commands which would be fine IFese if only you took out a bunch of
conversational nothing-words, or adverbial phrases that don't
correspond to anything in the world model. PHP Zork logs had a lot of
things like >KNOCK ON THE DOOR WITH MY FIST, etc., where "with my fist"
is the sort of addendum your average IF parser will choke on. It's easy
to throw this kind of information away entirely: what the player has
typed is a superset of what an experienced IF player would type. A
chatbot scripting language like AIML is good at the task of finding
keywords and trashing any intervening verbiage.

2) commands which make perfect sense with respect to the world model,
but aren't currently handled by most IF systems: >GO BACK, say, or >GO
TO THE KITCHEN. My hypothetical translator-bot wouldn't be much help
with these, but if we discovered what sorts of rational commands
newbies often type that *could* be supported, that suggests a direction
for feature improvements in individual games or in IF languages as a
whole. (Obviously, there are some games that do handle those specific
commands. In general, I get the impression that the whole "but why do I
have to use a flipping compass??" complaint could be reduced or avoided
if more games honored >GO TO THE ORRERY when the orrery is a listed
exit. It's not *hard* -- there's no pathfinding involved here, and the
idea goes all the way back to Adventure. Anyway, survey might turn up
more of this sort of thing.)

3) commands which are too vague ever to be converted to IFese, though
they do represent a logical sequence of actions in the game world. I
think of these as sort of macros for entire command sequences: >GO TO
THE KITCHEN when the kitchen is five rooms away; >CHECK INTO THE HOTEL
when the game expects you to spend five turns interacting with the desk
clerk; >PREPARE EGGS BENEDICT when there's a whole recipe to follow.
The first is a special case because, in fact, with a pathfinding
mechanism it is possible to work out what the command set would be.
Usually this sort of thing is harder, and couldn't be dealt with by the
translator-bot *or* general language improvements. What you would need,
if you actually wanted to honor this kind of command, is customization
on a per-game basis, perhaps using a goal-finding mechanism of some
sort. Alternatively, it might be enough to recognize the most common
over-general commands that might be issued in your game, and provide
good hints or refusal messages to let the player know how to rephrase
the request. The typical puzzlefest probably doesn't want to enable
such shortcuts anyway.

4) commands which are essentially garbage from the game's point of view
-- things like >YOU ARE VERY ANNOYING or >PONDER TRANSCENDENT REALITY
or >TRANSLATE THAT LAST PARAGRAPH INTO ESPERANTO. Most of the time,
though, even the newbie player probably realizes that this stuff is
nonsense. Impossible to cope with the full range of what might be said
here, but I'm also not too worried about this category.

Poster

unread,
Jun 14, 2005, 12:03:07 AM6/14/05
to
I really don't see the use/fun/desire to play IF online, unless it was
part of a shared world construct. I mean, if you could leave notes for
other players, or affect the scenery to a degree, or co-operate with
other players to get points, and so on, then sure. But at that point,
you're getting awfully close to MUD territory and I'm not sure how far
you could take the concept without encountering terminal complexity.
What makes sense for an individual player exploring an old white house
may not make sense for a group of people exploring the house,
especially if twenty people have been there ahead of you and eaten the
food, taken the bottle and so forth. Yes, you could have respawning
items, but the whole mimesis is strangled to death by that particular
thing. In short, I just don't think that multiplayer IFs work that
well, and that's the chief reason to going online. So I'll prefer
downloading a few K or Mb of a game and playing offline, thanks.

~Poster

Gary Leighton

unread,
Jun 14, 2005, 12:12:43 PM6/14/05
to
Thanks Viv.

I seems clear that most people would prefer to play offline. To me it
seems that there would be big advantages for an author if his game is
played online (and session logs stored away for analysis) but maybe
it's not that great for the player.

I suppose a compromise could be for works to be offered online at
first, and an offline version created later when the author is happy
that he has received enough feedback from the online version. However,
this would require an authoring system that can work online and
offline.

Gary

Gary Leighton

unread,
Jun 14, 2005, 12:23:36 PM6/14/05
to
Does the Z-terp at Ifiction.org save session logs for the author to
view?

I'm not sure I see why an online game would have trouble saving the
game state on the client computer, assuming the user gives permission.

I'm not sure what the appeal would be for a non-newcomer to play
online. I was thinking more about advantages to the author.

Gary

Rockersuke Moroboshi

unread,
Jun 14, 2005, 4:18:45 PM6/14/05
to

"Gary Leighton" <leight...@hotmail.com> escribió en el mensaje
news:1118766216.6...@g49g2000cwa.googlegroups.com...

> I'm not sure I see why an online game would have trouble saving the
> game state on the client computer, assuming the user gives permission.

In fact, Hamlet, listed in the Ifwiki page with online games, claims to do
so using cookies. I haven´t checked if it really works, but I've bookmarked
it to play when I have the time... It looks fun!

http://www.robinjohnson.f9.co.uk/adventure/hamlet.html

Josh Lawrence

unread,
Jun 14, 2005, 4:25:26 PM6/14/05
to
As a side note, the Java-based Cycon Online Gaming (GOG) Engine was
meant to allow the development of IF-style games (with or without
graphics) that would be playable in any compliant browser without
having to download anything.

It looks like development on the engine stalled back in 2002, though:
http://cogengine.sourceforge.net/news.html

But you can still play the unfinished test game in low- and
high-bandwidth versions:
http://cogengine.sourceforge.net/CogEngine/online_games-low.html
http://cogengine.sourceforge.net/CogEngine/online_games-high.html

I remember the game running really slow and clunky for me when I first
stumbled upon it in, hm, probably 2002. Now it seems better, but still
has some lag. Also the game is (as the site admits) rather
under-implemented. Never tried out the development tools.

Josh

David Kinder

unread,
Jun 14, 2005, 5:58:29 PM6/14/05
to
Gary Leighton wrote:
> I suppose a compromise could be for works to be offered online at
> first, and an offline version created later when the author is happy
> that he has received enough feedback from the online version. However,
> this would require an authoring system that can work online and
> offline.

Well, all you need is an interpreter that works online. You could use
(say) Inform and modify Zplet to record the player sessions for the game,
for example.

David

Mike Rozak

unread,
Jun 15, 2005, 5:33:51 PM6/15/05
to
> Well, all you need is an interpreter that works online. You could use
> (say) Inform and modify Zplet to record the player sessions for the game,
> for example.

I am working on a IF authoring tool that lets authors create either online
or offline IF. It will also be graphical, with still 360-degree images (like
Myst III), audio, and use text-to-speech.

An online game is technically the same as an offline game, but needs to be
designed to take into account the fact that multiple players might be in the
same world.

--

Mike Rozak
http://www.mxac.com.au


Mike Snyder

unread,
Jun 15, 2005, 5:55:23 PM6/15/05
to
"Mike Rozak" <Mike...@bigpond.com> wrote in message
news:391se.17925$F7....@news-server.bigpond.net.au...

I have a friend who has been developing pretty much the same thing. Games
are browser-deployed, and can be played online or offline. He has been
working on it for about 5 years. I keep encouraging him to wrap it up and
make it available.

---- Mike.


Gary Leighton

unread,
Jun 17, 2005, 5:40:17 AM6/17/05
to

ems...@mindspring.com wrote:
[snip]


> I think it has been taken down now, but for a while there was an online
> Zork which also featured the ability, not to view all the logs, but to
> look at a random sampling of commands other players had recently
> issued. I played with it for a while just to see whether any patterns
> emerged, especially any commands that seemed sensible from a newbie
> perspective but wouldn't be accepted by the average IF parser. So that
> was interesting, though obviously one could do a lot more if one had
> access to a full log file, and some way of sifting for the commands
> rejected by the interpreter.

I'm suprised that authors here are not crying out for this kind of
feature. On the one hand, an author could examine the session logs
(i.e. transcipts) and use this information to help fix bugs or make
improvements to their works. I would also expect these resources to be
extremely valuable to developers who are trying to improve or create
new authoring systems. If the resulting transcripts could be made
public, I'm sure there would be a good audience who would enjoy reading
them too. All in all I would expect this to be very good for IF. (I
seem to remember you (Emily) discussing the idea of a transcript
repository before).

In my case, I'm writing a new parser, and I need some way to get
feedback for it. I plan to make an online version initially. I'm not
sure what technology to use yet but Java is my favourite so far.

> I don't think that the people (person?) behind PHP Zork was doing any
> such thing, though. I've occasionally thought about trying more or less
> what you describe*, but it would take me a long time to teach myself
> what I'd need to know in order to build it, so other projects keep
> taking priority...
>
> * Actually, the full form of my idea was to try making what would
> basically be a chatbot interface to an IF interpreter. It would let
> people type, log the commands entered, use various filters to convert
> invalid forms of command into IFese, pass the fixed commands to the
> interpreter, and report the results to the player. Gradual refinements
> (so the theory went) would make it receptive to a much broader range of
> input than the average IF parser, by first applying a sort of AIML-like
> methodology for extracting the real content from the dross.

Sounds like a great idea. Would you make this technology available to
other authors?

> This may be completely cracked, but it struck me as an interesting idea
> to try. I suspect that the invalid commands would break down into
> several categories:
>
> 1) commands which would be fine IFese if only you took out a bunch of
> conversational nothing-words, or adverbial phrases that don't
> correspond to anything in the world model. PHP Zork logs had a lot of
> things like >KNOCK ON THE DOOR WITH MY FIST, etc., where "with my fist"
> is the sort of addendum your average IF parser will choke on. It's easy
> to throw this kind of information away entirely: what the player has
> typed is a superset of what an experienced IF player would type. A
> chatbot scripting language like AIML is good at the task of finding
> keywords and trashing any intervening verbiage.

I'm in two minds about this idea (in relation to my own parser). The
danger is that you lead the player to beleive that "KNOCK ON THE DOOR
WITH MY FIST" is an acceptable input and has been completely understood
by the program. I think it would be better if your UI somehow reports
that (WITH MY FIST) has been ignored.

The approach I plan to take is to introduce something like the
Intellisence that Microsoft use in Word and Visual Studio. As the
player types in a command, the program will underline and words it
doesn't recognise, any grammar errors and semantic misunderstandings in
real time (with a descriptive error message beneath the input text).
The input will not be accepted without correction. In this way, I hope
the player will quickly learn the subset of English that is in use,
without lots of error messages spoling the main story text.

Even so, I am hoping to significantly extend the range of allowable
input and I'm hoping that the Intellisense is not called into action
too frequently.

> 2) commands which make perfect sense with respect to the world model,
> but aren't currently handled by most IF systems: >GO BACK, say, or >GO
> TO THE KITCHEN. My hypothetical translator-bot wouldn't be much help
> with these, but if we discovered what sorts of rational commands
> newbies often type that *could* be supported, that suggests a direction
> for feature improvements in individual games or in IF languages as a
> whole. (Obviously, there are some games that do handle those specific
> commands. In general, I get the impression that the whole "but why do I
> have to use a flipping compass??" complaint could be reduced or avoided
> if more games honored >GO TO THE ORRERY when the orrery is a listed
> exit. It's not *hard* -- there's no pathfinding involved here, and the
> idea goes all the way back to Adventure. Anyway, survey might turn up
> more of this sort of thing.)

I strongly agree that the compass directions have to go!

But this is a great reason to support the idea of getting feedback from
the player. There are probably lots of plausible verbs, commands and
synonyms that could easily be implemented in a parser that we simply
haven't guessed at.

> 3) commands which are too vague ever to be converted to IFese, though
> they do represent a logical sequence of actions in the game world. I
> think of these as sort of macros for entire command sequences: >GO TO
> THE KITCHEN when the kitchen is five rooms away; >CHECK INTO THE HOTEL
> when the game expects you to spend five turns interacting with the desk
> clerk; >PREPARE EGGS BENEDICT when there's a whole recipe to follow.
> The first is a special case because, in fact, with a pathfinding
> mechanism it is possible to work out what the command set would be.
> Usually this sort of thing is harder, and couldn't be dealt with by the
> translator-bot *or* general language improvements. What you would need,
> if you actually wanted to honor this kind of command, is customization
> on a per-game basis, perhaps using a goal-finding mechanism of some
> sort. Alternatively, it might be enough to recognize the most common
> over-general commands that might be issued in your game, and provide
> good hints or refusal messages to let the player know how to rephrase
> the request. The typical puzzlefest probably doesn't want to enable
> such shortcuts anyway.

: SOLVE ADVENTURE :-)

I hadn't considered this before. I suppose I have become trained as a
player to always enter micro-commands.

I think the GO TO <distant location> using a map searching function
like A* should be a standard feature in authoring systems nowadays.
It's not that hard to do.

Gary

Howard

unread,
Jun 17, 2005, 10:07:10 PM6/17/05
to
There's some fantastic IF you can play online at
http://www.malinche.net/demos/index.html

Have at it and have a good time.

Howard Sherman
Implementor
http://www.malinche.net


ems...@mindspring.com

unread,
Jun 19, 2005, 4:46:01 PM6/19/05
to
Gary Leighton wrote:

Re. transcript logging of online games:


> I'm suprised that authors here are not crying out for this kind of
> feature. On the one hand, an author could examine the session logs
> (i.e. transcipts) and use this information to help fix bugs or make
> improvements to their works. I would also expect these resources to be
> extremely valuable to developers who are trying to improve or create
> new authoring systems. If the resulting transcripts could be made
> public, I'm sure there would be a good audience who would enjoy reading
> them too. All in all I would expect this to be very good for IF. (I
> seem to remember you (Emily) discussing the idea of a transcript
> repository before).

Yes; no one else seemed interested, though. Oh well.

> > * Actually, the full form of my idea was to try making what would
> > basically be a chatbot interface to an IF interpreter. It would let
> > people type, log the commands entered, use various filters to convert
> > invalid forms of command into IFese, pass the fixed commands to the
> > interpreter, and report the results to the player. Gradual refinements
> > (so the theory went) would make it receptive to a much broader range of
> > input than the average IF parser, by first applying a sort of AIML-like
> > methodology for extracting the real content from the dross.
>
> Sounds like a great idea. Would you make this technology available to
> other authors?

If I ever came up with it to start with, which at this point does not
seem likely to happen in the very near future, sure.

> I'm in two minds about this idea (in relation to my own parser). The
> danger is that you lead the player to beleive that "KNOCK ON THE DOOR
> WITH MY FIST" is an acceptable input and has been completely understood
> by the program. I think it would be better if your UI somehow reports
> that (WITH MY FIST) has been ignored.

Right. (I think in order not to drive the player mad it might be wise
for the bot to show as feedback what the player's command had been
translated *into*, as well. Also, useful for debugging the thing.)

> The approach I plan to take is to introduce something like the
> Intellisence that Microsoft use in Word and Visual Studio. As the
> player types in a command, the program will underline and words it
> doesn't recognise, any grammar errors and semantic misunderstandings in
> real time (with a descriptive error message beneath the input text).
> The input will not be accepted without correction. In this way, I hope
> the player will quickly learn the subset of English that is in use,
> without lots of error messages spoling the main story text.

Hunh. Interesting idea -- sounds challenging to do systematically, but
I can see the advantages.

> I strongly agree that the compass directions have to go!

I didn't quite say *that*. Just that I think extending the range of
ways to express movement might make things less annoying to newbie
players.

> I think the GO TO <distant location> using a map searching function
> like A* should be a standard feature in authoring systems nowadays.

It's been discussed before, and there are games that do it --
"Dangerous Curves", if I'm remembering correctly, and some others. But
not every game is really set up to make that desirable: some games have
disconnected maps where you can't move from one place to another via
normal compass directions, and some are organized more by plot than by
geography. So I think of this functionality as more something that
belongs in an extension set than as something part of every game.

If I remember right, T3 does have some built-in pathfinding stuff. (T3
seems to have built-in *everything*.) I assume, though, that it's up to
the author whether to allow that particular use of the pathfinding
mechanism. In "Return to Ditch Day", the player wasn't allowed to go to
random distant locations, but there was a handy map that gave
directions dynamically, which I thought was a neat effect.

Mike Rozak

unread,
Jun 20, 2005, 5:35:44 PM6/20/05
to
>> All in all I would expect this to be very good for IF. (I
>> seem to remember you (Emily) discussing the idea of a transcript
>> repository before).
>
> Yes; no one else seemed interested, though. Oh well.

I could use some transcripts for generating text-to-speech training
sentences. The closer the training sentences are to what text-to-speech will
speak, the better the text-to-speech becomes (for those types of sentences).
Does anyone have such a library?


>> > Gradual refinements
>> > (so the theory went) would make it receptive to a much broader range of
>> > input than the average IF parser, by first applying a sort of AIML-like
>> > methodology for extracting the real content from the dross.
>>
>> Sounds like a great idea. Would you make this technology available to
>> other authors?
>
> If I ever came up with it to start with, which at this point does not
> seem likely to happen in the very near future, sure.

I have it working, but I haven't produce the logging tools, and haven't
tested it with a full compliment of IF commands. It seems to work for the 50
or so basic command I have already implimented.

ems...@mindspring.com

unread,
Jun 20, 2005, 5:58:44 PM6/20/05
to
Mike Rozak wrote:
> >> All in all I would expect this to be very good for IF. (I
> >> seem to remember you (Emily) discussing the idea of a transcript
> >> repository before).
> >
> > Yes; no one else seemed interested, though. Oh well.
>
> I could use some transcripts for generating text-to-speech training
> sentences. The closer the training sentences are to what text-to-speech will
> speak, the better the text-to-speech becomes (for those types of sentences).
> Does anyone have such a library?

I seem to recall Lucian Smith having a bunch of transcripts, but I
think they were all his own play-sessions, rather than diverse players'
input. I assume for this purpose you'd want to survey a variety of
different player types.

[Re. AIML pre-parsing thing]


> I have it working, but I haven't produce the logging tools, and haven't
> tested it with a full compliment of IF commands. It seems to work for the 50
> or so basic command I have already implimented.

Yeah, I remember exchanging some email with you about this. I'd be
curious to see how it's going, if you've got something at a point you'd
like to share somewhere along the line. (Whether it runs online or not.)

markm

unread,
Jun 20, 2005, 7:01:57 PM6/20/05
to
If it helps, I've got several transcripts from my Comp98 play sessions,
warts and all. Feel free to use them.

http://www.ministryofpeace.com/if/comp1998/transcripts/

Mike Rozak

unread,
Jun 20, 2005, 7:34:34 PM6/20/05
to
Thanks. I won't get to recording the sentences for awhile, but I added a
link to my favorites.

The reason that actual text from the game (as displayed by the game's
narrator) helps is because my TTS system learns how to pronounce words and
read sentences from sample recordings of a real person. So far, I have
recorded about 4000 sentences. They are all from Project Gutenberg books and
my own text, though.

Gutenberg books present two problems:

a) The sentences are long (10-40 words), so I don't have speech samples of
short sentences like "You put the lantern in the bag." Consequently, my
short-sentence prosody sounds worse than my long-sentence prosody... not
that my long-sentence prosody is that great.

b) Words that are common in IF are not necessarily common in books. The more
recordings I have of a word, the better it will sound. Words that have never
been recorded and which have no morphs (roots, prefixes, suffixes) in common
with other words won't sound as good. "brillig", "slithy", "toves",
"jabberwocky", etc. won't sound too good since I haven't read in "Alice in
wonderland", although the algorithms will probably piece together a
recording of "jabber" + "wok" + "key" to produce "jabberwocky".

--

Mike Rozak
http://www.mxac.com.au
"markm" <mark.m...@gmail.com> wrote in message
news:1119308517.0...@g49g2000cwa.googlegroups.com...

Mike Rozak

unread,
Jun 20, 2005, 10:11:32 PM6/20/05
to
> [Re. AIML pre-parsing thing]
>> I have it working, but I haven't produce the logging tools, and haven't
>> tested it with a full compliment of IF commands. It seems to work for the
>> 50
>> or so basic command I have already implimented.
>
> Yeah, I remember exchanging some email with you about this. I'd be
> curious to see how it's going, if you've got something at a point you'd
> like to share somewhere along the line. (Whether it runs online or not.)

I was hoping to have a pre-release up around now, but everytime I sit down
and think about a pre-release, I see more flaws that I'd rather fix before
puting out a pre-release (even though the pre-release will be hugely flawed
despite my best efforts). I just spent a lot of time speeding up my
rendering (although it's still fairly slow), and need to put in some
multi-player socialization features like voice-chat, and many more things.

Graphics, text-to-speech, and multiplayer capability add significant coding
tasks for me. They also provide some really nice touches.. a month ago I got
torches, candles, and other lights working in a photometrically accurate
way. A torch provides more light than a candle, as well as differently
colored light. (Not to mention the subtle fire-crackling sound that a torch
makes.) The combination of physically-correct lighting and sound effects
signficantly increases the immersion of caves and night scenes. (And you can
create a puzzle that requires a properly colored light.)

This immersion is promptly dashed by the yellow smiley (and frowning) faces
I'm using as a filler for NPC head shots, as well as my inability to create
a female voice from my male text-to-speech voice. I need to work on these
next.

Robin Johnson

unread,
Jun 21, 2005, 7:43:56 AM6/21/05
to
Rokersuke Moroboshi wrote:
> "Gary Leighton" <leightong...@hotmail.com> escribió en el mensaje
> news:1118766216.6...@g49g2000cwa.googlegroups.com...

> > I'm not sure I see why an online game would have trouble saving the
> > game state on the client computer, assuming the user gives permission.

> In fact, Hamlet, listed in the Ifwiki page with online games, claims to
> do so using cookies. I haven´t checked if it really works

It does now, I think. In IE and Firefox, at least.

I just gave it a spangly new scrolling interface too.

Robin Johnson

Reply all
Reply to author
Forward
0 new messages