Survey about why some people don't like graphical IF...
Mike Rozak
been thinking about building an IF authoring system on top of it. About a
month ago I posted a few questions about graphical IFs and conversation
models, receiving many good answers. One response pointed me to "Twisty
little passages", which just arrived in the mail. (It takes 2-3 weeks to get
a book to Darwin, Australia from Amazon.com). TLP provides a good discussion
of IF work over the past few decades. It didn't list too many graphical IF
works though, which got me thinking about another issue...
Many IF authors/readers do not like graphical IF. (Conversely, many do like
graphical IF). My scan through previous years' usenet posts, and trying to
read between the lines, has come up with the following real or perceived
problems with graphical IF:
(Issues that also arise when books are converted to movies.)
- The plot/story suffers because so much work goes into the graphics.
- Graphics often cannot compete with imagination, especially for surreal
situations like IF objects that cannot exist in reality ("get the sound
wave"). Furthermore, some concepts just cannot be visualized and require a
narrator to express.
- The graphics distract from the mental image created by the text. And/or
the poor-quality graphics distract from the text... if you can't do it right
then don't do it at all.
- Association with large(r) companies: Graphical IF is expensive to produce,
so it's in the realm of large companies. This has political ramifications
(anti-estabilishment and anti-Goliath) and quality implications (large
company = mass market = McIF).
- Graphics are a lot of work and don't really add anything to the
experience. Or, it's not worth the work.
- Graphical IFs don't work well for blind or vision-impaired users.
(Issues more specific to IF.)
- In general, graphical IF developers (such as for Myst) have decided to
toss out language. They do so because a narrator doesn't seem to fit, and
recording narration in umpteen languaes is expensive. Ultimately, the lack
of narration reduces the depth of what can be expressed in the IF. (If you
don't think Myst if IF then call it an adventure game.)
- Similarly, graphical IFs tend to rely on mouse-clicks, reducing the
complexity of interaction. Even those graphical IF books/games that use
commands have a very limited parser.
- A graphical IF is going to be limited to a few platforms (due to graphical
API incompatability), will take a long time to download, and requires a fast
CPU with lots of memory (no palm pilot support).
Is there anything else I should add to the list? Before I undertake such a
huge task, I'd like to know what elements could cause (or has caused)
amateur graphical IF to fail. Many of the issues cannot be overcome (such as
handheld support and vision-impaired users), but some can be worked around
or mitigated. (I like learning from other people's mistakes so I can avoid
them, enabling me to make my own new and more-innovative mistakes.)
--
Mike Rozak
www.mXac.com.au
Rexx Magnus
> Is there anything else I should add to the list? Before I undertake such a
> huge task, I'd like to know what elements could cause (or has caused)
> amateur graphical IF to fail.
*snip*
Probably the worst point of all, in that it is tremendously difficult to
cater graphically for the plethora of actions that the player could take -
this is why most games that have graphics do so only to supplement location
descriptions, rather than depict the action.
--
UO & AC Herbal - http://www.rexx.co.uk/herbal
To email me, visit the site.
JAN THORSBY
Mike Rozak <Mike...@bigpond.com> skrev i
meldingsnyheter:61uDb.54299$aT.2...@news-server.bigpond.net.au...
> - Similarly, graphical IFs tend to rely on mouse-clicks, reducing the
> complexity of interaction. Even those graphical IF books/games that use
> commands have a very limited parser.
This is the biggest problem for me. I liked the first Police Quest game, it
has a fairly good parser. But later games (Not all of them I think) use just
point and click, and are not much fun.
Another problem with graphical games is that it can be hard to tell what
objects one can interact with. In Goblins 3 you are supposed to interact
with a small rock on the ground. It looks just like an ordinary rock. I
think many people thought it was just scenery. You get the same problem in
IF, but to a much smaller degree, because every object is in the room
description. (or in other descriptions). There can also sometimes be a
problem to realise that two objects is not just one object. Like if the head
of a monster is another object than the body of the monster.
I guess a solution to this would be if there was a button on the interface
that showed you what things where objects. Like when you pushed it small
signs would appear on each object with the name of the object, then
dissapear again when youn cliked again. But quite possible many people would
think this took away the exploring aspect or something, and would not like
it.
Rexx Magnus
> I guess a solution to this would be if there was a button on the
> interface that showed you what things where objects. Like when you
> pushed it small signs would appear on each object with the name of the
> object, then dissapear again when youn cliked again. But quite possible
> many people would think this took away the exploring aspect or
> something, and would not like it.
NWN has this, and it's quite a useful feature, though you're right - it does
detract from the exploration aspect if you overuse it.
Paul Drallos
adventure, I think the graphics are very distracting, and reduce the impact of the
text.
I also think that text IF requires much more attention and concentration to play,
while graphical games can be played more casually. So when I'm playing a text IF
game, I find *anything* else distracting. I even find the auto-mapping and
description box distracting in Zork Zero. I think the game would have been much
better if it had used the standard z5 format.
On the other hand, many times I am more in the mood for a graphic game because they
don't require as much concentration and I often don't want to have to work so hard
as I do in a text game. In that case, the text is a distraction.
Text and graphics are both wonderful, but the sum is less than the parts.
Kevin Forchione
news:61uDb.54299$aT.2...@news-server.bigpond.net.au...
Fundamentally, you're dealing with different psychological effects on the
audience. If you think about the ways that movies and books satisfy you,
you'll come closer to an answer that will explain the phenomenon. Also, with
text adventures there is an intermediate step, a translation that the brain
must make. This intermediate step engages the brain differently from, say,
what happens when you watch an action played out in a movie or play. It may
be that these two modes are not very compatible.
Although I have no data to support my hypothesis, I suspect that the
difficulty is similar to that of reading a book while trying to watch
television. Students do claim to do it, but there is a loss of concentration
on both sides. The brain-wave shift between the translation process and the
non-translation process is probably non-trivial. Hence the observation made
by many players that they found one element or the other _distracting_.
The graphical IF hybrid can probably be made to work in the same way that
experimental theatre works - on the fringe, and with a limited audience.
--Kevin
Mike Rozak
> objects one can interact with.
I've experienced this too. I also experienced the reverse, where the change
in cursor as you move it over the scene makes it way too obvious what should
be done. But, I had similar problems with text IF where it isn't obvious if
an element of the description is actually an object or not, since some
objects can be hidden from the object list. One possible solution is to
provide a list of objects, and/or leave it up to the author whether the
objects are listed in his/her work.
Elder Scrolls:Morrowind, while a RPG, had an interesting solution to this.
Everything in the room/scene was an object. If you walked into a kitchen,
there were random bowls, jars, and other bits lying around. You could pick
any of them up, and wander around the world with as many useless trinkets as
you wished. They didn't have rocks lying around (like the game you cited),
but the principle could be extended. Unfortunately, such a solution spawns
other problems.
--
Mike Rozak
www.mXac.com.au
Mike Rozak
I agree that reading text and trying to look at images is pointless. Would
you say the same about "NARRATION and graphics"? I don't have an IF
precedent to go on, but TV documentaries and slide-shows are a combination
of narration and graphics, and they to work well for me.
--
Mike Rozak
www.mXac.com.au
Mike Rozak
> Probably the worst point of all, in that it is tremendously difficult to
> cater graphically for the plethora of actions that the player could take -
> this is why most games that have graphics do so only to supplement
location
> descriptions, rather than depict the action.
Completely agree. Animation is a real pain, and some events that can be
described in a few words take an awful lot of animation.
> Fundamentally, you're dealing with different psychological effects on the
> audience. If you think about the ways that movies and books satisfy you,
> you'll come closer to an answer that will explain the phenomenon.
There definitely are different psychological effects happening. I'm not sure
how to deal with this, or if I pull out the old artistic phrase, "Every
medium has its place." Some scenes are best portrayed in oil paintings while
other scenes work better as water colors.
> Although I have no data to support my hypothesis, I suspect that the
> difficulty is similar to that of reading a book while trying to watch
> television. Students do claim to do it, but there is a loss of
concentration
> on both sides. The brain-wave shift between the translation process and
the
> non-translation process is probably non-trivial. Hence the observation
made
> by many players that they found one element or the other _distracting_.
I agree about the shifting between reading and seeing. However, if the text
is narrated is it still a problem? If there is a clash between narration and
interactive visuals then the prospect of a graphical IF is hurt
sigificantly.
Maybe I should more clearly define what I'm thinking about since that will
allow people to more accurate assess it:
Possible IF UI #1:
- Start with a text IF, but have the computer speak the text, either with
pre-recorded audio or text-to-speech. (The text could also be displayed, but
I suspect that won't be necessary unless the player is deaf or playing in
quiet mode.). Also throw in sound effects.
- While the narration is going on, bring up "slides" (taking up all/most of
the screen) that illustrate key points in the narration. The slides are JPEG
images, or automatically generated by the 3D engine. (Animations are
possible, but they're a lot of work for an author to create, unless they're
simple camera or object movements.) Example: "You say hello to the dragon
but it glares back at you silently." could be accompanied by a still-image
slide of the dragon glaring at the user.
- If the user examines an object, he/she could be presented with a narration
"It's a long wooden stick with some fancy carving at the end", along with a
3d model of the object that could be rotated and zoomed in on.
- If the user talks with a NPC then show an animated talking head. I have
seen OK-looking automatically animated heads (lip sync based on what
speaking, automatic blinks, automatic eye movement, automatic head movment).
An author would be able to add extra cues, such as emotions or specific
movements to improve the quality.
- When just looking around a room provide a 360 degree pan capability like
Zork Nemesis or Myst III. Perhaps provide small thumbnail images of objects
in the room or the user's inventory.
- Clicking on hot-spots in the room image could act like a traditional
graphcs aventure. (You could do the same for examining objects or the
talking head, but poking a NPCs eyes would only annoy him/her.)
- Provide an edit field for typing commands, and/or provide speech
recognition so a user could speak the commands. (The edit field vs. clicking
provides its own UI dilemma since it's a mental shift to switch between
clicking and typing/speaking. I think this is part of the reason why
graphical IF systems don't provide a command edit field anymore.)
- Alternatively, provide a menu of choices.
- At any stage an author could decide NOT to show images and just end up
with text IF that's read by the computer.
Possible UI #2: (Although not exclusive to UI #1)
- Begin with a 3d model of the environment and let the user wander around
the environment, like in any MMORPG or first-person shooter.
- Provide narration only when the user sees a new object or enters a new
location. (Or if requested). MMORPG and other first person games do NOT
provide narration, but I think it's important, especially when the PC has
visited a location before but the player has not. For example: The player
walks into a bedroom an hears the PC/narrator say, "This is my brother's
bedroom. He's a real idiot. Etc." Similar information is common for text IF
to give out in a room description.
- Unfortunately, this approach requires more animation, signfnicantly
increasing the amount of work for the author to do. Still scenes require no
extra work. Walking (and other common tasks) would be easy, but if the
author wants a juggling NPC he/she would have to animate it. (Which would
mean that authors wouldn't include juggling NPCs, or other
complex-to-animate objects.)
- Clicking acts like a traditional graphical adventure. Edit field for
commands, as above. Or a menu, as above.
Possible UI #3:
- I haven't come up with one, but I'm sure others are out there. A 3rd
person view (like Syberia) could work, but I personally find them less
engrossing than 1st person graphical IF. I suspect most people prefer 3rd
person graphical IF since they seem more common than the 1st person view.
Just to complicate matters: UI #1 works really well for tablet PCs and
notebooks, allowing you to lie in bed and experience the IF. UI#2 works
really well for 3D virtual-reality glasses and data gloves.
--
Mike Rozak
www.mXac.com.au
M.D. Dollahite
>look at images is pointless. Would
>you say the same about "NARRATION
>and graphics"? I don't have an IF
>precedent to go on, but TV >documentaries and slide-shows are a >combination
of narration and graphics,
>and they to work well for me.
First of all, Mike, I'd like to point out that the majority of the larger
population does not feel as the majority (or at least the loudest minority, I
haven't been here long enough to know which) of the people in this group do.
If most people thought graphics distracted from text and vice-versa, then we'd
hardly have as many picture books, illustrated novels, magazines, and web pages
as we do, eh? Not to mention that two very famous and popular (in their time)
IF games, Shadowgate and Deja Vu, both combined text descriptions with
first-person graphics and a point-n-click interface.
Correctly mixing text and graphics can have a powerful effect -- people in fact
major in the subject in college -- incorrectly mixing them can indeed be
amateurish and distracting. But I firmly believe that the author of a
game-making engine should assume that the users of that engine will know what
they're doing; after all, no tool can instantly turn a novice into a master, so
don't limit the master by trying to make his tools novice-proof.
On the subject of narration, check out Zork: Grand Inquisitor. It has a
narrator (who also happens to be a visible character, but that doesn't really
matter). It also renders death scenes as transcripts of what a text version of
the game might display (including the fatal command typed by the player), which
might interest to you.
>- Clicking on hot-spots in the room image could act like a traditional
>graphcs aventure. (You could do the same for examining objects or the
>talking head, but poking a NPCs eyes would only annoy him/her.)
>- Provide an edit field for typing commands, and/or provide speech
>recognition so a user could speak the commands. (The edit field vs. clicking
>provides its own UI dilemma since it's a mental shift to switch between
>clicking and typing/speaking.
>- Alternatively, provide a menu of choices.
Two words: Verb Coin.
The Verb Coin class of interfaces provides the versatility of a text parser in
a form that works smoothly with a mouse-driven interface. I personally
consider it the best possible interface for graphical adventure games.
In fact, I've been thinking about ways to adapt the verb coin interface to text
IF. It would make a nice complement to the TADS 3 "topic inventory" system,
eliminating guess-the-verb the same way the topic inventory eliminates
guess-the-noun.
I believe you mentioned verb coin interfaces yourself in an earlier thread, so
I shouldn't have to say any more about it.
OKB (not okblacke)
> Maybe I should more clearly define what I'm thinking about since
> that will allow people to more accurate assess it:
>
>
> Possible IF UI #1:
> - Start with a text IF, but have the computer speak the text,
> either with pre-recorded audio or text-to-speech. (The text could
> also be displayed, but I suspect that won't be necessary unless the
> player is deaf or playing in quiet mode.). Also throw in sound
> effects.
>
> - While the narration is going on, bring up "slides" (taking up
> all/most of the screen) that illustrate key points in the
> narration. The slides are JPEG images, or automatically generated
> by the 3D engine. (Animations are possible, but they're a lot of
> work for an author to create, unless they're simple camera or
> object movements.) Example: "You say hello to the dragon but it
> glares back at you silently." could be accompanied by a still-image
> slide of the dragon glaring at the user.
>
> - If the user examines an object, he/she could be presented with a
> narration "It's a long wooden stick with some fancy carving at the
> end", along with a 3d model of the object that could be rotated and
> zoomed in on.
These last two bullet points describe the kind of thing I find
annoying in graphical IF -- and it's basically one of the points you
identified in your original post. It's annoying when the text and the
graphics are redundant. It's not so much that one distracts from the
other as that both seem to be annoyingly jostling one another for
position.
Most of the King's Quest games I played were like this: I go to
move a log, I see my character pick up the log and move it, and then the
game says to me "Alexander moves the log to the side." Well, thanks for
telling me: I can SEE that he's moved the log. Similarly, showing me a
picture of a glaring dragon and simultaneously saying "the dragon is
glaring" is going to annoy me.
I think your point about having the game speak the text rather than
printing it on the screen is a good one -- it generally is less
distracting if one element is auditory rather than visual -- but I think
that this can only be a convenience, never a cure. If you have so much
text that printing it on the screen would be distracting, you cannot get
around this by having the game say it out loud.
> - At any stage an author could decide NOT to show images and just
> end up with text IF that's read by the computer.
I think that if this is an option, the game isn't really what I
would call graphical IF -- it's text IF with graphics.
> Possible UI #2: (Although not exclusive to UI #1)
> - Begin with a 3d model of the environment and let the user wander
> around the environment, like in any MMORPG or first-person shooter.
>
> - Provide narration only when the user sees a new object or enters
> a new location. (Or if requested). MMORPG and other first person
> games do NOT provide narration, but I think it's important,
> especially when the PC has visited a location before but the player
> has not. For example: The player walks into a bedroom an hears the
> PC/narrator say, "This is my brother's bedroom. He's a real idiot.
> Etc." Similar information is common for text IF to give out in a
> room description.
This is, I think a much better mechanism, but. . .
> - Unfortunately, this approach requires more animation,
> signfnicantly increasing the amount of work for the author to do.
> Still scenes require no extra work. Walking (and other common
> tasks) would be easy, but if the author wants a juggling NPC he/she
> would have to animate it. (Which would mean that authors wouldn't
> include juggling NPCs, or other complex-to-animate objects.)
. . . right.
> Possible UI #3:
> - I haven't come up with one, but I'm sure others are out there. A
> 3rd person view (like Syberia) could work, but I personally find
> them less engrossing than 1st person graphical IF. I suspect most
> people prefer 3rd person graphical IF since they seem more common
> than the 1st person view.
Personally I like a third-person view, but I can see either one
working.
I just sort of came up with this thought, so maybe it's half-baked,
but it occurs to me that what's important here is that the visual and
text/speech streams have to be describing things that are
fundamentally along different axes. That is, the words must tell you
something that you cannot see in the picture, and the picture must show
you something that would be cumbersome to describe verbally.
For instance, a good use of words in a graphical game is for
dialogue between characters, or internal monologue of the type you
outlined above ("this is my brother's room, he's a dork. . ."), because
these are things outside the purvey of the image itself. Having a voice
narrate things like "the dragon glares at you" is not so good -- it's
better conveyed by a picture of a glaring dragon. Now, it may be the
case that many objects don't warrant a full image, so for them a verbal
description is fine, but there's no point in having a verbal description
AND a picture.
Basically, for me, it comes down to, what makes me like the
graphical IF that I like. This then basically comes down to what makes
the classic LucasArts graphical adventures (Indiana Jones and the Fate
of Atlantis, Full Throttle, Day of the Tentacle, etc.) seem so much
better to me than most other graphical IF. In particular, I find it
helpful to compare them to the Sierra games (King's Quest, Laura Bow,
etc.), which are in many ways quite similar to the LucasArts ones, but
which always seemed much cheesier and less good through and through than
the LucasArts games. Part of the difference is, I think, what I
outlined above -- and these are the differences that have most to do
with the interface.
--
--OKB (not okblacke)
"Do not follow where the path may lead. Go, instead, where there is
no path, and leave a trail."
--author unknown
Paul Drallos
is passive and works very well with graphics.
Paul Drallos
> If most people thought graphics distracted from text and vice-versa, then we'd
> hardly have as many picture books, illustrated novels, magazines, and web pages
> as we do, eh? Not to mention that two very famous and popular (in their time)
There is a big difference between illustrated novels, magazines, etc., and
illustrated IF. In non-interactive illustrated text, the outcome or advancement
of the plot does not depend on how carefully the text is read or the illustrations
scanned. In IF, the smallest detail may be critical to the advancement of the
work and missing it may be fatal. In books, the story goes on exactly the same
regardless of whether the text or graphics are carefully examined.
I love pictures, I love text, I love pictures and text. But I find one or
the other distracting in IF.
mad...@bmrb.wisc-antispam.edu
It's not the passivity that matters. It's the fact that in one case you
are trying to push two information streams through a single sense (sight
for graphics and sight for reading the narrative) and in the other you
are pushing the two information streams, each through their own sense
(sight for graphics and audio for narration).
Your eyes cannot be focused on the words and the graphics at the same
time, so you have to toggle back and forth between them. The *same*
problem would happen if you made a game for the blind where everything
was audio, and you tried having a voice describe the room layout
while another voice narrated at the same time. The listener would
have to choose which to pay attention to and which to filter out
as background noise, because your brain can't easily parse to two
streams of language at the same time.
Mike Rozak
parser in
> a form that works smoothly with a mouse-driven interface. I personally
> consider it the best possible interface for graphical adventure games.
It seems to be the optimum if the author can only pick one of the possible
UI devices. What if the author could use a verb coin for the most
common/obvious actions, and a command line (or really-really long menu so it
would work with SR) for everything else? How much would two possible input
methods confuse the user and/or weaken the experience? I don't know. (Mouse
and keyboard (arrows, and ilk) seems to work well, but mouse and typed
command line doesn't - from my experience.)
From another post:
> I just sort of came up with this thought, so maybe it's half-baked,
> but it occurs to me that what's important here is that the visual and
> text/speech streams have to be describing things that are
> fundamentally along different axes. That is, the words must tell you
> something that you cannot see in the picture, and the picture must show
> you something that would be cumbersome to describe verbally.
"Orthogonality" is the mathematical term for what you're describing. (I took
far too much math.) Totally agree. My examples were flawed.
> First of all, Mike, I'd like to point out that the majority of the larger
> population does not feel as the majority (or at least the loudest
minority, I
> haven't been here long enough to know which) of the people in this group
do.
> If most people thought graphics distracted from text and vice-versa, then
we'd
> hardly have as many picture books, illustrated novels, magazines, and web
pages
> as we do, eh?
Good point.
This brings up another design issues: No matter what feature is proposed
some people will be for and some against. In a broader sense, this applies
to the basic system design. The majority of potential IF users in the world
"obviously" prefer graphics like Myst or Syberia because they sell more
copies. Text IF definitely doesn't sell as much, and probably doesn't
distribute as many free copies either.
However:
1) I don't really want to cater for the majority. There are plenty of large
companies that do that, and they do it well. Plus, I have never really been
part of the majority.
2) While I prefer graphics, text IF titles have something about them that's
lost in graphical IF. The style of IF changes depending upon the medium,
with some things lost and some gained. I'm trying to wrap my mind around the
reasons why this change occurs in the hope that I can create a graphical IF
with (some of) the depth of text. (Which may end up being a quest for the
holy grail.)
As a result, it's good to hear from people that don't like graphical IF
because they are more in-tune with what's lost in the transition from text
to graphics. I suppose that before choosing a design, I want to know why it
works, and why it doesn't work. Since I'm naturally biased towards graphical
IF, I need to hear an alternate POV. (Hearing from people that like
graphical IF is also good, since some like different aspects than I do.) So
far, responses to my posts have brought up many good issues for me to
ponder.
Design is full of hidden pitfalls, so other people's experience is very
useful. Here is example of another idea I've been thinking about, but which
isn't really related to IF. It does illustrate the design issues that keep
popping up:
I like the concept in a MMORPG where users can interact with one another,
and even share the plot. If I had never played a MMORPG, I'd be gung-ho
about this idea. However, when the concept is applied to reality, the
multi-user aspect doesn't work well because:
1) There are 100K people wandering around in the virtual world, many of
which you won't like. Running into likable people is fine, but the odd
encounters with griefers puts some people (or at least me) off.
2) Because there are 100K people, it's impossible to have any plots that
involve one person changing the world. It's impossible to have puzzles and
not have people "cheat" by posting all the solutions. (Although they're
really cheating themselves). And, it's very difficult to have anything other
than just pure combat because 100K people are running all over one another
stepping on each others' toes. Even combat is flawed because you may get to
the dragon's lair to find it slain only a few moments before. The net is a
static world with regenerating hoards of monsters.
3) There are some 100+ MMORPG in development, many with $10M+ budgets. The
budgets will only increase exponentially. I can't compete.
I have come up with one solution to this, although it has flaws: Allow
players to play offline, but they can always go online for some
socialization. The only thing users can do online is chat with one another
when in nearby pieces of the IF. Other users will not be able to affect each
other's worlds - they're all in independent, so you won't have griefers
getting to the treasure before you. Furthermore, the IF subject matter
chosen by the author could be used to self-select certain types of users.
For example: if there isn't any combat in the IF people that like killing
will go elsewhere. Lots of puzzles will encourage puzzle-solvers, etc.
... This also has a bizarre metaphysical(?) aspect to it, which is totally
off subject but rather interesting: Multiple time-lines will converge. To
illustrate: Assume that Harry Potter is turned into IF. Harry Potter goes
into Hogsmead several times per book, so that could be one of the places
where people interested in a chat might hang out. What happens when you have
several hundred players chatting away in Hogsmead when they are all on
DIFFERENT timelines? Some of the people are still in book 2, while some are
on book 7. (Of course, some idiot who has reached book 7 will give away the
plot.) MMORPG and MUDs only have one time line, so this is never a problem.
Unfortunately, such a solution just turns the IF into a large chat room. I
don't know this for sure, but I've been reading about MUDs. They started out
as multi-user IF. From what I've read, players quickly consume the content
and then need to find other ways to entertain themselves. They either leave,
go on repetitious moster hunts, try to find flaws with the system, beome
wizards/builders in the virtual world, or socialize. With a single-user
system, the user just shelves the program and pulls out another one. (Many
IF readers also become builders, as this newsgroup shows.)
Life for the author seems a lot simpler in IF than a MUD. But, if IF were
more MUD-like it would be easier for the author to see how people experience
their work, enabling them to do a better job. (I REALLY like this aspect
multiuser systems. Feedback from beta testers is good, but invisibly
watching a beta tester try to use your product is equally (or more) useful.
Being able to step in and ask "Are you having a problem?" is also useful
because a) it saves the user from total frustration, and b) you can get
feedback about the beta tester's current state-of-mind which he/she would
never remember to send you. Not to mention the opportunity for machine
learning from all the data that could be collected.)
And so the pros and cons go on ad infinitum... If I ever did a multiuser
system it would be a LONG time from now. Knowing that I might impliment
multiuser ability allows me to make some simple changes to the design now
that make implimenting multiuser much easier later on, though. Looking a few
years out is always a good thing.
--
Mike Rozak
www.mXac.com.au
Daniel Barkalow
> I agree about the shifting between reading and seeing. However, if the text
> is narrated is it still a problem? If there is a clash between narration and
> interactive visuals then the prospect of a graphical IF is hurt
> sigificantly.
Two games you should look at: "City of Secrets" and "Words of Power"; both
are GLULX games with significant graphics (and optional sound).
> Maybe I should more clearly define what I'm thinking about since that will
> allow people to more accurate assess it:
>
>
> Possible IF UI #1:
> - Start with a text IF, but have the computer speak the text, either with
> pre-recorded audio or text-to-speech. (The text could also be displayed, but
> I suspect that won't be necessary unless the player is deaf or playing in
> quiet mode.). Also throw in sound effects.
Good so far, except that speech is much slower than reading. Also, IF
tends to repeat things a lot, since the information is often useful and is
easy to skip if it's not needed.
> - While the narration is going on, bring up "slides" (taking up all/most of
> the screen) that illustrate key points in the narration. The slides are JPEG
> images, or automatically generated by the 3D engine. (Animations are
> possible, but they're a lot of work for an author to create, unless they're
> simple camera or object movements.) Example: "You say hello to the dragon
> but it glares back at you silently." could be accompanied by a still-image
> slide of the dragon glaring at the user.
Sure, although I think it would be best to have some convention for what's
on the screen while the game tells you things that don't make good images
or that the game doesn't have an image of. Consider a puzzle where you
have to throw something (it doesn't matter what, so long as it's not
something fragile) at a target; this is going to be a pain to illustrate
convincingly.
> - If the user examines an object, he/she could be presented with a narration
> "It's a long wooden stick with some fancy carving at the end", along with a
> 3d model of the object that could be rotated and zoomed in on.
This could certainly be nice for objects with interesting details (many of
the items from So Far, some from Slouching, etc.); I'd leave most objects
to the player's imagination, and avoid issues like how everything fits
into a backpack, and things that don't look like much (a spherical seed
pod, e.g.).
> - If the user talks with a NPC then show an animated talking head. I have
> seen OK-looking automatically animated heads (lip sync based on what
> speaking, automatic blinks, automatic eye movement, automatic head movment).
> An author would be able to add extra cues, such as emotions or specific
> movements to improve the quality.
I think a static image might be just as good, especially if you're not
going to have voices for each NPC. A set of images with different emotions
could be useful.
> - When just looking around a room provide a 360 degree pan capability like
> Zork Nemesis or Myst III. Perhaps provide small thumbnail images of objects
> in the room or the user's inventory.
I'm not sure this is worthwhile. I'd save the graphics for scenic
vistas. The main case in which I'd render a whole room would be if the
player was looking into it from a different room through a window or
something. The point is essentially to give the player a good idea of what
it's like to be somewhere, and a general view is not really that
useful. You really want to use some photographic composition to convey
what you want to convey.
> - Clicking on hot-spots in the room image could act like a traditional
> graphcs aventure. (You could do the same for examining objects or the
> talking head, but poking a NPCs eyes would only annoy him/her.)
This is, of course, rather limited.
> - Provide an edit field for typing commands, and/or provide speech
> recognition so a user could speak the commands. (The edit field vs. clicking
> provides its own UI dilemma since it's a mental shift to switch between
> clicking and typing/speaking. I think this is part of the reason why
> graphical IF systems don't provide a command edit field anymore.)
I think that typing text is necessary in order to have interesting
interaction with the world. And switching between modalities is worse than
just sticking with text for everything. It might be useful to allow
pointing at pictures to identify the things in them, though.
> Possible UI #2: (Although not exclusive to UI #1)
> - Begin with a 3d model of the environment and let the user wander around
> the environment, like in any MMORPG or first-person shooter.
This is actually relatively difficult to write; you need to come up with a
substantial amount of architecture for even a modest game. Furthermore, IF
maps are designed in a highly stylized way, which omits a long of
connective areas which aren't interesting and focuses on the important
parts. They ignore differences in scale, and let you walk to town in a
single easy action, while requiring a number of actions and more time to
walk through a house. It would be difficult to design a 3D model which was
both reasonably realistic and at all useful for story-telling.
-Iabervon
*This .sig unintentionally changed*
Michael Bechard
> On Tue, 16 Dec 2003 02:24:02 GMT, Mike Rozak scrawled:
>
>
>>Is there anything else I should add to the list? Before I undertake such a
>>huge task, I'd like to know what elements could cause (or has caused)
>>amateur graphical IF to fail.
>
>
> *snip*
>
> Probably the worst point of all, in that it is tremendously difficult to
> cater graphically for the plethora of actions that the player could take -
> this is why most games that have graphics do so only to supplement location
> descriptions, rather than depict the action.
>
Well, actually there are plenty of graphic adventure games that animate
all the PC's actions rather well. PC's bend down to pick things up,
bring objects to their mouth to eat them, etc. The Gabriel Knight series
has some examples of this kind of animation. The real issue is that,
generally, not that many actions are implemented in modern graphic
adventure games, compared to IF. When I sit down to play IF, I can type
in literally anything, and it can be a candidate for a player's action.
With graphic adventures, you can only make do with what verb icons you
are supplied with.
Michael
Rexx Magnus
>
> Well, actually there are plenty of graphic adventure games that animate
> all the PC's actions rather well. PC's bend down to pick things up,
> bring objects to their mouth to eat them, etc. The Gabriel Knight series
> has some examples of this kind of animation. The real issue is that,
> generally, not that many actions are implemented in modern graphic
> adventure games, compared to IF. When I sit down to play IF, I can type
> in literally anything, and it can be a candidate for a player's action.
> With graphic adventures, you can only make do with what verb icons you
> are supplied with.
>
> Michael
>
Well, yes, but I'd assumed we were talking about an IF interface with a
graphical representation of what was going on - which causes it to be a
more complicated kettle of fish. :)
Mike Roberts
> The real issue is that, generally, not that many actions are
> implemented in modern graphic adventure games, compared
> to IF. When I sit down to play IF, I can type in literally anything,
> and it can be a candidate for a player's action. With graphic
> adventures, you can only make do with what verb icons you
> are supplied with.
To be fair, the claim that "I can type literally anything" could just as
well hold true in a graphical game. If you think about it, most of the
"literally anything" you can type in a text game is just random stuff you're
making up that isn't understood by the game; you could just as well make up
your own random control inputs that are equally not understood in a
graphical game. For example, while playing a graphical adventure, you could
say to yourself: "if I move the mouse really fast to the right, I want it to
mean 'run east.'" That would be every bit as meaningful as saying to
yourself, while playing some text game: "if I type in RUN REALLY FAST EAST,
it means 'run really fast east." In each case, you're just making up your
own command that has nothing to do with what the game accepts as valid
inputs, and in each case the game responds, in its own way, "I don't
understand."
The only reason the two cases *seem* different is that the graphical game
tells you up front what control inputs it accepts, while the text game
leaves it up to you to figure out what inputs are valid. The graphical
game's up-front enumeration doesn't in any way prohibit you from making up
your own inputs, but once you have the valid ones laid out for you, it just
never occurs that you would try anything else. The text game's use of a
(tiny) subset of human natural language also obscures what's going on. The
text game is actively misleading; it encourages the naive player to make a
generalization error, specifically "valid commands look like natural
language -> natural language yields valid commands." Some people claim to
like that illusion of control, but I claim that only an experienced player
is capable of maintaining the illusion, because only an experienced player
intuitively limits herself to the valid command set; the inexperienced
player is quickly disillusioned by the fifty "I don't understand that
command" errors in a row, and goes to find a nice graphical adventure to
play.
You're right that the *real* issue is the range of valid inputs actually
implemented, not the amount of illusion inherent in the UI. The bang for
the buck in implementation effort is obviously much higher working in text
than in animated graphics; even a hobbyist author working alone on a text
game can run circles around a highly paid team of professionals building a
graphical game, in terms of the number of valid inputs and distinct events
implemented.
--Mike
mjr underscore at hotmail dot com
Mike Rozak
> are GLULX games with significant graphics (and optional sound).
Thanks. I'll check these out.
> Good so far, except that speech is much slower than reading. Also, IF
> tends to repeat things a lot, since the information is often useful and is
> easy to skip if it's not needed.
Good issue. Hadn't thought about that.
> Consider a puzzle where you
> have to throw something (it doesn't matter what, so long as it's not
> something fragile) at a target; this is going to be a pain to illustrate
> convincingly.
As you point out, animation would be a major pain. Animating the thing being
thrown, while in mid-air, isn't terribly difficult (newton's laws). The act
of a character throwing and the bounces off the ground are. My inital guess
at how to solve this is to just narrate, "You throw X at Y. It misses", and
leave it at that. Maybe throw a slide in to show object X sitting outside of
Y to reinforce that it misses.
Of course, it would all be up to the author, but I would weigh off how much
work an animation vs. how much it improves the story. If the throwing
weren't an integral part of the story, it's not worth spending the 1+ hours
to animate it.
> I think a static image might be just as good, especially if you're not
> going to have voices for each NPC. A set of images with different emotions
> could be useful.
A static image would work.See http://www.curiouslabs.com for poser. They
provide with "poses" for different facial expressions; I'll get that
impimented at some point. I also do lip sync from audio, BUT doing a 30 fps
animation might be taxing depending upon the detail that the author choses
to use for the character models. I have seen a 500 polygon head animated at
30 fps real-time, but a head whose quality is the same as poser is
5000-10,000 polygons = 3 fps. In time this will be possible. (This assumes
that the hair is drawn game-style, and not realistically. To draw completely
realistic hair, like you see in Monsters Inc., can take 1 hr. per frame.)
So, the author might be forced to use a static image.
I do have the ability to make different voices, either from TTS or from a
recording of an original speaker. See http://www.mxac.com.au/m3d/tts.htm.
The algorithm isn't going to replace actors, but it will allow an amateur to
create a variety of different voices from their own. If a different TTS
voice is acquired by the author, even more voices can be created. Plus,
authors can always call up their friends and have their friends record.
> I'm not sure this is worthwhile. I'd save the graphics for scenic
> vistas. The main case in which I'd render a whole room would be if the
> player was looking into it from a different room through a window or
> something. The point is essentially to give the player a good idea of what
> it's like to be somewhere, and a general view is not really that
> useful. You really want to use some photographic composition to convey
> what you want to convey.
I appreciate your point, but don't agree entirely. Ideally, the author
wouldn't have to show 360 room views for rooms that aren't visually
interesting, but once you show an image from one room, the user is going to
expect it for all of them. This brings up complications of how much detail
to render into an image. In pure text, if the author omits certain details
the reader magically isn't bothered by their omission. In graphics, if
they're omitted and expected by the user (such as a kitchen scene lacking a
sink) the reader will notice. This is a flaw of graphics, and one reason why
it's so much work.
> I think that typing text is necessary in order to have interesting
> interaction with the world. And switching between modalities is worse than
> just sticking with text for everything. It might be useful to allow
> pointing at pictures to identify the things in them, though.
This is one of my worries. I too suspect that switching modalities is bad.
One alternative is to rely on object-specific interaction menus (like many
Lucas games). The other is to use speech recognition, which I don't think
will have a modality problem with mouse click. But SR has accuracy probolems
which may (or may not) cause a problem. The only way to tell if SR will work
is for me to try it out. If I organize the text parser properly then SR will
be easy to impliment. SR also has the advantage of working on tablet PCs
(aka: PDAs with enough CPU to run graphics IF), which would be a good win.
SR is a must when using VR glasses.
> This is actually relatively difficult to write; you need to come up with a
> substantial amount of architecture for even a modest game. Furthermore, IF
> maps are designed in a highly stylized way, which omits a long of
> connective areas which aren't interesting and focuses on the important
> parts. They ignore differences in scale, and let you walk to town in a
> single easy action, while requiring a number of actions and more time to
> walk through a house. It would be difficult to design a 3D model which was
> both reasonably realistic and at all useful for story-telling.
There are two point here:
I'm trying to make the construction of 3d models as easy as possible. But
not matter what I do it will always be more difficult to construct 3d than
write the lines of text for description. Another flaw in graphics.
Also agree with your comment about how a nodal representation of space (as
in text IF) has a convenient way of skipping the dull bits of scenery which
a spatial representation (MMORPG, 1st person shooter) does not. To some
extent the author can work around the problem, but not completely.
--
Mike Rozak
www.mXac.com.au
Jeff
Hello Mike!
I'm one of those people who tends to dislike graphical IF games. The
main reason is that they don't draw me into the story like an all text
game can. Consider the following bit of text, taken from the
description of the first room in Michael Gentry's classic
"Anchorhead".
"Outside the Real Estate Office.
A grim little cul-de-sac, tucked away in a corner of the
claustrophobic tangle of narrow, twisting avenues that largely
constitute the older portion of Anchorhead. Like most of the streets
in this city, it is ancient, shadowy, and leads essentially nowhere.
The lane ends here at the real estate agent's office, which lies to
the east, and winds it's way back to the center of town to the west. A
narrow, garbage-choked alley opens to the southeast."
This is a great example of the author taking advantage of the text
medium. Not only do I know what's there, I know exactly how I'm
supposed to feel about what's there, how it looks to the protagonist.
This is something a picture can't do. If this was a picture instead of
a description I likely wouldn't think "grim" when looking at the
cul-de-sac, "claustrophobic" when looking at the streets, etc. What I
would probably do is glance briefly at the picture, see there's no
object to pick up, and click the door. Once I found that locked I'd
try the alley. Even if I did stop to try to get a feel from the
picture before moving on, who's to say I'd see it the way the author
wanted me too? Worse still, if there's a narration along with the
picture I might disagree with the author. It's very hard for the
author to convince me to see the cul-de-sac as grim if I look at the
picture and think it looks cozy. At that point, I'm no longer into the
story so much as I'm now disbelieving it.
Anyway, good luck with your project.
JAN THORSBY
Mike Roberts <mjrUND...@hotmail.com> skrev i
meldingsnyheter:Z4oEb.18$Da4...@news.oracle.com...
> "Michael Bechard" <not...@nothing.com> wrote:
> The only reason the two cases *seem* different is that the graphical game
> tells you up front what control inputs it accepts, while the text game
> leaves it up to you to figure out what inputs are valid.
There is a point you overlook. When you are told the all the commands it can
be considered spoilers. What if the soulution to a puzzle is to dance? (I
don't know why, maybe to impess some woman at a disco or something.) If you
where told upfront that it is possible to dance, it would be a very easy
puzzle.
mad...@bmrb.wisc-antispam.edu
:
:Mike Roberts <mjrUND...@hotmail.com> skrev i
:meldingsnyheter:Z4oEb.18$Da4...@news.oracle.com...
:
:> The only reason the two cases *seem* different is that the graphical game
:> tells you up front what control inputs it accepts, while the text game
:> leaves it up to you to figure out what inputs are valid.
:
:There is a point you overlook. When you are told the all the commands it can
:be considered spoilers. What if the soulution to a puzzle is to dance? (I
:don't know why, maybe to impess some woman at a disco or something.) If you
:where told upfront that it is possible to dance, it would be a very easy
:puzzle.
Yes, but what frustrates a lot of players is when they think they
know what they need to do, but the game doesn't recognize it so
they assume it was wrong when it was actually on the right track.
(For example, if the player knew the character had to dance, but
it turns out it's not "dance", but "waltz" that's the right verb.
So after a lot of time spent trying to "dance" or "sashay" or
"rhumba" or "boogie" or "gyrate" or "mamba", the player eventually
gives up, assuming that the solution must be something other than
dancing. Then the player looks up in some hint book and finds out
that dancing *was* the right answer, but had to use the word "waltz".
That's going to anger the player since he knows he had the answer
right and a human being would have recognized it as correct even
though the computer doesn't.)
Another annoying feature (this is one that was very common in
Hitchiker's Guide to the Galaxy) is when the text description
of things is insufficient to tell you the info to solve the
problem. For example, there's a tool that you are supposed to
use to remove part of your brain. But when you look at the tool,
it never says what it's for, never says what it looks like,
and nowhere is there any hint telling you that this tool
is somehow a medical instrument. Then there was the "thing your
aunt gave you which you don't know what it is" (that's the
actual text description). It's an item that you can't get rid of,
because it keeps coming back to your inventory in a few turns.
It turns out that it's a container, and the ONLY way to discover
this is to type the command to use it like a container: "put foo in
thing", which you won't do if you don't realize it's a container.
(I never realized it was a container until I looked at some hint
book.)
And the worst part was trying to get the babel fish. It was a
complex puzzle during which one part of the puzzle involved a
trapdoor set into the floor, which flips up and a cleaning robot
pops out of it, sweeps stuff up, and then goes back in the
compartment and the trapdoor closes. To solve the puzzle you
had to realize that an object placed on the trapdoor would be
flung up into the air if the cleaning robot came out and the
door popped open. The big problem with this puzzle was that
nothing in the text descriptions helped tell me that the trapdoor
was a trapdoor in the floor. It was always described as a "panel",
and it was described in a manner that made me think it was set
in the wall near the floor, not in the floor itself. (This was
reinforced by the fact that the floor cleaning robot was
described as rolling on wheels (not flying), and so I naturally
was picturing a panel on the wall that the robot drives out of,
not something it would pop UP out of.) So anyway, because of
this description, I could never solve the puzzle without a
hintbook. I was trying to put things "in front of" the panel
instead of "on top of" the panel, and it kept telling me that
made no sense, without telling me why. That was particularly
frustrating because I was *right* in my guess that the answer
involved obstructing the panel in some way. It's just that
it gave an ambigious description of the panel that led me to
using the wrong prepositions every time I tried.
Michael Coyne
> this description, I could never solve the puzzle without a
> hintbook. I was trying to put things "in front of" the panel
> instead of "on top of" the panel, and it kept telling me that
> made no sense, without telling me why. That was particularly
> frustrating because I was *right* in my guess that the answer
> involved obstructing the panel in some way. It's just that
> it gave an ambigious description of the panel that led me to
> using the wrong prepositions every time I tried.
Just a little correction here. The panel was described properly; it was
set into the wall, and you *were* supposed / allowed to put things in
front of it.
SPOILERS!
The robot came OUT from somewhere and went INTO the panel in the wall.
You had to put the satchel in front of the panel. The robot would hit
it, and the fish would fly through the air and then get caught by the
flying cleaning robot. Then, you had to think of adding something to
the top of the satchel, so that more than just the fish would fly into
the air, confusing the flying robot. So you put the mail on top of the
satchel.
It sounds like you may be mis-remembering the puzzle.
Michael
Julio Juliopolis
That's just simply bad programming. If the command sought was "Waltz", there
should be a reply to "Dance" that would lead you to guess it. Such as...
"What kind of dance do you want to perform?" True, it takes more effort on
the designer's part to consider that than simply offering up a list of all
the usable commands, but I think it's worth it. In addition to the "spoiler
effect" mentioned above, (which could be minimized by include many
non-essential verbs), there's also a certain joy in trying something that
you're not sure if the parser will allow you to do and having it work. I'd
personally rather see...
>rhumba
You don't know how to rhumba very well.
>boogie
Don't be ridiculous, this isn't the 70s.
>sashay
I don't understand that command.
>gyrate
You do your best impression of the water buffalos you saw going at it on
"Wild Kingdom" last night. Unfortunately, this doesn't seem to impress
anyone. Tough crowd.
>mamba
I don't understand that command.
>waltz
You dance a perfect waltz, gaining admirers as you move about the floor. The
woman you noticed earlier seems to be watching you with a renewed interest
now.
...as a result of playing "guess the verb" than simply going down the list
of commands and typing each one in turn.
>
> Another annoying feature (this is one that was very common in
> Hitchiker's Guide to the Galaxy) is when the text description
> of things is insufficient to tell you the info to solve the
> problem. For example, there's a tool that you are supposed to
> use to remove part of your brain. But when you look at the tool,
> it never says what it's for, never says what it looks like,
> and nowhere is there any hint telling you that this tool
> is somehow a medical instrument.
Actually you don't need a tool to take your "common sense".
> Then there was the "thing your
> aunt gave you which you don't know what it is" (that's the
> actual text description).
It's a pouch. Arthur isn't exactly the brightest bulb on the tree.
> It's an item that you can't get rid of,
> because it keeps coming back to your inventory in a few turns.
My current theory on that is that the ghost of Arthur's aunt is following
him around returning it to him whenever he foolishly drops it somewhere.
> It turns out that it's a container, and the ONLY way to discover
> this is to type the command to use it like a container: "put foo in
> thing", which you won't do if you don't realize it's a container.
Well, since you didn't know what it is, surely you tried manipulating it in
a variety of ways, right?
> (I never realized it was a container until I looked at some hint
> book.)
>
> And the worst part was trying to get the babel fish. It was a
> complex puzzle during which one part of the puzzle involved a
> trapdoor set into the floor, which flips up and a cleaning robot
> pops out of it, sweeps stuff up, and then goes back in the
> compartment and the trapdoor closes. To solve the puzzle you
> had to realize that an object placed on the trapdoor would be
> flung up into the air if the cleaning robot came out and the
> door popped open. The big problem with this puzzle was that
> nothing in the text descriptions helped tell me that the trapdoor
> was a trapdoor in the floor. It was always described as a "panel",
> and it was described in a manner that made me think it was set
> in the wall near the floor, not in the floor itself. (This was
> reinforced by the fact that the floor cleaning robot was
> described as rolling on wheels (not flying), and so I naturally
> was picturing a panel on the wall that the robot drives out of,
> not something it would pop UP out of.)
Er, actually you were thinking correctly. It is a panel in the wall. You
block the panel with Ford's satchel. Then you put the mail on top of the
satchel, so when the robot hits that the mail flies through the air, along
with the fish which lands with a loud squish in your ear.
> So anyway, because of
> this description, I could never solve the puzzle without a
> hintbook. I was trying to put things "in front of" the panel
> instead of "on top of" the panel, and it kept telling me that
> made no sense, without telling me why. That was particularly
> frustrating because I was *right* in my guess that the answer
> involved obstructing the panel in some way. It's just that
> it gave an ambigious description of the panel that led me to
> using the wrong prepositions every time I tried.
I can easily see how a player might not get the correct image of what the
puzzle looks like from that description. I agree that for certain puzzles a
picture might prove invaluable, however I'd personally prefer the mainstay
of the game remain all text.
mad...@bmrb.wisc-antispam.edu
:
:SPOILERS!
I don't think you need to worry about spoilers for a game that's
been out of print now for over 15 years.
:
:
:The robot came OUT from somewhere and went INTO the panel in the wall.
:You had to put the satchel in front of the panel. The robot would hit
:it, and the fish would fly through the air and then get caught by the
:flying cleaning robot. Then, you had to think of adding something to
:the top of the satchel, so that more than just the fish would fly into
:the air, confusing the flying robot. So you put the mail on top of the
:satchel.
:
:It sounds like you may be mis-remembering the puzzle.
It must be. I might have had it backward - perhaps I was trying to
put things ON it and I had to put them in FRONT of it. But I do
remember trying over and over to PUT <thing> <preposition> PANEL in
a number of ways and getting the frustrating error message that it
doesn't make any sense to do that. So I never got to the part about
the mail on top of the satchel because I could never put anything
<preposition> the panel in the first place. When I finally looked
at a hintbook I can remember being very mad at the game for letting
me get so close and yet giving me an error message that made it sound
like I was miles away from the correct answer.
I wonder if a parser can be constructed that is aware of all the
legal prepositions for a verb-object pair and can thus give
better error messages for cases like this. Instead of
"That doesn't make any sense", it could say, "I can't do
that, but I could put something in front of it, or next to it."
Or, if that's too much of a hint, perhaps just "It has no
top surface to put things on" would at least hint to the
player that the problem is with the choice of the "on top of"
preposition phrase (as opposed to problem being because the
panel won't let you put things relation to it, or the problem
being that the satchel isn't an object you can put somewhere
in relation to the panel.
The problem was that as a player, the response from "put satchel
on panel" was the same as for "put guide on panel" or "put satchel
on guide" - so the response didn't clue me in that "satchel" and
"panel" can actually go together some way.
mad...@bmrb.wisc-antispam.edu
Yes, and I'd like a pony. The thing is, that level of detail takes
entirely too much time to expect the author to include all of it.
(Each and every puzzle would have to include each and every possible
alternate solution for it mapped out with proper messages. That
turns into an exponential progression pretty fast. A reasonable
solution is to reduce the options so you don't have to know the
*exact* verb to make it work. Make 'dance' and 'rhumba' and 'gyrate'
and so on all map to do the same as just "dance". This is essentailly
what graphical IF makers are doing with the mouse-driven interfaces.
There are only a limited set of actions - walk to X, Look at X,
take X, drop X, use X, and use X with Y. So anything you try to do
has to be expressed as one of those. (In the dancing example, it
would probably be some mouse action that equates to "use dance
floor area".)
:>
:> Another annoying feature (this is one that was very common in
:> Hitchiker's Guide to the Galaxy) is when the text description
:> of things is insufficient to tell you the info to solve the
:> problem. For example, there's a tool that you are supposed to
:> use to remove part of your brain. But when you look at the tool,
:> it never says what it's for, never says what it looks like,
:> and nowhere is there any hint telling you that this tool
:> is somehow a medical instrument.
:
:
:Actually you don't need a tool to take your "common sense".
:
:
:> Then there was the "thing your
:> aunt gave you which you don't know what it is" (that's the
:> actual text description).
:
:
:It's a pouch. Arthur isn't exactly the brightest bulb on the tree.
:
:
:> It's an item that you can't get rid of,
:> because it keeps coming back to your inventory in a few turns.
:
:
:My current theory on that is that the ghost of Arthur's aunt is following
:him around returning it to him whenever he foolishly drops it somewhere.
:
:
:> It turns out that it's a container, and the ONLY way to discover
:> this is to type the command to use it like a container: "put foo in
:> thing", which you won't do if you don't realize it's a container.
:
:
:Well, since you didn't know what it is, surely you tried manipulating it in
:a variety of ways, right?
If I took that approach with every item that is in the game, I'd still
be playing it 15 years later. It's unreasonable to expect people to
try every action in the game with every object until they find the
combinations that work.
Julio Juliopolis
> Julio Juliopolis <Ju...@juliopolis.com> wrote:
>
<snipping>
I think you overestimate the difficulty in that. I've used several IF
authoring systems which allowed for easy implementation of synonyms.
> (Each and every puzzle would have to include each and every possible
> alternate solution for it mapped out with proper messages.
That is the "ideal", but it's probably not a realistic one. Even in the
example I gave above I used the default "I don't understand that command"
twice.
> That
> turns into an exponential progression pretty fast. A reasonable
> solution is to reduce the options so you don't have to know the
> *exact* verb to make it work. Make 'dance' and 'rhumba' and 'gyrate'
> and so on all map to do the same as just "dance".
Just seperate those words with commas in the synonyms list for dance and
you're done. But if it really didn't matter what type of dance then simply
having one command "dance" should be good enough. I don't think the player
will consider rhumba and sashay yet miss simply typing dance when those
come back as unknown commands.
> This is essentailly
> what graphical IF makers are doing with the mouse-driven interfaces.
> There are only a limited set of actions - walk to X, Look at X,
> take X, drop X, use X, and use X with Y. So anything you try to do
> has to be expressed as one of those. (In the dancing example, it
> would probably be some mouse action that equates to "use dance
> floor area".)
I've used those types of interfaces and for me they just aren't as
entertaining. The point and click feels too much like hyperlink and the
graphics don't set the mood the way text descriptions do. (See my post as
"Jeff" in this thread).
<snip>
> :> It turns out that it's a container, and the ONLY way to discover
> :> this is to type the command to use it like a container: "put foo in
> :> thing", which you won't do if you don't realize it's a container.
> :
> :
> :Well, since you didn't know what it is, surely you tried manipulating it
> :in a variety of ways, right?
>
> If I took that approach with every item that is in the game, I'd still
> be playing it 15 years later. It's unreasonable to expect people to
> try every action in the game with every object until they find the
> combinations that work.
I just meant that the item's description actually said you didn't know what
it was, which should have encouraged you to play with it in a multitude of
ways. And remember this is Douglas Adams we're talking about here. How much
do you really expect a picture of the G'gvunts Ultra-plasmic Vacuum Awl to
help? You're not supposed to understand what those things are in that game,
because Arthur has, (much like yourself), never seen them before, with the
humorously ironic exception of the pouch his aunt gave him.
Most games however have items whose uses you can guess pretty easily. If you
find a game with a flashlight that doesn't understand the command "turn on
flashlight", odds are that author wouldn't have written a sensible game
with a graphical interface anyway.
Andrew Plotkin
> :
> :[various synonyms with different results]
> :
> :...as a result of playing "guess the verb" than simply going down the list
> :of commands and typing each one in turn.
>
> Yes, and I'd like a pony. The thing is, that level of detail takes
> entirely too much time to expect the author to include all of it.
> (Each and every puzzle would have to include each and every possible
> alternate solution for it mapped out with proper messages. That
> turns into an exponential progression pretty fast. A reasonable
> solution is to reduce the options so you don't have to know the
> *exact* verb to make it work. Make 'dance' and 'rhumba' and 'gyrate'
> and so on all map to do the same as just "dance". This is essentailly
> what graphical IF makers are doing with the mouse-driven interfaces.
But it's *not* essentially the same; that's Julio's point, and I
agree.
The particular example of the dance may be misleading. One can imagine
an IF scenario in which it matters which kind of dance you're doing;
and a different scenario in which it doesn't. One would implement them
differently. (One would also implement *the rest of the game*
differently, so that the player has enough sense of the world to
understand what the interesting range of action is.)
I have argued in the past that the essence of IF lies in *not* giving
a menu-list of actions, but instead giving a free range of action
which the player has room to explore. Whenever I encounter a menu
interface -- and this includes dialogue menus in text games as well as
verb pop-up menus -- I find that my sense of the game-world
immediately vanishes. I have a mechanical strategy available to me
("try everything on everything"), and it's easier to follow that than
to visualize myself as a *person* in a *situation*.
I decided that the reason I like Myst-style graphical IF is that it
*acknowledges* that selecting an action for an object is
uninteresting. In this sort of game, there's only one action -- click,
which means "use this". The explorable range of action becomes *the
visual world which is presented to you*, which you can explore to
discover which objects are reactive and interesting.
--Z
"And Aholibamah bare Jeush, and Jaalam, and Korah: these were the borogoves..."
*
* Make your vote count. Get your vote counted.
Julio Juliopolis
> Whenever I encounter a menu
> interface -- and this includes dialogue menus in text games as well as
> verb pop-up menus -- I find that my sense of the game-world
> immediately vanishes. I have a mechanical strategy available to me
> ("try everything on everything"), and it's easier to follow that than
> to visualize myself as a *person* in a *situation*.
I appreciate where you're coming from because that is exactly the same thing
that happens to me on those sorts of games. If it's a text-based puzzle
I'll wrap my mind about it much like a riddle. In a graphical format, I'll
try pointing and clicking for about 45 seconds before I bother trying to
figure out what the puzzle even is.
Another major advantage text games have over graphical for me is that the
text helps me to empathize with the character. This could be well
illustrated with your own game "Hunter in Darkness". If that game were
ported to a graphical interface and the text and graphics versions could
stand side by side for comparison I think it would be very obvious that the
text conveys certain aspects of the story in a way that you just can't get
from images.
J. Robinson Wheeler
> Michael Coyne <coy...@mtsdot.net> wrote:
> :
> :SPOILERS!
>
> I don't think you need to worry about spoilers for a game that's
> been out of print now for over 15 years.
Yes, you do. It's an especially polite gesture when one is about to reveal
pretty much the entire solution to an all-time classic puzzle, one which
is really rewarding to solve on one's own.
There may be readers of this newsgroup who have the Hitchhiker's game
on their "will play that someday" list, who really ought to get the chance
to solve it on their own steam.
The sensibility of including spoiler warnings remains valid for other games
as well, regardless.
--
J. Robinson Wheeler Games - http://raddial.com/if/
JRW Digital Media Movie - http://thekroneexperiment.com
j...@jrwdigitalmedia.com Comic - http://adamcadre.ac/comics.html
Rikard Peterson
news:b5b859fd.03121...@posting.google.com:
> mad...@bmrb.wisc-antispam.edu wrote:
>>
>> I don't think you need to worry about spoilers for a game that's
>> been out of print now for over 15 years.
>
> Yes, you do. It's an especially polite gesture when one is about
> to reveal pretty much the entire solution to an all-time classic
> puzzle, one which is really rewarding to solve on one's own.
>
> There may be readers of this newsgroup who have the Hitchhiker's
> game on their "will play that someday" list, who really ought to
> get the chance to solve it on their own steam.
There is at least one such reader.
M.D. Dollahite
>Worse still, if there's a narration along with the
>picture I might disagree with the author. It's very hard for the
>author to convince me to see the cul-de-sac as grim if I look at the
>picture and think it looks cozy. At that point, I'm no longer into the
>story so much as I'm now disbelieving it.
That's an issue of bad authoring, not a general problem with mixing text and
graphics. It could certainly happen if the game author used a bad artist, but
a good artist should be able to make the picture convey exactly what it's
supposed to to anyone who isn't completely insane.
In regards to all the discussion of the limitation of graphical interfaces and
the freedom of a text parser: This is why I like the verb coin class of
interfaces so much.
The reason graphical games have such a limited range of action is not that the
programmers are lazy or anything like that, it's that there's simply not enough
room on the screen to display icons for 300+ verbs, most of which are just
going to respond "That action is totally illogical to even attempt." As a
result, they boiled all the verbs down to just a few basic categories, and even
then you get a lot of "illogical" responses.
The verb coin system, popularized by one of the later Monkey Islands and also
used in other forms by Return to Zork and Leisure Suit Larry 7, is designed to
allow graphical games to handle large numbers of verbs without having to crowd
them all onto one menu. In this system, you select an object to interact with
first, and then the verb icons are displayed. This allows the game to display
only the verbs that make sense with that object. It doesn't guarantee that all
displayed verbs have a useful result, and in some cases it doesn't even
guarantee that all verbs are displayed (see LSL7). It does guarantee that you
don't waste screen real estate on a "pour" icon when the direct object is a
Rubik's Cube, and thus you can implement a verb set of arbitrary size and
expressiveness just like text games can.
With the verb coin system, the only verbs that are removed from consideration
by the enumeration of options are the ones that you probably weren't
considering in the first place. In fact, a verb coin system could have an even
larger range of verbs than text games, because the interface itself helps the
player narrow the search space to only the logical options, and also allows for
special commands that only work at one particular point in the story (much like
the special topics in TADS 3's conversation system). Unless the programmers
are in fact lazy, the verb coin system should usually provide even more options
than you thought of yourself. Yes, there will always be a couple verbs the
programmers didn't think of, but that isn't any different from text games,
really, as Mr. Roberts already pointed out.
:> Then there was the "thing your
:> aunt gave you which you don't know what it is" (that's the
:> actual text description).
:
:
:It's a pouch. Arthur isn't exactly the brightest bulb on the tree.
Actually, it's not a pouch, at least not as an Earthman would know it. Note
that it has infinite capacity but negligible bulk and weight, even when every
other item in the game is inside of it. That's certainly not your everyday
briefcase. I always figured that item was described like that because Mr.
Adams didn't feel like coming up with some technobabble B.S. to explain it when
he could just make a joke instead.
Incidentally, I don't remember how exactly I first figured out what the thing
was for, but I don't think it was either random experimentation (I'm not the
kind who does that) or a hintbook (I didn't have one at the time). I think
there was a clue somewhere in Arthur's bedroom about it, or in the description
of the item itself. I could be mistaken, though.
And a direct response to the guy who started all this:
"It seems to be the optimum if the author can only pick one of the possible
UI devices. What if the author could use a verb coin for the most
common/obvious actions, and a command line (or really-really long menu so it
would work with SR) for everything else? How much would two possible input
methods confuse the user and/or weaken the experience? I don't know. (Mouse
and keyboard (arrows, and ilk) seems to work well, but mouse and typed
command line doesn't - from my experience.)"
Making people switch between two different interfaces is a bad idea from the
start. Some people like to type commands, some people like to click icons,
some people like to do one or the other depending on their mood. I don't know
anyone who likes having to switch back and forth at the whim of the game
author. Providing both text parser and mouse interfaces is a good idea -- I
myself am one of those people who changes preference depending on my mood, so
it'd be great to have a game that gave me both options -- but you need to make
sure that the entire game can be played start to finish using only one of the
interfaces, so that it remains entirely up to the player to choose the input
method. And since you're leaving it up to the player to decide on the input
method, they both need to have the same level of specificity, otherwise the
less specific input method will become like a hint system due to it's ability
to automatically select the correct verb from among several equally logical
candidates.
You might argue that having a "simple" versus "complex" input method allows
players to choose their difficulty level, but that would restrict their choice
of input method if they wanted a particular difficulty. And I can't speak for
anyone else, but I think having an "easy" mode so readily available would be
just little too tempting. Even online hint systems have the overhead of
calling up the menu and searching for the correct hint; getting the game to
solve the puzzle for me just by grabbing the mouse presents no real overhead at
all.
From: Paul Drallos pdra...@tir.com
"There is a big difference between illustrated novels, magazines, etc., and
illustrated IF. In non-interactive illustrated text, the outcome or
advancement
of the plot does not depend on how carefully the text is read or the
illustrations
scanned. In IF, the smallest detail may be critical to the advancement of the
work and missing it may be fatal. In books, the story goes on exactly the same
regardless of whether the text or graphics are carefully examined."
But what about the simple pleasure of exploring the game world? That's what
the graphics and the larger part of the room descriptions are really there for,
after all. If solving the puzzle and advancing the plot was all that mattered,
room descriptions would be nothing but a list of interactive objects. The sole
purpose for flowery prose is to let you experience the game's locales as if
they were real places, and if graphics can add to that experience, why not use
them? If we are to take your logic that anything which distracts the player
from important details is to be eliminated, then we'd better get rid of those
room descriptions as well, and reduce all objects descriptions down to "this is
what this is, this is how you use it, and this is what you do with it."
I can certainly see how graphics could detract from a text game, if the author
had weak art or composition skills. But that's not really any different from
the text-only disasters resulting from people with weak writing skills. Poor
design skills, also: if you have to scan a graphic pixel for pixel, or a text
letter for letter, to find that critical detail, then the game is pretty poorly
designed to begin with.
I can also see how people who simply don't like looking at beautiful things
would dislike graphics, just as much as someone who doesn't like to read would
dislike text, but neither is an inherent property of either medium, nor of
their mixture, just a personal preference of a particular individual. And I
think Mr. Rozak is looking for objective opinions on game design, not
subjective preferences of what kind of game a specific individual personally
prefers.
And "fatal"? We're not talking about Super Mario Bros. here, where a single
misstep forces you to start all over from the beginning. A well-designed IF or
adventure game should never present a situation that can't be gotten out of
without giving clear warning, unless that situation is simply part of the story
and therefore cannot be avoided except by not even playing the game.
To sum up my position on all of this:
I think graphical games have some qualities that text games lack, and text
games have some qualities that graphical games lack; only when the proponents
of each finally stop fighting and work together will The Ultimate Game have a
chance of being realized.
Adrien Beau
>
>> There may be readers of this newsgroup who have the
>> Hitchhiker's game on their "will play that someday" list, who
>> really ought to get the chance to solve it on their own steam.
>
> There is at least one such reader.
There are at least two.
--
spam....@free.fr
You have my name and my hostname: you can mail me.
(Put a period between my first and last names).
Arthur Milliken
> Two words: Verb Coin.
>
> The Verb Coin class of interfaces provides the versatility of a text parser in
> a form that works smoothly with a mouse-driven interface. I personally
> consider it the best possible interface for graphical adventure games.
>
> In fact, I've been thinking about ways to adapt the verb coin interface to text
> IF. It would make a nice complement to the TADS 3 "topic inventory" system,
> eliminating guess-the-verb the same way the topic inventory eliminates
> guess-the-noun.
>
> I believe you mentioned verb coin interfaces yourself in an earlier thread, so
> I shouldn't have to say any more about it.
"Verb Coin"? Is that when a context-sensitive hierarchichal menu
appears whenever you click on an object or character, as in Neverwinter
Nights?
Mike Rozak
> medium. Not only do I know what's there, I know exactly how I'm
> supposed to feel about what's there, how it looks to the protagonist.
> This is something a picture can't do.
Thanks for the info. Good point about emotions and thoughts being implanted
by the author. This is something that words can do fairly easily, but
visuals either can't do or find it very difficult.
--
Mike Rozak
www.mXac.com.au
Mike Rozak
> it turns out it's not "dance", but "waltz" that's the right verb.
> So after a lot of time spent trying to "dance" or "sashay" or
> "rhumba" or "boogie" or "gyrate" or "mamba", the player eventually
> gives up, assuming that the solution must be something other than
> dancing.
Does anyone know if there's a public domain thesaurus available? (I have
found a public domain lexicon = word pronunciations, but haven't found a
thesaurus) A freely available thesaurus would solve some of the
guess-the-verb problems, but not all.
L. Ross Raszewski
<arthur_...@hotmail.com> wrote:
>
>"Verb Coin"? Is that when a context-sensitive hierarchichal menu
>appears whenever you click on an object or character, as in Neverwinter
>Nights?
>
>
While that would probably be covered, the term also applies to a
context insensitive menu of static choices (typically 'examine',
'use', 'talk to') which pops up when an interactive object is clicked
upon (And which, in many games, any particular object responds
meaningfully to no more than two of these, typically only one, but
it's a popular interface because, well, um. Probably because some
LucasArts games used it. Though I imagine they used it better)
Michael Coyne
> Does anyone know if there's a public domain thesaurus available? (I have
> found a public domain lexicon = word pronunciations, but haven't found a
> thesaurus) A freely available thesaurus would solve some of the
> guess-the-verb problems, but not all.
Did you try the rather obvious solution?
You've got to love the Internet. : )
Michael
Gene Wirchenko
>Mike Rozak wrote:
>
>> Does anyone know if there's a public domain thesaurus available? (I have
>> found a public domain lexicon = word pronunciations, but haven't found a
>> thesaurus) A freely available thesaurus would solve some of the
>> guess-the-verb problems, but not all.
>
>Did you try the rather obvious solution?
>
>http://www.thesaurus.com
What is another word for "thesaurus", and would that not be more
appropriate?
>You've got to love the Internet. : )
I do?
Sincerely,
Gene Wirchenko
Computerese Irregular Verb Conjugation:
I have preferences.
You have biases.
He/She has prejudices.
Rick Clark
news:61uDb.54299$aT.2...@news-server.bigpond.net.au...
I think that this is your real answer. The differences between text IF and
graphical IF are much the same as between a novel and a movie, IMO. Each one
is a different experience, and though a movie may be based on a book, it
isn't the book. The same can be said for books based on movies. When you
take the same material and express it in different mediums, you end up with
different products.
Although people love to complain how a particular movie wasn't as good as
the book, they are really comparing apples and oranges. You simply can't
compare a book to a movie, or visa versa. They are totally distinct
products, each with their own formats and requirements. They are truly as
different as apples and oranges.
The reason is very simple. Reading is active and viewing is passive. When
you read a book, you are actually doing something, and this translates into
a certain experience for the reader. When you watch a movie, you are passive
and this creates a completely different experience. One isn't better than
the other, just different, and each has their own rules and methods of
execution.
I think the same could be said for text based IF and graphical based IF. I
have played both and each medium is a different experience, using different
rules and operating in different ways. For me it is very much like the book
and movie. I experience each one differently.
Could you combine the two? You could, but personally, I think you would end
up something much less than the parts. You would have to dilute both aspects
in order to combine them, and loose the impact of both.
--
Rick Clark
http://www.rickclark.org
Seebs
J. Robinson Wheeler <j...@jrwdigitalmedia.com> wrote:
>Yes, you do. It's an especially polite gesture when one is about to reveal
>pretty much the entire solution to an all-time classic puzzle, one which
>is really rewarding to solve on one's own.
Speaking of spoilers:
I once saw a claim that there's a "do this thing early on" puzzle in that
game, where if you don't do something meaningless very early in the game,
you lose later - except I saw a claim that there was an elaborate,
incomprehensible, way around it.
So...
SPOILER WARNING
(rot13)
Vs lbh qba'g tvir gur fnaqjvpu gb gur qbt, vf gurer fbzrguvat ryfr lbh pna
qb gb xrrc vg sebz rngvat gur syrrg? Vs fb, jung vf vg?
-s
--
Copyright 2003, all wrongs reversed. Peter Seebach / se...@plethora.net
http://www.seebs.net/log/ - YA blog. http://www.seebs.net/ - homepage.
C/Unix wizard, pro-commerce radical, spam fighter. Boycott Spamazon!
Consulting, computers, web hosting, and shell access: http://www.plethora.net/
Andrew Plotkin
>
> The reason is very simple. Reading is active and viewing is passive. When
> you read a book, you are actually doing something, and this translates into
> a certain experience for the reader. When you watch a movie, you are passive
> and this creates a completely different experience.
This is, I would say, both simplistic and a lousy parallel for
text/graphical IF. When I play graphical IF I *am* doing something --
far more than when I'm reading, in which my only level of control is
to slow down, speed up, or skip.
Rick Clark
news:bs4avf$46o$1...@reader2.panix.com...
> This is, I would say, both simplistic and a lousy parallel for
> text/graphical IF. When I play graphical IF I *am* doing something --
> far more than when I'm reading, in which my only level of control is
> to slow down, speed up, or skip.
>
> --Z
Simplistic? Hardly. There have been numerous studies that show that reading
engages the brain in an active manner while passive viewing does not. A
simple Google search will bring up numerous cases.
Far more than reading? Again, hardly. Yes, deciding which icon to click does
involve some brain power, but it hardly produces the mental interconnections
that reading produces. How many times will an icon invoke a memory or
emotion that reading sometimes produces? Images will always fail when
compared with imagination, and images do not stimulate the imagination they
way finely crafted prose can do. There is a reason why we are still reading
Dickens and Shakespeare in this modern world.
That being said, my point was that the mediums are different, and invoke
different responses. I don't feel that one is better than the other, each
have their strengths and their weaknesses. The original question was
(paraphrased) how could graphical and text IF exist within the same package?
My view is that in order to do it, you have to compromise, sacrifice some of
the strengths of both, in order to marry the disparate mediums.
This can be done, and done successfully. Comics are a good example. As a
designer you simply have to understand the strengths and limitations of each
medium, understand he impact of the medium on the player and then try to
maximize the strengths toward whatever major goal is trying to be achieved
by the package. It will almost certainly come down to compromises and
sacrifice, but that doesn't mean a good product cannot be produced.
J. Robinson Wheeler
> Speaking of spoilers:
>
> I once saw a claim that there's a "do this thing early on" puzzle in that
> game, where if you don't do something meaningless very early in the game,
> you lose later - except I saw a claim that there was an elaborate,
> incomprehensible, way around it.
>
> So...
>
> SPOILER WARNING
>
> (rot13)
>
> Vs lbh qba'g tvir gur fnaqjvpu gb gur qbt, vf gurer fbzrguvat ryfr lbh pna
> qb gb xrrc vg sebz rngvat gur syrrg? Vs fb, jung vf vg?
>
Hfvat gur vzcebonovyvgl qevir yngre va gur tnzr, lbh nccrne nf
qvssrerag punenpgref ng enaqbz, naq va bar bs gurfr fpranevbf,
lbh ner Sbeq sebz gur gvzr bs gur cebybthr. Nf Sbeq, lbh pna
znxr fher gb tvir gur purrfr fnaqjvpu gb gur qbt.
Gung'f jung V erzrzore, naljnl.
mad...@bmrb.wisc-antispam.edu
What I described amounts to just plain synonyms. But what *you*
described, didn't. There were seperate messages for different types
of dance.
:
:
:> (Each and every puzzle would have to include each and every possible
The point and click wasn't the point. The point was that they *reduce*
rather than increase the command set and that removes the frustrating
tedious typing part of solving problems while retaining the thinking part.
The same effect could be had by reducing the verbs to a generic
"use X, Y" that would almost always act as an appropriate
synonym for other things (like "shoot X at Y" or "attach X to Y").
You could still do that and keep text descriptions.
:
:
:<snip>
:
:
:> :> It turns out that it's a container, and the ONLY way to discover
:> :> this is to type the command to use it like a container: "put foo in
:> :> thing", which you won't do if you don't realize it's a container.
:> :
:> :
:> :Well, since you didn't know what it is, surely you tried manipulating it
:> :in a variety of ways, right?
:>
:> If I took that approach with every item that is in the game, I'd still
:> be playing it 15 years later. It's unreasonable to expect people to
:> try every action in the game with every object until they find the
:> combinations that work.
:
:
:I just meant that the item's description actually said you didn't know what
:it was, which should have encouraged you to play with it in a multitude of
:ways.
No, actually it encouraged me to say, "ahh, this is an item Douglas
Adams threw in here for it's silliness rather than it's usefulness."
(Not an unreasonable thing to assume, since there *were* tools strewn
about that you didn't know what *they* were either, and some of *them*
never did anything.)
: And remember this is Douglas Adams we're talking about here. How much
Quintin Stone
> Seebs wrote:
>
> > Speaking of spoilers:
> >
> > I once saw a claim that there's a "do this thing early on" puzzle in that
> > game, where if you don't do something meaningless very early in the game,
> > you lose later - except I saw a claim that there was an elaborate,
> > incomprehensible, way around it.
> >
> > So...
> >
> > SPOILER WARNING
> >
> > (rot13)
> >
> > Vs lbh qba'g tvir gur fnaqjvpu gb gur qbt, vf gurer fbzrguvat ryfr lbh pna
> > qb gb xrrc vg sebz rngvat gur syrrg? Vs fb, jung vf vg?
> >
>
> Hfvat gur vzcebonovyvgl qevir yngre va gur tnzr, lbh nccrne nf
> qvssrerag punenpgref ng enaqbz, naq va bar bs gurfr fpranevbf,
> lbh ner Sbeq sebz gur gvzr bs gur cebybthr. Nf Sbeq, lbh pna
> znxr fher gb tvir gur purrfr fnaqjvpu gb gur qbt.
>
> Gung'f jung V erzrzore, naljnl.
Gur svefg fprar vaibyivat gur jneevat syrrgf cbcf hc enaqbzyl fbzr gvzr
nsgre lbhe svefg glcb/aba-erpbtavmrq pbzznaq. Vg'f lbhe "vafhyg" gung
frgf gur gjb syrrgf svtugvat, naq gura va frnepu bs Rnegu.
Jung unccraf vs lbh arire znxr n zvfgnxr?
/====================================================================\
|| Quintin Stone O- > "You speak of necessary evil? One ||
|| Code Monkey < of those necessities is that if ||
|| Rebel Programmers Society > innocents must suffer, the guilty must ||
|| st...@rps.net < suffer more." -- Mackenzie Calhoun ||
|| http://www.rps.net/QS/ > "Once Burned" by Peter David ||
\====================================================================/
DI Gunther Schmidl
>
> Jung unccraf vs lbh arire znxr n zvfgnxr?
>
Gura vg whfg pubbfrf fbzr aba-zvffcryyrq jbeq.
-- Gunther
Michael Bechard
> I have argued in the past that the essence of IF lies in *not* giving
> a menu-list of actions, but instead giving a free range of action
> which the player has room to explore. Whenever I encounter a menu
> interface -- and this includes dialogue menus in text games as well as
> verb pop-up menus -- I find that my sense of the game-world
> immediately vanishes. I have a mechanical strategy available to me
> ("try everything on everything"), and it's easier to follow that than
> to visualize myself as a *person* in a *situation*.
I'm going to play devil's advocate here and make a point opposite of
what I made earlier in this thread. As Mike Roberts pointed out, the
range of actions in a text-adventure game isn't really free. The game's
parser is setup to only accept commands in a semi-grammatical format,
and the verbs that an author chooses to implement are rather limited,
given the full range of the English language. This, compounded with the
"standards" of the IF commandset given by most parsers that IF veterans
are quite familiar with (x, up, north, take, etc.) makes for a
more-or-less spelled-out interface; the trick for newbies is to get
acquainted with it. It seems to me that the immersion and
free-formedness that most IF players feel is really an illusion.
>
> I decided that the reason I like Myst-style graphical IF is that it
> *acknowledges* that selecting an action for an object is
> uninteresting. In this sort of game, there's only one action -- click,
> which means "use this".
<snip>
I guess, as a non-Myst fan, this is one thing I have never understood
about those games and why people prefer them. You have one possible
course of action for everything; click. It's anybody's guess as to what
"click" exactly does, but that's all you get. Granted, for most
situations it's easy to deduce what clicking on an object will do (click
on a door and it opens, etc.), but it still seems exceedingly bland to
me. What if IF used this approach? Just "use" everything. Neato.
As in IF, having a set of verbs to use *increases* my sense of
interaction and also serves to further pull me into the game. Not only
do I know that I'm opening the door, I'm *choosing* to open the door.
And what if I want to listen to the door? To knock on it? To force it
open by pushing on it or attacking it? Granted, those verbs must be
implemented in the game in order to be an option, but it's possible.
Objects in Myst, to me, have a very one-dimentional aspect to them,
since I can't do any of those things to the door.
My biggest gripe with Myst style games is the lack of *any* NPC
interaction, but that's another conversation...
Michael
Mike Rozak
an IF authoring system to automagically build some of the thesaurus for the
author and/or user. If a freely distributable dictionary (with work
definitions) were availalble then a word-net could be constructed that could
further remove some of the guess-the-verb.
You'd never eliminate all the guess the verb problems though using a
thesaurus or word-net. Another approach is to get 100+ beta testers and scan
their play logs looking for unexpected (but correct) commands.
--
Mike Rozak
www.mXac.com.au
"Michael Coyne" <coy...@mtsDOS.net> wrote in message
news:Tc7Fb.8781$Tb7....@news1.mts.net...
Mike Rozak
Yes and no. Comics are a good example because they include graphics and
text, but bad because on a computer you could have graphics and recorded
speech. I think TV documentaries (or newscasts) might be a better example of
the ways to combine them. (Although they're not fiction. Cross comics with
documentary?)
> Although people love to complain how a particular movie wasn't as good as
> the book, they are really comparing apples and oranges. You simply can't
> compare a book to a movie, or visa versa. They are totally distinct
> products, each with their own formats and requirements. They are truly as
> different as apples and oranges.
> The reason is very simple. Reading is active and viewing is passive. When
> you read a book, you are actually doing something, and this translates
into
> a certain experience for the reader. When you watch a movie, you are
passive
> and this creates a completely different experience. One isn't better than
> the other, just different, and each has their own rules and methods of
> execution.
> Far more than reading? Again, hardly. Yes, deciding which icon to click
does
> involve some brain power, but it hardly produces the mental
interconnections
> that reading produces.
I agree that it's apple and oranges. I don't necessarily agree about passive
and active. (Although I suspect TV broadcasts are designed to be passive and
inane, especially morning news shows.) Maybe it's a left brain vs. right
brain issue? (To pull out a cliche.)
> Could you combine the two? You could, but personally, I think you would
end
> up something much less than the parts. You would have to dilute both
aspects
> in order to combine them, and loose the impact of both.
To use the fruit analogy: Some people like cran-apple drink. Personally, I
like orange-mango and orange-tangerine, but not orange-apple. Seriosly, I am
worried about the possibility that combining oranges and apples results in
bad combination though. I keep doing mind experiments, starting with pure
text IF and then adding only a few images/sounds and seeing how well that
works, or conversely, starting with a first-person shooter UI and adding IF
components.
I know that when text and graphics were combined in the late 1980's the
experience did NOT work. The apple II graphics adventures were fairly awful
because of the poor graphics, only 4 lines of text, and the difficulty of
reading the text and trying to look a the pitcutres. I only played the early
king's quest for a few minutes, and it didn't appeal to me either.
Nor does text superimposed on graphics (such as all the diaries in Myst
I-IV) work. Not only are they difficult to read, the text diaries feel
completely out of place in the graphics world.
Neither tried recorded speech. Lucas arts uses recorded speech, combined
with graphics. I am ambilent towards them. The speech and graphics seems to
work OK, but they don't use a narrator (except for occasional comments by
the character) so some concepts (as pointed out in other posts) cannot
easily be commuinicated. Plus, their context-sensative command menus, while
making life easier, reduce some of the open-endedness of the world.
--
Mike Rozak
www.mXac.com.au
Papillon
As I often ramble in this sort of discussion...
I still fail to understand why the best hybrid of graphical adventure games
and IF (best, at least, in as far as combining the most elements of both)
seems to be not very popular.
That would be the Sierra engine right before it gave up the text parser -
when mouse clicking was starting to be meaningful, but not yet the only
method of interaction.
You had the graphical display of what was going on, the visualisation of the
scene.
You had, thankfully, a game that would PAUSE when you were typing something,
so you couldn't be killed because you couldn't finish your command fast
enough. :)
You had the fairly-necessary addition of mouse-look, so that if you failed
to understand what a pixelised-graphical-object WAS, you could always look
at it, figure out what it was called and get a general idea of whether or
not it was important.
And you had the open-endedness of the text parser, allowing you to pursue
topics/information at your own whim instead of, say, clicking through a
dialog menu (given a list of topics to talk about, it seems quite obvious
that the correct approach is to click ALL of them, making it dull)
---
Hanako Games
http://www.hanakogames.com/
Mike Rozak
> free-formedness that most IF players feel is really an illusion.
One nice feature about a parser (even if it is an illusion) is that the
player doesn't know where the limits of its understanding lie. (This is also
the frustrating part.) The inability to easily perceive the limits creates
the illusion that anything can be done. It's like people that live on a 1/2
acre block and plant thick, bushy trees around the edge of their lot so they
can't see their neighbors or the road. It gives the illusion that the house
is isolated in a forest, even though it isn't. Personally, I prefer the
illusion of a huge forest surrounding my house. Facing the reality of living
in a suburb is too much. (I have since moved into a huge forest/savanah, and
must now face the reality that I can't get pizza delivery.)
The point-and-click interface completely destroys this illusion since it's
obvious where the limits are. (But it also removed the guess-the-verb
frustration.)
> My biggest gripe with Myst style games is the lack of *any* NPC
> interaction, but that's another conversation...
Agree. They had to do this because pointing and clicking on people is
considered rude. Or maybe because the work of creating a conversation tree,
recording all the speech, and then trying to animate/video it is too much
work. (I have to admit, the world they created has a clever excuse for no
NPCs, but you can only wander around so many deserted landscapes.) The
conversation trees in all the graphical adventures that I've played have
been sparse because of the costs.
--
Mike Rozak
www.mXac.com.au
Rick Clark
> I know that when text and graphics were combined in the late 1980's the
> experience did NOT work. The apple II graphics adventures were fairly
awful
> because of the poor graphics, only 4 lines of text, and the difficulty of
> reading the text and trying to look a the pitcutres. I only played the
early
> king's quest for a few minutes, and it didn't appeal to me either.
It seems to me that the problem lies in the transition between the two
medias when they are combined, not necessarily the media format themselves.
Take a movie that has sub-titles, for example. Switch from watching the
movie to reading the subtitles, for me at least, diminishes the movie
experience. I cannot get immersed into the movie because I have to mentally
stop, read the subtitle and then try to relate that to what I am seeing on
the screen.
For me, this combination process of viewing/reading applies to games as
well. When I first played Doom, I was so immersed into the game, that I
actually jumped when my wife entered the computer room. If, while playing
the game, I had to stop and read a bunch of text info, it would have ruined
the experience.
To bring it closer to IF, I have read some of the TADS stories, both with
and without the graphics. I found myself much more absorbed in the text-only
versions, rather than the graphic versions, because I didn't have to break
out of reading mode, look at a picture, and then get back into reading mode.
The graphic versions just ended up being annoying (while still being good
stories) because I had to switch my brain between the two modes. For me, the
problem is the transition between the two modes that draw me out of the
story and remind me that I am indeed just playing a game.
Rick Clark
> You'd never eliminate all the guess the verb problems though using a
> thesaurus or word-net. Another approach is to get 100+ beta testers and
scan
> their play logs looking for unexpected (but correct) commands.
There has been some success in this problem by using AI techniques such as
neural nets or genetic algorithms (GA)/classifier systems(CS). A GA system
might be a better approach since it lends itself to wildcards within the
coded gene that greatly reduce the number of rules that must be kept within
the system. By bulding a GA/CS it could automatically adjust itself based on
the user input (assuming you have a text interface) and handle unexpected
situations (once the system has generated enough rules) that would be
impossible to hard-code.
John W. Kennedy
> Seebs wrote:
>
>
>>Speaking of spoilers:
>>
>>I once saw a claim that there's a "do this thing early on" puzzle in that
>>game, where if you don't do something meaningless very early in the game,
>>you lose later - except I saw a claim that there was an elaborate,
>>incomprehensible, way around it.
>>
>>So...
>>
>>SPOILER WARNING
>>
>>(rot13)
>>
>>Vs lbh qba'g tvir gur fnaqjvpu gb gur qbt, vf gurer fbzrguvat ryfr lbh pna
>>qb gb xrrc vg sebz rngvat gur syrrg? Vs fb, jung vf vg?
>>
>
>
> Hfvat gur vzcebonovyvgl qevir yngre va gur tnzr, lbh nccrne nf
> qvssrerag punenpgref ng enaqbz, naq va bar bs gurfr fpranevbf,
> lbh ner Sbeq sebz gur gvzr bs gur cebybthr. Nf Sbeq, lbh pna
> znxr fher gb tvir gur purrfr fnaqjvpu gb gur qbt.
>
> Gung'f jung V erzrzore, naljnl.
Actually, it isn't random.
--
John W. Kennedy
"But now is a new thing which is very old--
that the rich make themselves richer and not poorer,
which is the true Gospel, for the poor's sake."
-- Charles Williams. "Judgement at Chelmsford"
Andrew Plotkin
> > I decided that the reason I like Myst-style graphical IF is that it
> > *acknowledges* that selecting an action for an object is
> > uninteresting. In this sort of game, there's only one action -- click,
> > which means "use this".
> <snip>
>
> I guess, as a non-Myst fan, this is one thing I have never understood
> about those games and why people prefer them. You have one possible
> course of action for everything; click. It's anybody's guess as to what
> "click" exactly does, but that's all you get. Granted, for most
> situations it's easy to deduce what clicking on an object will do (click
> on a door and it opens, etc.), but it still seems exceedingly bland to
> me. What if IF used this approach? Just "use" everything. Neato.
You snipped the completion of my point, which is that these games
give you a *different* range of action, to replace the missing one of
the verb choice.
In graphical games, this range is a cluttered and interesting visual
world, which must be parsed and understood. Much more so than the
prose of a text game.
I'm not saying that reading prose is *not* mind-engaging work. But
it's to a different end. For the purpose of understanding what's going
on around you, prose is pre-analyzed -- already in the language of
thought, in a sense. (Adam Cadre posted an interesting and related
point on his web site -- http://adamcadre.ac/calendar/ if you don't
follow it, the recent entries on autism.)
(Yes, I'm leaving aside here the sub-group of text IF which *is*
obscure and hard to read, right down in the prose of room descriptions
and such. People do this, but it's not common, any more than it is in
novels.)
Seebs
Rick Clark <rickc...@yahoo.com> wrote:
>Take a movie that has sub-titles, for example. Switch from watching the
>movie to reading the subtitles, for me at least, diminishes the movie
>experience. I cannot get immersed into the movie because I have to mentally
>stop, read the subtitle and then try to relate that to what I am seeing on
>the screen.
Interesting. I don't have that. I have the problem that I cannot control my
reading behavior; if words pass in front of my eyes, they are fed to me by the
part of my brain which reads. I'm one of the people who needs about a
paragraph of text in front of a spoiler to give my conscious brain enough
time to stop reading before the text spooler gets to the spoiler.
mad...@bmrb.wisc-antispam.edu
:> The reason is very simple. Reading is active and viewing is passive. When
:> you read a book, you are actually doing something, and this translates
:into
:> a certain experience for the reader. When you watch a movie, you are
:passive
:> and this creates a completely different experience. One isn't better than
:> the other, just different, and each has their own rules and methods of
:> execution.
:
:> Far more than reading? Again, hardly. Yes, deciding which icon to click
:does
:> involve some brain power, but it hardly produces the mental
:interconnections
:> that reading produces.
:
:I agree that it's apple and oranges. I don't necessarily agree about passive
:and active. (Although I suspect TV broadcasts are designed to be passive and
:inane, especially morning news shows.) Maybe it's a left brain vs. right
:brain issue? (To pull out a cliche.)
I think the problem is far more that you are trying to cram two different
streams of communication down one single sense that must be focused - sight.
That one is active and one passive isn't the problem. The problem is that
both require you to focus your eyes, and doing so is a mutually exclusive
process. You can't be focused on two different parts of the screen at the
same time. At any one instant of time your eyes can be pointed either at
the text or at the picture, but not both. So you have to flip back and
forth between them. If the narration is audio rather than textual, then
the problem goes away because you are now using a different sense for each
input stream, and the physical sensing apparatus CAN do both at once. It
has nothing to do with what is happening in your mind, and everything to do
with what is happening in the physical sensing devices (ears, eyes). They
measure different media, and so can each simultaneously feed it's own
information to your brain. Your brain is actaully VERY good at seperating
the two streams in parallel *if* your body can actually pick them up in
the first place. But, your eyes cannot aim in two different directions at
once and form a coherent picture.
Adrien Beau
>
> Interesting. I don't have that. I have the problem that I
> cannot control my reading behavior; if words pass in front of my
> eyes, they are fed to me by the part of my brain which reads.
I would have said exactly the same a few weeks ago, but since then
I've noticed two things.
First, this depends for me on the kind of subtitles: I read
laser-cut movie subtitles effortlessly, but DVD or fansub
subtitles are a bit more challenging. This is probably a matter
of habit and of subtitle quality.
Second, if you throw more than image+subtitle to my brain, I have
to really concentrate to digest all that is going on. I saw a
Japanese movie two weeks ago, which was laser-subtitled in
English and video-projector subtitled in French. Since I
understand both languages, my brain was trying to read both of
them (plus the images), and that was a hard work I was very
conscious of.
--
spam....@free.fr
You have my name and my hostname: you can mail me.
(Put a period between my first and last names).
Graham Holden
<papillo...@bigfoot.com> wrote:
>
>As I often ramble in this sort of discussion...
>
>
>I still fail to understand why the best hybrid of graphical adventure games
>and IF (best, at least, in as far as combining the most elements of both)
>seems to be not very popular.
>
>That would be the Sierra engine right before it gave up the text parser -
>when mouse clicking was starting to be meaningful, but not yet the only
>method of interaction.
<snip good bits about early Sierra games>
I too think these were some of the best combination of text and
graphics; both the text and graphics seemed to enhance the
game-playing: Sometimes the text gave clues/hints/information that
wasn't in the graphics; sometimes the graphics helped to clarify
something that might have been a little unclear with text alone. From
my memories (admittedly not perfect), there was very little detriment
or feelings of "kludging" them together...
Even the first game or so in each series that went fully mouse-driven
wasn't too bad; although the downward spiral had started. Their later
efforts, which I almost totally avoided, in comparison seemed little
more than watching a video that would randomly pause, and you had to
spend half an hour looking for the hidden "play" button.
I strongly suspect that the reason why the good-games died out is NOT
that they weren't popular, but because of the all too common "we must
be as 'flashy' as possible" syndrome... once it became possible to
have super-slick-looking graphics, digitised from actors' movements,
etc. etc., the "powers that be" insisted that they be used, and
gameplay just seemed to crawl into a corner to curl up and die...
</moan>
--
Regards,
Graham Holden
g DASH holden AT dircon DOT co DOT uk
(to reply by email, replace DASH, AT and DOT as appropriate).
There are 10 types of people in the world;
those that understand binary and those that don't.
Michael Bechard
I see; the range of graphical objects depicted replaces the range of
verbs. It still seems to be a bit lacking, since you could have so much
more interactivity with all those juicy graphics if you had a range of
verbs to use, but I see your point.
Interesting point, too, about IF prose being pre-analyzed in the context
of its own grammar. However, graphic adventure games only gain an
advantage there if they ARE Myst games, a.k.a. what the hell do I do
with all these objects. Less gadgety games like the Lucas Arts games are
more akin to IF in the respect of the contexts in which the graphics are
represented. It's more familiar.
Michael
Rick Clark
> > You'd never eliminate all the guess the verb problems though using a
> > thesaurus or word-net. Another approach is to get 100+ beta testers and
> scan
> > their play logs looking for unexpected (but correct) commands.
"Rick Clark" <rickc...@yahoo.com> wrote:
> There has been some success in this problem by using AI techniques such as
> neural nets or genetic algorithms (GA)/classifier systems(CS).
Just for grins I worked up a small bibliography of the GA work in this area:
Genetic Algorithms and Natural Language (pdf. This is called the LUST
system. :))
http://www.ils.unc.edu/~losee/gene1.pdf
Learning Algorithm for Keyphrase Extraction (Article Abstract Page, links to
pdf)
http://cogprints.ecs.soton.ac.uk/archive/00001797/
GA in Hyperlink Clustering (similar domain problem)
http://www.austlii.edu.au/au/other/CompLRes/1997/15/
Searching for Features (Word Types) Using a GA (pdf) (Includes some
pseudo-code examples)
http://ling.uib.no/BREDT/publikationer/joh_scai2.pdf
Who knows if any of it could really be implemented successfully in an IF
system, but it at least offers some interesting ideas about the
player/computer interface.
Rick Clark
Daniel Barkalow
> On Mardi 23 Décembre 2003 10:45, Seebs wrote:
> >
> > Interesting. I don't have that. I have the problem that I
> > cannot control my reading behavior; if words pass in front of my
> > eyes, they are fed to me by the part of my brain which reads.
>
> I would have said exactly the same a few weeks ago, but since then
> I've noticed two things.
>
> First, this depends for me on the kind of subtitles: I read
> laser-cut movie subtitles effortlessly, but DVD or fansub
> subtitles are a bit more challenging. This is probably a matter
> of habit and of subtitle quality.
Just because something is automatic and unconscious doesn't mean it
doesn't require mental effort. It's just that you can't avoid expending
the mental effort and are not directly aware of it. There is a lot more
mental effort involved in reading some text than others, up to the point
where your brain fails to do it (I remember some Russian movies with
horrendous subtitles; I could only understand the dialogue when the scene
was dark).
> Second, if you throw more than image+subtitle to my brain, I have
> to really concentrate to digest all that is going on. I saw a
> Japanese movie two weeks ago, which was laser-subtitled in
> English and video-projector subtitled in French. Since I
> understand both languages, my brain was trying to read both of
> them (plus the images), and that was a hard work I was very
> conscious of.
There are also some mental resources used in understanding language which
aren't more generally used. This means that adding the first set of
subtitles partially demands resources which would otherwise be
idle. Adding the second set is pure extra work, and is also more likely to
give the conscious perception of effort.
-Iabervon
*This .sig unintentionally changed*
Cedric Knight
> Actually, I mean public domain as in freely distributable. This would
> allow an IF authoring system to automagically build some of the
> thesaurus for the author and/or user. If a freely distributable
> dictionary (with work definitions) were availalble then a word-net
> could be constructed that could further remove some of the
> guess-the-verb.
WordNet (from Princeton Uni) is probably what you're after - it's
arranged hierarchically, so you can find similar, but not synonymous
words. However, it's about 20MB so more a compiler add-on than an
interpreter add-on, I'd think. It is copyright and does have a licence,
but it's very loose freeware, which is why online dictionary sites use
it.
"armoire" turns up wardrobe, closet, clothespress, but not "dresser"
Also the 1911 US Edition of Roget's is out-of-copyright and available
for download. I've parsed it to make it more easily machine-readable.
Email me (without the spamblock) if you want a copy. Or maybe I'll put
a verb database on my site sometime (after Christmas).
Actually, I think what'd happen in your proposed scenario is that
players would soon adapt by not phraseing their commands so precisely,
and would be annoyed if the IF work occasionally didn't intuit what they
wanted to do, or did so wrongly. However, we wouldn't know this till we
try it.
CK
Mike Rozak
One of the papers discusses POS tagging: I already have a POS tagger that I
wrote for text-to-speech. It uses tri-grams, and gets 80%-90% accuracy. I
have a paper that discusses a system for more accurate POS detection (97%
accurate), BUT it requires a tagged corpus of about 60,000 words. The
tri-gram requires user input on only a few hundred sentences (plus the POS
possibilities in the lexicon). I chose the tri-gram because it makes
localization easier, and 80-90% accracuy is fine for TTS. I don't do a parse
tree, which would be the next step. Parse trees are more difficult to
localize to other languages.
I don't know how well POS tagging and parse trees would be for IF because 1)
the sentences are fairly simple, making the parse tree obvious, and 2)
people don't type in complete sentences anyway. Parse trees don't work well
for SR dictation because people don't always use gramatically correct
sentences when they dictate either.
Keyphrase extraction - This might be interesting as a way of generating
conversations. It wouldnt do much for commands parsing. You could impliment
command parsing using some of the algorithms used in SR (dictation). You
could also use statistics to determine which words are most common, such as
"the" and "it" and lower their importance for phrase. Hence, if only one
command example used the word "innoculate", then if the user typed in
"innoculate it", the "innoculate" would have a higher importance than "it".
I have started to design my IF OOP system, and have put in a few
modifications so that machine learning could be added to the command and
conversation systems.
--
Mike Rozak
www.mXac.com.au
Mike Rozak
Thanks for that link. I saw Microsoft's equivalent, but didn't know
Princeton had one that was easily distributable. I'll have to look at the
code to see how it's hooked up, but if/when I get around to it, something
like WordNet could work well. There is one major problem though... it's only
for English.
As far as 20MB... I view things somewhat differently than most people in IF.
If I undertake a graphical IF authoring kit it will target a 10 gHz machine
with 1 gB RAM and a 100MB-200MB IF content download. A high end consumer
machine today is 3.5 gHz with 512 MB RAM and broadband, so a factor of 3x in
speed, and 2x in memory is only 2 years away. (I wouldn't have anything
worth using before then.) It's lifespan would be another 5-10 years, by
which point another 10x-20x improvements in computers is possible. (Unless
Moore's law breaks, which it could.) Plus, to do the graphics right, you
need a fast machine and a large download for the textures. Audio also
requires a large file.
> Actually, I think what'd happen in your proposed scenario is that
> players would soon adapt by not phraseing their commands so precisely,
> and would be annoyed if the IF work occasionally didn't intuit what they
> wanted to do, or did so wrongly. However, we wouldn't know this till we
> try it.
Of course they would. Give 'em an inch and they'll take a mile. But, I
suspect the inch is better than nothing.
> Also the 1911 US Edition of Roget's is out-of-copyright and available
> for download. I've parsed it to make it more easily machine-readable.
Thanks for this link. I just found it on project gutenberg.
--
Mike Rozak
www.mXac.com.au
someone
On Wed, 24 Dec 2003 22:11:23 GMT, "Mike Rozak" <Mike...@bigpond.com>
posted:
>> WordNet (from Princeton Uni) is probably what you're after - it's
>
>Thanks for that link. I saw Microsoft's equivalent, but didn't know
>Princeton had one that was easily distributable. I'll have to look at the
>code to see how it's hooked up, but if/when I get around to it, something
>like WordNet could work well. There is one major problem though... it's only
>for English.
>
>As far as 20MB... I view things somewhat differently than most people in IF.
>If I undertake a graphical IF authoring kit it will target a 10 gHz machine
>with 1 gB RAM and a 100MB-200MB IF content download. A high end consumer
>machine today is 3.5 gHz with 512 MB RAM and broadband, so a factor of 3x in
>speed, and 2x in memory is only 2 years away. (I wouldn't have anything
>worth using before then.) It's lifespan would be another 5-10 years, by
>which point another 10x-20x improvements in computers is possible. (Unless
>Moore's law breaks, which it could.) Plus, to do the graphics right, you
>need a fast machine and a large download for the textures. Audio also
>requires a large file.
who will host it? the bandwidth cost will be substantial.
Mike Rozak
1) Moore's law also affects bandwidth costs.
2) A sub-set of the full IF (basically the first "chapter") could be put up
for download. If users get through that they can download the rest. This
would ensure that people browsing only download the smallest chunk.
3) A sub-set is downloaded, and the rest is delivered via a CD-Rom. There
are companies that automatically duplicate and mail CD-ROMs for about $5.
(They also do all the credit card charges)
3a) A few dozen IF titles are packaged onto a DVD and similarly distributed.
4) The author makes sure not to many texture maps or sounds.
Oh, and I forgot to mention in my previous mail: The 20 mb for the wordnet
(if it were used) wouldn't need to be shipped in the IF title since only the
word-net portions necessary for the IF would be needed.
--
Mike Rozak
www.mXac.com.au
"someone" <eri...@hotmail.com> wrote in message
news:3fea50bc...@news.singnet.com.sg...
Stas Starkov
>
> As I often ramble in this sort of discussion...
>
>
> I still fail to understand why the best hybrid of graphical adventure
> games
> and IF (best, at least, in as far as combining the most elements of both)
> seems to be not very popular.
>
> That would be the Sierra engine right before it gave up the text parser -
> when mouse clicking was starting to be meaningful, but not yet the only
> method of interaction.
I've played some of the old "Sierra"'s games (several years ago), and I
wasn't impressed. The reason is not a bad parser or interface, but the
plot -- it was not very original, illogic and simplistic. Can you name a
"Sierra"'s game that is quite original and gives impression that the
game is not a buffoonery? -- maybe I missed one?
> You had the fairly-necessary addition of mouse-look, so that if you failed
> to understand what a pixelised-graphical-object WAS, you could always look
> at it, figure out what it was called and get a general idea of whether or
> not it was important.
In really early "Sierra"'s games one have to type "examine toad" to get
the description of the object. But... one should have been aware, that
the small green block of pixels is called "toad".
Next generation games of the software giant haven't had a
"mouse-over-object" behavior -- one have to chose verb "examine" from
the game menu and then click on the suspicious object just to understand
what's that. Often I was quite annoyed doing that over and over again
(especially in time constraints).
Nah, "Sierra" is not my favorite graphic adventure game maker.
Mike Roberts
> Mike Roberts <mjrUND...@hotmail.com>:
> > The only reason the two cases [text vs graphical UI] *seem*
> > different is that the graphical game tells you up front what control
> > inputs it accepts, while the text game leaves it up to you to figure
> > out what inputs are valid.
>
> There is a point you overlook. When you are told the all the
> commands it can be considered spoilers. What if the soulution to a
> puzzle is to dance? [...] If you where told upfront that it is possible
> to dance, it would be a very easy puzzle.
That's not really relevant to the point I was making, which was that nothing
in a graphical UI prevents you from making up your own commands that don't
work, just as in a text UI. For your specific example, consider:
- In Zork 1 (a text game), you, the player, can decide that typing >DANCE
means "the player character dances." Zork 1 will respond: "I don't know the
word 'dance'."
- In Day of the Tentacle (a graphical game), you, the player, can decide
that wiggling the joystick back and forth in time with the music means "the
player character dances." DotT will respond by moving the player character
back and forth slightly, and nothing more.
You can make up your own invalid commands in either case, and the net
effects are the same. The only difference is that you know in advance they
won't work in a graphical game, because the instructions tell you everything
that does work; whereas in a text game, you're often actively invited to try
anything you like, because the instructions suggest or say outright that
they *don't* tell you everything that does work. *That* was my point.
You're talking about something separate from what I was commenting on,
although it's certainly relevant to the poster's original survey question
(probably moreso than my nitpicking technicality). You're talking about the
way a text game designer can exploit the player's lack of knowledge of the
set of valid inputs to create hidden actions that the player has to think
of - from hints, intuition, or just plain guesswork. Many text IF fans
clearly see as a virtue; it can occasionally lead to satisfying puzzles in
which some non-obvious action comes to you in a flash of inspiration and
just works.
For what it's worth, though, the exact same thing is seen as a huge, even
fatal flaw by a lot of non-fans, and would certainly be a top answer to the
complementary survey question, Why Do So Few People Like Text IF? The
existence of hidden actions leads much more often to guess-the-verb and
guess-the-action puzzles than to satisfying flash-of-inspiration puzzles. A
lot of non-fans see this sort of thing as an unfair, one-sided guessing
game, where the player is obliged to make random guesses from the whole
universe of real-world actions, while the author just sits back, smirks, and
doles out such unhelpful hints as "I don't know that word" and "You can't do
that... yet!" and "How bizarre!"
My own personal feeling, having fooled around with these games for quite a
long time now, is that the whole hidden-action angle is a novelty that can
be amusing the first couple of times you see it, but that it quickly wears
thin. I'm continually surprised at how much weight people give it when
ranking the reasons they like text games, but there's no question that it's
a big part of the appeal for a lot of people.
--Mike
mjr underscore at hotmail dot com
Mike Roberts
> JAN THORSBY <jtho...@c2i.net> wrote:
> :There is a point you overlook. When you are told the all the
> :commands it can be considered spoilers. What if the soulution
> :to a puzzle is to dance?
>
> know what they need to do, but the game doesn't recognize it so
> they assume it was wrong when it was actually on the right track.
That's essentially the definition of a guess-the-verb puzzle, which everyone
agrees is bad. This sort of thing is essentially a bug that can be usually
be fixed with enough play-testing, though.
To me, the more insidious problem with hidden actions is that any situation
will have innumerable logical actions that *aren't* on the right track.
Okay, so the solution to the present puzzle is that we have to DANCE. But
in the same situation, there could be a hundred or a thousand perfectly
sensible real-world alternatives that *won't* be accepted by the game - not
because the parser doesn't know enough synonyms for verbs, but because the
game's world model doesn't have enough depth of implementation to support
the actions at all. If dancing doesn't occur to you as the solution, you'll
just get a blanket rejection ("I don't know that word") for the other
logical-but-wrong things you try. To the novice player, certainly, this
could be reasonably taken as a meta-hint that a hidden action isn't what's
needed - you could read it as the game telling you to stick to the
"standard" actions listed in the instructions. Or, as a hint that you're
playing a random guessing game:
Pick a number, any number.
>17
Wrong! Pick again.
>123
Oh, so close! Try again.
>124
Not even close! Pick another number.
>198930809
Getting warmer! Care for another try?
>198930810
Oops! Try again.
>18400
Bzzt! Thanks for playing! Pick again.
In either case, there's no way to correct this through playtesting; no
matter how hard you try as an author, you simply can't anticipate every
possible action a player might attempt, because the set of possible actions
is literally infinite. This is what I don't like about coyly inviting the
player to try anything they can think of - *most* of what they can think of
will get a blanket rejection, which is no fun for the player. Even if the
instructions explicitly encourage random exploration of the verb space, the
game will de facto discourage it by responding with blanket rejection to
most attempts at exploration.
Mike Roberts
> I have argued in the past that the essence of IF lies in *not*
> giving a menu-list of actions, but instead giving a free range of
> action which the player has room to explore. Whenever I
> in text games as well as verb pop-up menus -- I find that my
> sense of the game-world immediately vanishes.
I tend to agree about this psychological effect of menus, but...
> I have a mechanical strategy available to me ("try everything
> on everything"), and it's easier to follow that than to visualize
> myself as a *person* in a *situation*.
...I don't think this is true in general. An obvious counter-example is
Chess: at any given time, all of the available moves could be enumerated in
a menu, but exhaustive search of the available moves is not a viable
strategy. I don't mean to hold Chess up as a paradigm for IF, since it has
the central element of symmetrical competition that's absent in IF, but it's
an example of how a constrained range of options can combine to create an
effectively infinite range of play.
A more IF-y example would be a cipher machine that has, say, two inputs
(representing bit values 0 and 1), and does something to each bit according
to some secret internal rotor that changes upon entering each bit; the
"menu" of actions would amount to PUSH 0 BUTTON, PUSH 1 BUTTON, and STOP
PLAYING WITH MACHINE, but it would be next to trivial to design a puzzle
along these lines that isn't soluble by a brute-force attack. I also don't
hold this up as a particularly interesting kind of IF puzzle, at least when
presented so transparently as a cipher machine, but I do think the principle
is important: specifically that interesting interactivity can come from the
range of the game world's internal state rather than from the range of
possible inputs.
> I decided that the reason I like Myst-style graphical IF is that
> it *acknowledges* that selecting an action for an object is
> uninteresting. In this sort of game, there's only one action --
> click, which means "use this". The explorable range of action
> becomes *the visual world which is presented to you*, which
> you can explore to discover which objects are reactive and
> interesting.
A couple of people have posted replies with the inference that you mean the
illusion of freedom from hidden verbs in text IF maps to the illusion of
freedom from hidden pixels in graphical IF. I take it from reading some of
the reviews you've written that you don't think graphical-mode
find-the-pixel puzzles are any more interesting than text-mode
guess-the-verb puzzles, so I'm guessing this isn't what you meant to imply.
If find-the-pixel puzzles aren't interesting, then it seems to me that
Myst-style graphical IF has to amount to menu-based IF, not in visual
presentation style but in the essential UI design principle that all options
are plainly visible at all times. (Options don't have to be highlighted
with yellow rectangles to be plainly visible - if there's a part of the
picture that's obviously a separate object, then clicking on it is a plainly
visible option.)
My conjecture - and this is hard to prove experimentally, but I'll propose
it anyway - is that Myst-style graphic games tend to work in part because of
the design discipline that the UI imposes on the authors. When the authors
resist the temptation to play find-the-pixel, there's nothing left in the UI
to make puzzles obscure, so the authors are forced to create puzzles based
on statefulness and interactions among objects. When I think about the
parts of Myst-style games that have worked for me, this seems to hold true -
the interesting puzzles are the ones where you not only have to figure out
what an object does, but think strategically about how to get a desired
effect by adjusting the states of several such objects to combine in the
right way. The same thing often seems to be true in text IF - puzzles where
I have to come up with a new verb to try are rarely as interesting as those
requiring piecing together some information and performing actions involving
the basic, standard verbs.
Daniel Dawson
written by Mike Rozak <Mike...@bigpond.com>. It says:
>worth using before then.) It's lifespan would be another 5-10 years, by
>which point another 10x-20x improvements in computers is possible. (Unless
>Moore's law breaks, which it could.) Plus, to do the graphics right, you
Moore's Law must break soon (in about a decade, IIRC) because circuits can only
be made so dense before tunneling becomes a problem. Of course, I'm sure some
workarounds can be found by then, but those will only go so far.
--
| Email: Daniel Dawson <ddawson at icehouse.net> ifMUD: DanDawson |
| Web: http://www.icehouse.net/ddawson/ PGP key: 'about' on web site |
-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----
Richard Bos
> You pick up and read article <faoGb.64427$aT.5...@news-server.bigpond.net.au>,
> written by Mike Rozak <Mike...@bigpond.com>. It says:
> >worth using before then.) It's lifespan would be another 5-10 years, by
> >which point another 10x-20x improvements in computers is possible. (Unless
> >Moore's law breaks, which it could.) Plus, to do the graphics right, you
>
> Moore's Law must break soon (in about a decade, IIRC) because circuits can only
> be made so dense before tunneling becomes a problem. Of course, I'm sure some
> workarounds can be found by then, but those will only go so far.
Moore's Law has "had to" break down because of fundamental problems that
could not possibly be overcome by electronic trickery a couple of times
already. Every single time, the fundamental problem was overcome.
Richard
Fredrik Ramsberg
> You pick up and read article <faoGb.64427$aT.5...@news-server.bigpond.net.au>,
> written by Mike Rozak <Mike...@bigpond.com>. It says:
> >worth using before then.) It's lifespan would be another 5-10 years, by
> >which point another 10x-20x improvements in computers is possible. (Unless
> >Moore's law breaks, which it could.) Plus, to do the graphics right, you
>
> Moore's Law must break soon (in about a decade, IIRC) because circuits can only
> be made so dense before tunneling becomes a problem. Of course, I'm sure some
> workarounds can be found by then, but those will only go so far.
The point is that Moore's law has already outlived many older
technologies. When a technology can't be stretched any further,
something new replaces it.
/Fredrik
Daryl McCullough
>You'd never eliminate all the guess the verb problems though using a
>thesaurus or word-net. Another approach is to get 100+ beta testers and scan
>their play logs looking for unexpected (but correct) commands.
I think you're right. Synonyms for individual words is not the only problem,
there is also (and I think more importantly) various phrasings for synonymous
actions. For example, for hiding a key under a doormat, you could say
Put the doormat over the key.
Put the key under the doormat.
Cover the key with the doormat.
Hide the key under the doormat.
"Put" "Cover" and "Hide" are not normally synonyms, but they can
be used to convey synonymous actions.
--
Daryl McCullough
Ithaca, NY
L. Ross Raszewski
>
>The point is that Moore's law has already outlived many older
>technologies. When a technology can't be stretched any further,
>something new replaces it.
>
>/Fredrik
But that new technology isn't bound by the same limits. Yes, we've
seen superscalar and hyperthreading technologies which allow PCs to
get even faster than predicted, but there's simply a thermodynamic
limit on how fast the system can get. When we hit that, we've got to
make a much more radical paradigm shift than has been done before --
something like moving to totally optical systems. When that happens,
there's no reason the moore's law constants should still apply.
And before you point out how many naysaysers have been proved wrong
before, remember: Moore's law *did not* apply in the past. It didn't
start applying until some time in the 1970s (or possibly a little
earlier. But I don't think you'll find a Moore's law correspondance
between the difference engine, eniac, and the PDP-11, say), and you
can't expect it to apply to computers based on different fundamental
technologies.
Mike Roberts
> >You'd never eliminate all the guess the verb problems though
> >using a thesaurus or word-net.
>
> problem, there is also (and I think more importantly) various
> phrasings for synonymous actions.
That's definitely a big part of the problem.
Yet another is that if you create too many synonyms, at some point false
positives will start to become an issue - the user will type a command with
a particular meaning in mind, but the parser will consider the verb to be an
obscure synonym of some completely unintended action; or the user will refer
to an object with a particular noun or adjective, and the parser will find
the word surprisingly ambiguous because it's an obscure synonym of nouns and
adjectives for lots of other in-scope objects. The best natural language
parsers available today still aren't nearly as good as humans at sorting
these things out from context, so there's a delicate balance between
programming a parser with too few and too many synonyms.
Daryl McCullough
>The point is that Moore's law has already outlived many older
>technologies. When a technology can't be stretched any further,
>something new replaces it.
Somebody told me that Moore's law has become a self-fulfilling
prophecy:
Suppose you're a chip manufacturer, and it takes you T months
to create the next generation of your processor. For that next
generation to sell, it must be significantly better than the
current generation. Also, it must be approximately as good as,
or better than, the competition. But how do you know where the
competition will be in T months? You use Moore's law (or industrial
spies, I guess) to make a prediction. So you decide that you must
shoot for a chip that is a factor of F = e^{MT} better than your
current chip (where M is whatever exponent Moore's Law says it is).
There's no magic going on here. You could very well fail to
make your chip a factor of F better than your current one in time T.
But if it is *possible* to reach your goal, then you will devote
whatever resources it takes to make sure you *do* reach it.
Moore's Law will hold up until the time it no longer holds.
Mike Rozak
problem,
> there is also (and I think more importantly) various phrasings for
synonymous
> actions. For example, for hiding a key under a doormat, you could say
>
> Put the doormat over the key.
> Put the key under the doormat.
> Cover the key with the doormat.
> Hide the key under the doormat.
>
> "Put" "Cover" and "Hide" are not normally synonyms, but they can
> be used to convey synonymous actions.
This has probably been discussed in the archives, but I suspect that
typed-command interaction problems/frustrations occur in the following
cases: (All of which are more-or-less related to "guess the verb")
1) Synonyms - "put" vs. "place", or "doormat" vs. "door mat" vs. "rug"
2) Synonymous actions - as above
3) Actions not modelled because the author doesn't think of it - "Put the
key down. Put the mat down." leads to "Do you want to put the mat over the
key?"
4) Actions not modelled because it's too much work to model all
possibilities, even if they're predictable - "Bury the key in the soil next
to the door"
5) Actions not modelled because they would mess up the plot/game - "Slide
the key underneath the door" or "Break the key in half"... so much for
openeing the door ever again.
6) The command to solve the puzzle is properly parsed and handled, but not
really obvious. - "Slide the key underneath the door" ends up opening the
door for no obvious/predictable reason. (This isn't really a parsing issue,
but it is a source of frustration.)
Case (1), synonyms, can be solved 80% of the time with a thesaurus, and the
other 19% of the time by the author ("Zurg" = "Lord Zurg" = "Zurg the
mighty" = "the evil emperor"). There is always the 1% case that can't be
solved.
Case (2), synonymous actions, can be mostly solved by looking at logs from
beta testers. (Assuming there are enough, 100+.)
You can actually apply statistics to figuring out how well you solve this
problem: Take the logs from 100 beta testers and split into 4 groups of 25.
Find all of the ways that users phrase a given command. Take the first 25
(at random) and generate a heursitc based on them. Produce a % parsed
properly by passing the phrasing from all 100 users into the heuristic from
only 25 users. (For example: score w/25 = 80%) Then, modify the heuristic so
it works for the first 25 plus another 25 users. Produce another % parsed
properly. (For eample: score w/50 = 90%) Repeat. (w/75=93% and w/100=95%)
You end up with a graph of accuracy (parsed properly) vs. the training set.
From this graph you can guess how much more data you need to reach 100%
accuracy, and if it's even possible. (With speech recognition training,
you'll notice that the graph asymtotes, and SR can never be 100% accurate,
even with infinite amounts of training data.) Machine learning is
particularly useful here, since it saves the author from generating all the
algorithms. Not that anyone would do this, because it's a lot of work.
Although, once you take this line of thinking, the overall quality of
command parsing can be measured as the percent chance that a command does
what the user expects it to do. (Barring intentional plot twists, such as
putting the key under the mat is supposed to make it disappear.) Over the
course of beta testing, an author can plot this accuracy. When it reaches
99% (for example), the author decides that the command-parsing ability of
the IF is "good enough" and ships it.
Case (3), "didn't think of modelling it", can be solved similarly to case 2.
Case (4), "not worth modelling", will always happen. The data collection
from case (2) will inform an author how likely a user is to try a solution.
Unforunately, there are so many possible actions that you'd need 10,000
users to get a really good statistical reading.
Case (5), "would mess up the game", is just as open-ended as case (4), but
worrisome because once the author models all the game-breaking actions
properly, the author must then come up with ways to un-break the game, or
tell the user that they've just done (or are about to do) something very
stupid.
Case (6), "non-obvious solutions". I don't know what to say about this one.
Graphical IF (with menus or point-and-click) don't have to worry about case
(1) and case (2). Case (3) is less likely, but I have played graphical IF,
most recently URU, where either the authors didn't think of my solution
(case 3) or their laws of phyics worked differently than I'd expect (case
6). Case (4) and case (5) are less of a problem with menus and clicks
because users accept that if it's not on the menu or clickable they can't do
it. Users interpret limited menus as "rules" (or hints) that are in addition
to "the laws of physics". Such rules ends up abstracting the IF experience
so it becomes more of a game (with artifical rules, such as chess) than a
simulated reality (with just the laws of physics, such as a war-game - even
though war-games also have rules).
--
Mike Rozak
www.mXac.com.au