Defining quests with natural language?

 
Post new topic   Reply to topic    mudlab.org Forum Index -> Design
View previous topic :: View next topic  
Author Message
ide



Joined: 21 Feb 2006
Posts: 105
Location: Seattle

PostPosted: Mon May 01, 2006 5:18 am    Post subject: Defining quests with natural language? Reply with quote

There's been an interesting new release in the interactive fiction (IF) community today, a development language based on a natural language subset. So instead of writing a quest as a script, like many people here are used to (using DG, mob progs, embedded language, etc.) you would write many basically declarative sentences to do the same thing; the language processor turns these sentences into your script.

Not being a programmer I don't know how well mud coders would take to such a thing, but I can guess that many mud builders might enjoy a similar language for writing quests. Anyway, for your amusement the links are below. Thoughts?,

http://www.inform-fiction.org/I7/Inform%207.html

Check:


http://brasslantern.org/writers/howto/i7intro.html

for more information. No doubt the rec.arts.int-fiction group will have ongoing discussion as well.
Back to top
View user's profile Send private message Send e-mail
Author Message
Nornagest



Joined: 08 Jan 2006
Posts: 12
Location: California

PostPosted: Mon May 01, 2006 11:46 pm    Post subject: Reply with quote

I can't make any definitive judgments without playing around with the system for a while, but, judging from the links you provide, I don't think this would make quest scripting significantly easier.

There's some effort involved in learning the syntax for a scripting language, to be sure. But the real difficulty of writing a working script comes from organizing your thoughts in a way that the parser will accept. Since I very much doubt this system can handle obtuse phrasing, idiomatic expressions, lack of clarity, and all the other unfortunate realities of natural language in an elegant way, it basically just defines another scripting language; as such, it'd likely take just as long to gain a good understanding of what it accepts as it would to learn an (unnatural?) scripting language in the first place. It might even be more difficult, since I can see a natural-language subset being significantly harder to debug.

I hate to sound like a prophet of doom here, but there are serious problems with getting a computer to understand natural language. If this system's managed to overcome them, more power to it, but that's an exceptionally difficult proposition.
Back to top
View user's profile Send private message
Author Message
MikeRozak



Joined: 27 Nov 2005
Posts: 15
Location: Australia

PostPosted: Tue May 02, 2006 6:04 am    Post subject: Reply with quote

It looks very interesting, but I'm not sure about the NLP approach. (I might change my mind.) Some of it might be better handled with a GUI to enter values, and then have the values displayed as NLP derived from the GUI input. However, the editor does look to be well designed, which should make the author's life a lot easier.

I've been playing around with my tools for http://www.mxac.com.au/mif to make them a bit easier, while still trying to maintain flexibility. This is tricky, since if you make the editor so easy that anyone can create a world in 10 seconds flat, you also limit how flexible their worlds can be. The net result is 10,000 user-created worlds that are all tripe. Many IF editors take this approach. Inform has always opted for more flexibility (even with 7), which also increases the authoring difficulty. The net is only 100's (not 10,000) of worlds, but they're fun to play.

I suppose the same could be said for MUD quests... I can make a quest editor that makes it really easy to create kill N monsters, but it wouldn't allow the flexibility that quests in Oblivion are displaying. I've scanned through Oblivion's documentation and think their AI/quests could be a bit more flexible.

I wish Inform 7 had spent more time on AI (which I don't see much mention of). I think that without a solid AI foundation, IF titles will suffer, even with a really nice editor. Quests (and just about everything) get more complex as AI complexity increases.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Author Message
Massaria



Joined: 14 May 2005
Posts: 31
Location: Denmark

PostPosted: Tue May 02, 2006 11:35 pm    Post subject: Reply with quote

I feel a little silly replying to this topic, as I think the previous two replies have got it covered pretty damn good. Especially Nornagast, when he says:

Quote:
There's some effort involved in learning the syntax for a scripting language, to be sure. But the real difficulty of writing a working script comes from organizing your thoughts in a way that the parser will accept.


I've played around a little with I7, and I've found the syntax only marginally easier than Ruby to understand.
It seems like you'd be giving up a lot of freedom for a system with an only slightly gentler learning curve.

He's a little off the mark here, though:

Quote:
It might even be more difficult, since I can see a natural-language subset being significantly harder to debug.


It seems you might be right, for now. The 'Transcript' function of I7, which seems an integral part of testing, isn't done for the windows version - and I actually ran into some errors I couldn't locate. But assuming that the parser can tell you at least where the problem is, I think NLP is easier to learn, however slightly.
It just seems less foreign, I think, to someone unfamiliar with any kind of scripting. It should prove an excellent stepping stone towards more versatile applications/languages.

I think the creator of I7 has done a really fine job at what he set out to do: to make interactive fiction available to writers. The validity of the claim to be helping programmers write, is more dubious, I think.

I think I very well might have used this a couple of years ago, but after you've seen the world through Ruby... Wink
Back to top
View user's profile Send private message
Author Message
thyrr



Joined: 18 Apr 2006
Posts: 7

PostPosted: Fri May 05, 2006 10:04 am    Post subject: Reply with quote

"The anti-Perl" would indeed be a good description. Easy to read, annoying to write, at least from my cursory glance over the articles. Natural language works in Interactive Fiction because it's limited to commands. I7 seems to have a degree of logic relations implemented.

Still, it's very impressive. Maybe a good first step. The slick IDE is definitely a good idea; being able to see how the parser interprets your writing by showing a new room on the map or an item in the room goes a long away. It'd be neat if it translated your writing to normal programming code on-the-fly in a panel so you could see a concise representation of what you're writing. Since you're able to describe things in multiple ways, an authoritative interpretation by the parser would be useful for debugging and learning.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    mudlab.org Forum Index -> Design All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum

Powered by phpBB © 2001, 2002 phpBB Group
BBTech Template by © 2003-04 MDesign