intfiction.org

The Interactive Fiction Community Forum
It is currently Tue Jul 25, 2017 9:50 am

All times are UTC - 6 hours [ DST ]




Post new topic Reply to topic  [ 33 posts ]  Go to page Previous  1, 2, 3, 4  Next
Author Message
PostPosted: Sun Nov 06, 2016 3:35 am 
Offline

Joined: Thu May 19, 2016 3:39 am
Posts: 65
cvaneseltine wrote:
For people who love parser, there is a parser-specific event (ParserComp) which existed in 2014/2015 and will exist again in 2016/2017. (Stay tuned!)


This sounds cool! I'm in!


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 5:00 am 
Offline

Joined: Mon Nov 19, 2012 4:51 pm
Posts: 191
Location: UK
The parser comp page lists certain formats as "Tools for writing parser games":
Quote:
    Inform 7
    TADS 3
    Inform 6
    TADS 2
    ADRIFT
    Quest
    Hugo
    Alan

I notice that Twine isn't included. Can we have a commitment that ParserComp will allow any game which accepts text input and won't discriminate against games based on format?

The winners of IF Comp in the recent past were all adventures. It's just a coincidence that they were made with traditional parser systems. The 2015 winner, Brain Guzzlers from Beyond!, partly uses a choice-based interface. It's only a slight exaggeration to say that it could have been made in a non-traditional system. I would like to see "ParserComp" broaden out to accept the adventures we love to play and want to make in any format.

This year Detectiveland, by Robin Johnson, is very much an adventure in the traditional style, yet you can't even type anything! I'm not sure about the rules, would Detectiveland be banned from Parsercomp? People who have never used an interpreter program are playing Detectiveland. Games made with the traditional systems are welcome. But don't we want to encourage new ideas and innovation?


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 8:19 am 
Offline
User avatar

Joined: Thu May 19, 2016 1:42 am
Posts: 108
Location: The Planet of Ocherous
heartless zombie wrote:
The parser comp page lists certain formats as "Tools for writing parser games":


The list in question (found here) is a list of SUGGESTED tools, with nothing implying it is an exhaustive list. It is also a list specifically aimed at people who want to write a parser game but has no idea where to start. Trimming the context away from a quote until it appears to say whatever you want it to say for your argument to work is quote-mining and it is extremely dishonest. More to the point, it's just frustrating to be on the receiving end of, as I know from experience. You end up having to explain things people very well know (such as the difference between a prescriptive list and a descriptive list of common systems) while whatever point anyone might have had gets lost in the shuffle.

_________________
"He never really loved her. Or her money. He wanted her secrets."


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 8:25 am 
Offline
User avatar

Joined: Tue Apr 20, 2010 2:48 pm
Posts: 935
Location: Greece
If it was up to me, I would simply ban everyone who deals in these (non)visual novels.


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 11:41 am 
Offline

Joined: Tue Nov 08, 2011 8:11 am
Posts: 2000
Location: US - Central
heartless zombie wrote:
The parser comp page lists certain formats as "Tools for writing parser games":
Quote:
    Inform 7
    TADS 3
    Inform 6
    TADS 2
    ADRIFT
    Quest
    Hugo
    Alan

I notice that Twine isn't included. Can we have a commitment that ParserComp will allow any game which accepts text input and won't discriminate against games based on format?

The issue here is that Twine is not a parser engine. Accepting text input is not the sole criteria for a parser game.

While Twine and other choice-based engines will accept text and use it in the story, it is just a variable input and is not being "parsed" in the technical sense. While some talented Twine programmers could do some nifty tricks to simulate that the game is parsing their input, it would still be simulation of an actual parser.

Robin Johnson has stated that he wrote his own parser in Java for Detectiveland, and pressing the buttons actually creates a sentence that goes into the transcript. He can verify if this input is actually being parsed or if each button command is built as 1:1 recognition

Saying that games which are not built using a parser engine and don't qualify for a parser competition are being "discriminated against" because they don't use a parser is like saying that the Westminster Kennel Club competition discriminates against cats.

"Discriminate" is a thorny loaded word which (IMHO) should be limited to situations involving discrimination against people, and not rules for game jams. Nothing prevents anyone from writing a parser game and entering it.

_________________
http://hanonondricek.wix.com/pyramidif


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 12:32 pm 
Offline

Joined: Sat Jan 23, 2010 4:56 pm
Posts: 5390
Quote:
While some talented Twine programmers could do some nifty tricks to simulate that the game is parsing their input, it would still be simulation of an actual parser.


To forestall the obvious retort, it would be *possible* to implement a full IF-style parser in Javascript, build enough of a world model to react to the parser commands, and then embed *that* in the Twine framework.

(It would be a lot more work than using an existing, well-tested parser engine, so I don't recommend it.)

However. *Backing up.* The original ParserComp site has, as its first question, "What is a parser game?" And the nut of the answer is: "To play a parser game, players type commands at a prompt, and the game attempts to parse those commands and carry out actions."

This seems like a perfectly reasonable common-sense description. If there are judgement calls to be made, Carolyn is capable of making them as necessary.


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 12:59 pm 
Offline

Joined: Tue Nov 08, 2011 8:11 am
Posts: 2000
Location: US - Central
Agreed. And I will say if someone did that, it would fall into the "experimental roll your own parser" category that would be welcome and a great experimental showpiece for ParserComp.

One of my controversial statements last year was "A parser can reject a player's command based on its evaluation of rules and the game state." which is arguable that Twine can offer a choice but then disallow it if a flag isn't set. My counter was that's a 1:1 simulation of parser that the author sets up. Parsers can account for unexpected input and make a judgement. Choice-based systems, by definition, don't allow for unexpected input.

It's similar to the difference between a multiple choice test where answers are phrased explicitly and can be instantly scored right or wrong, and an essay question which needs the answer input read, comprehended, and judged based on interpretation.

_________________
http://hanonondricek.wix.com/pyramidif


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 1:15 pm 
Offline

Joined: Fri Dec 04, 2015 5:23 pm
Posts: 150
Location: Scotland
HanonO wrote:
While Twine and other choice-based engines will accept text and use it in the story, it is just a variable input and is not being "parsed" in the technical sense. While some talented Twine programmers could do some nifty tricks to simulate that the game is parsing their input, it would still be simulation of an actual parser.

Well, you could say that Inform 7 merely "simulates" that the game is parsing the input, since it doesn't have any knowledge of what a brass lantern actually is, but that way lies philosophy.

Quote:
Robin Johnson has stated that he wrote his own parser in Java for Detectiveland, and pressing the buttons actually creates a sentence that goes into the transcript. He can verify if this input is actually being parsed or if each button command is built as 1:1 recognition

(Javascript, not Java.) It doesn't create the sentence and then "parse" it as a string, no. It creates a model action containing a verb and (usually) an object - the concept of getting a lamp, which can then be run through the item-object "lamp" and the verb-object "get" - the same sort of model as a simple two-word parser would produce from an input like "get lamp". You could say that the parser is precisely what is missing, since the player is directly inputting the action through a button without the game needing to parse text to find out what the action is. The rest of the game is structured similarly to a parser game, in terms of rooms, items, declarations of what happens when you apply a particular verb to a particular item, and so on.

I've written parser games in Javascript too (see links in my signature), but I wrote that engine 13 years ago as a much worse programmer. It pretends to be a lot cleverer than it actually is, and doesn't really create a sensible action-object internally. For some definitions of 'parser', it's a much worse parser than Detectiveland. But I'd hope a game made with that engine would be eligible for Parsercomp.

For the record, I would not expect Detectiveland to be eligible for ParserComp. If someone were crazy, stupid, masochistic and/or brilliant enough to write a parser in Twine - or a convincing thing that pretends to be a parser, though I'm not sure what the difference is - it's not up to me of course, but I'd be disappointed if it wasn't allowed. (And I don't think Parsercomp's listing of some sensible parser authoring systems implicitly forbids it.)

_________________
IFComp 2016: Detectiveland
My games: The Xylophoniad, Draculaland, Portcullis, Aunts and Butlers, and Hamlet

Tweets: @rdouglasjohnson


Last edited by robinjohnson on Sun Nov 06, 2016 2:22 pm, edited 1 time in total.

Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 2:02 pm 
Offline

Joined: Sat Jan 23, 2010 4:56 pm
Posts: 5390
Quote:
It doesn't create the sentence and then "parse" it as a string, no. It creates a model action containing a verb and (usually) an object - the concept of getting a lamp, which can then be run through the item-object "lamp" and the verb-object "get" - the same sort of model as a simple two-word parser would produce from an input like "get lamp".


Glad to hear it. :)

(I figured that's how Detectiveland was constructed.)

(Because including a full parser would have been both harder and more error-prone.)


Top
 Profile Send private message  
 
PostPosted: Sun Nov 06, 2016 2:28 pm 
Offline

Joined: Fri Dec 04, 2015 5:23 pm
Posts: 150
Location: Scotland
It's just occurred to me that Detectiveland's engine has an unparser, since the action creates a string (usually, but not always, the verb and the object noun separated by a space) to put into the transcript window.

It can even do ambiguation.

_________________
IFComp 2016: Detectiveland
My games: The Xylophoniad, Draculaland, Portcullis, Aunts and Butlers, and Hamlet

Tweets: @rdouglasjohnson


Top
 Profile Send private message  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 33 posts ]  Go to page Previous  1, 2, 3, 4  Next

All times are UTC - 6 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 2 guests


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 post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group