~ubuntu-branches/ubuntu/vivid/inform/vivid

<TR><TD Valign="top"><A HREF="contents.html">Contents</A> <A HREF="section23.html">Back</A> <A HREF="section25.html">Forward</A><TD bgcolor="#F5DEB3"><BLOCKQUOTE><H3>24. How nouns are parsed</H3></BLOCKQUOTE><TR><TD><TD>

The Naming of Cats is a difficult matter,

It isn't just one of your holiday games;

You may think at first I'm as mad as a hatter

When I tell you, a cat must have THREE DIFFERENT NAMES.

...T. S. Eliot (1888--1965), The Naming of Cats</BLOCKQUOTE>

Bulldust, coolamon, dashiki, fizgig, grungy, jirble, pachinko,

poodle-faker, sharny, taghairm

...Catachrestic words from Chambers English Dictionary</BLOCKQUOTE>

Suppose we have a tomato defined with

<PRE>

name "fried" "green" "tomato",

</PRE>

but which is going to redden later and need to be referred to as "red

tomato''. It's perfectly straightforward to alter the <TT>name</TT> property of

an object, which is a word array of dictionary words. For example,

<PRE>

[ Names obj i;

for (i=0:2*i<obj.#name:i++) print (address) (obj.&name)-->i, "^";

];

</PRE>

prints out the list of dictionary words held in <TT>name</TT> for a given

object. It's perfectly possible to write to this, so we could just set

<PRE>

(tomato.&name)-->1 = 'red';

</PRE>

but this is not a flexible or elegant solution, and it's time to

begin delving into the parser.

<TR><TD Valign="top"><IMG SRC="icons/dbend.gif" ALT="/\"><TD bgcolor="#EEEEEE"> Note that we can't change the size of the <TT>name</TT> array.

To simulate this, we could define the object with

<TT>name</TT> set to, say, 30 copies of an 'untypeable word' (see below)

such as <TT>'blank.'</TT>.

The Inform parser is designed to be as "open-access'' as possible,

because a parser cannot ever be general enough for every game without

being highly modifiable. The first thing it does

is to read in text from the keyboard and break it up into a

stream of words:

so the text "wizened man, eat the grey bread'' becomes

<TT>wizened</TT> / <TT>man</TT> / <TT>,</TT> / <TT>eat</TT> / <TT>the</TT> / <TT>grey</TT> / <TT>bread</TT>

</BLOCKQUOTE>

and these words are numbered from 1. At all times the parser keeps

a "word number'' marker to keep its place along this line, and this

is held in the variable <TT>wn</TT>. The routine <TT>NextWord()</TT> returns

the word at the current position of the marker, and moves it forward,

i.e. adds 1 to <TT>wn</TT>. For instance, the parser may find itself at

word 6 and trying to match "grey bread'' as the name of an object.

Calling <TT>NextWord()</TT> gives the value <TT>'grey'</TT> and calling it again

gives <TT>'bread'</TT>.

Note that if the player had mistyped "grye bread'', "grye'' being a word

which isn't mentioned anywhere in the program or created by the library,

<TT>NextWord()</TT> returns 0 for 'misunderstood word'. Writing something like

<TT>if (w=='grye') ...</TT> somewhere in the program makes Inform put "grye''

into the dictionary automatically.

<TR><TD Valign="top"><IMG SRC="icons/dbend.gif" ALT="/\"><TD bgcolor="#EEEEEE"> Remember that the game's dictionary only has 9-character

resolution. (And only 6 if Inform has been

told to compile an early-model story file: see <A HREF="section31.html">Section 31</A>.)

Thus the values of <TT>'polyunsaturate'</TT> and

<TT>'polyunsaturated'</TT> are equal.

Also, upper case and lower case letters are considered the same.

Words are permitted to contain numerals or symbols (but not at

present to contain accented

characters).

<TR><TD Valign="top"><IMG SRC="icons/ddbend.gif" ALT="/\/\"><TD bgcolor="#EEEEEE"> A dictionary word can even contain spaces, full stops or commas.

If so it is 'untypeable'. For instance, <TT>'in,out'</TT> is an untypeable

word because if the player does type it then the parser cuts it into

three, never checking the dictionary for the entire word. Thus the

100

constant <TT>'in,out'</TT> can never be anything that <TT>NextWord</TT> returns.

101

This can actually be useful (as it was in <A HREF="section16.html">Section 16</A>).

102

103

104

105

106

<TR><TD Valign="top"><IMG SRC="icons/dbend.gif" ALT="/\"><TD bgcolor="#EEEEEE"> It can also be useful to check for numbers. The library

107

routine <TT>TryNumber(wordnum)</TT> tries to parse the word

108

at <TT>wordnum</TT> as a number (recognising decimal numbers and English

109

ones from "one'' to "twenty''), returning -1000 if it fails

110

altogether, or else the number. Values exceeding 10000 are rounded

111

down to 10000.

112

113

114

115

<TR><TD Valign="top"><IMG SRC="icons/ddbend.gif" ALT="/\/\"><TD bgcolor="#EEEEEE"> Sometimes there is no alternative but to actually look at

116

the player's text one character at a time (for instance, to check

117

a 20-digit phone number). The routine <TT>WordAddress(wordnum)</TT>

118

returns a byte array of the characters in the word, and

119

<TT>WordLength(wordnum)</TT> tells you how many characters there are in it.

120

Thus in the above example,

121

<PRE>

122

thetext = WordAddress(4);

123

print WordLength(4), " ", (char) thetext->0, (char) thetext->2;

124

</PRE>

125

126

127

128

prints the text "3 et''.

129

130

131

132

133

An object can affect how its name is parsed by giving a <TT>parse_name</TT>

134

routine. This is expected to try to match as many words as possible

135

starting from the current position of <TT>wn</TT>, reading them in one at

136

a time using the <TT>NextWord()</TT> routine. Thus it must not stop just because

137

the first word makes sense, but must keep reading and find out how many

138

words in a row make sense.

139

It should return:

140

141

0<SAMP> </SAMP> if the text didn't make any sense at all,

142

k<SAMP> </SAMP> if k words in a row of the text seem to refer to the object, or

143

-1<SAMP> </SAMP> to tell the parser it doesn't want to decide after all.

144

</BLOCKQUOTE>

145

146

The word marker <TT>wn</TT> can be left anywhere afterwards. For example:

147

<PRE>

148

Object -> thing "weird thing"

149

with parse_name

150

[ i; while (NextWord()=='weird' or 'thing') i++;

151

return i;

152

];

153

</PRE>

154

155

This definition duplicates (very nearly) the effect of having defined:

156

<PRE>

157

Object -> thing "weird thing"

158

with name "weird" "thing";

159

</PRE>

160

161

Which isn't very useful. But the tomato can now be coded up with

162

<PRE>

163

parse_name

164

[ i j; if (self has general) j='red'; else j='green';

165

while (NextWord()=='tomato' or 'fried' or j) i++;

166

return i;

167

168

</PRE>

169

170

so that "green" only applies until its <TT>general</TT> attribute has

171

been set, whereupon "red'' does.

172

173

174

<TR><TD Valign="top"><IMG SRC="icons/exercise.gif" ALT="??"><TD bgcolor="#FBB9AC"><A NAME="ex56">EXERCISE 56: (link to <A HREF="answers2/answer56.html">the answer</A>)<TR><TD><TD> Rewrite this to insist that the adjectives

175

must come before the noun, which must be

176

present.

177

178

179

<TR><TD Valign="top"><IMG SRC="icons/exercise.gif" ALT="??"><TD bgcolor="#FBB9AC"><A NAME="ex57">EXERCISE 57: (link to <A HREF="answers2/answer57.html">the answer</A>)<TR><TD><TD> Create a musician called Princess who, when kissed,

180

is transformed into "<TT>/?%?/</TT> (the

181

artiste formerly known as Princess)''.

182

183

184

<TR><TD Valign="top"><IMG SRC="icons/exercise.gif" ALT="??"><TD bgcolor="#FBB9AC"><A NAME="ex58">EXERCISE 58: (link to <A HREF="answers2/answer58.html">the answer</A>)<TR><TD><TD> (Cf. 'Cafè Inform'.) Construct a drinks machine capable

185

of serving cola, coffee or tea, using only one object for the buttons

186

and one for the possible drinks.

187

188

189

<TR><TD Valign="top"><IMG SRC="icons/dbend.gif" ALT="/\"><TD bgcolor="#EEEEEE"> <TT>parse_name</TT> is also used to spot plurals: see <A HREF="section25.html">Section 25</A>.

190

191

192

193

194

Suppose that an object doesn't have a <TT>parse_name</TT> routine, or that

195

it has but it returned -1. The parser then looks at the <TT>name</TT>

196

words. It recognises any arrangement of some or all of these words

197

as a match (the more words, the better). Thus "fried green tomato'' is

198

understood, as are "fried tomato'' and "green tomato''. On the other

199

hand, so are "fried green'' and "green green tomato green fried green''.

200

This method is quick and good at understanding a wide variety of

201

sensible inputs, though bad at throwing out foolish ones.

202

203

204

However, you can affect this by using the <TT>ParseNoun</TT> entry point.

205

This is called with one argument, the object in question, and should

206

work exactly as if it were a <TT>parse_name</TT> routine: i.e., returning

207

-1, 0 or the number of words matched as above. Remember that it

208

is called very often and should not be horribly slow. For example,

209

the following duplicates what the parser usually does:

210

<PRE>

211

[ ParseNoun obj n;

212

while (IsAWordIn(NextWord(),obj,name) == 1) n++; return n;

213

];

214

[ IsAWordIn w obj prop k l m;

215

k=obj.&prop; l=(obj.#prop)/2;

216

for (m=0:m<l:m++)

217

if (w==k-->m) rtrue;

218

rfalse;

219

];

220

</PRE>

221

222

In this example <TT>IsAWordIn</TT> just checks to see if <TT>w</TT> is one of the

223

entries in the word array <TT>obj.&prop</TT>.

224

225

226

<TR><TD Valign="top"><IMG SRC="icons/dexercise.gif" ALT="??/\"><TD bgcolor="#FBB9AC"><A NAME="ex59">EXERCISE 59: (link to <A HREF="answers2/answer59.html">the answer</A>)<TR><TD><TD> Many adventure-game parsers split object names into

227

'adjectives' and 'nouns', so that only the pattern

228

<0 or more adjectives> <1 or more nouns>

229

is recognised. Implement this.

230

231

232

<TR><TD Valign="top"><IMG SRC="icons/dexercise.gif" ALT="??/\"><TD bgcolor="#FBB9AC"><A NAME="ex60">EXERCISE 60: (link to <A HREF="answers2/answer60.html">the answer</A>)<TR><TD><TD> During debugging it sometimes helps to

233

be able to refer to objects by their internal numbers, so that

234

"put object 31 on object 5'' would work. Implement

235

this.

236

237

238

239

240

<TR><TD Valign="top"><IMG SRC="icons/dexercise.gif" ALT="??/\"><TD bgcolor="#FBB9AC"><A NAME="ex61">EXERCISE 61: (link to <A HREF="answers2/answer61.html">the answer</A>)<TR><TD><TD> How could the word "<TT>#</TT>'' be made a wild-card,

241

meaning "match any single object''?

242

243

244

245

<TR><TD Valign="top"><IMG SRC="icons/ddexercise.gif" ALT="??/\/\"><TD bgcolor="#FBB9AC"><A NAME="ex62">EXERCISE 62: (link to <A HREF="answers2/answer62.html">the answer</A>)<TR><TD><TD> And how could "<TT>*</TT>'' be a wild-card for

246

"match any collection of objects''?

247

248

249

250

<TR><TD Valign="top"><IMG SRC="icons/ddexercise.gif" ALT="??/\/\"><TD bgcolor="#FBB9AC"><A NAME="ex63">EXERCISE 63: (link to <A HREF="answers2/answer63.html">the answer</A>)<TR><TD><TD> There is no problem with calling a container

251

"hole in wall'', because the parser will understand

252

"put apple in hole in wall'' as "put (apple) in (hole in wall)''.

253

But create a fly in amber, so that "put fly in amber in

254

hole in wall'' works properly and isn't misinterpreted as

255

"put (fly) in (amber in hole in wall)''.

256

(Warning: you may need to know about the <TT>BeforeParsing</TT> entry

257

point (see <A HREF="section26.html">Section 26</A>) and the format of the <TT>parse</TT> buffer (see

258

<A HREF="section27.html">Section 27</A>).)

259

260

261

<TR><TD Valign="top"><IMG SRC="icons/refs.gif" ALT="*"><TD bgcolor="#EEEEEE">REFERENCES: Straightforward <TT>parse_name</TT> examples are the chess-pieces object

262

and the kittens class of 'Alice Through The Looking-Glass'. Lengthier

263

ones are found in 'Balances', especially in the white cubes class.

264

</TABLE>

265

<HR><A HREF="contents.html">Contents</A> / <A HREF="section23.html">Back</A> / <A HREF="section25.html">Forward</A>

266

<A HREF="chapter1.html">Chapter I</A> / <A HREF="chapter2.html">Chapter II</A> / <A HREF="chapter3.html">Chapter III</A> / <A HREF="chapter4.html">Chapter IV</A> / <A HREF="chapter5.html">Chapter V</A> / <A HREF="chapter6.html">Chapter VI</A> / <A HREF="chapterA.html">Appendix</A><HR>Mechanically translated to HTML from third edition as revised 16 May 1997. Copyright © Graham Nelson 1993, 1994, 1995, 1996, 1997: all rights reserved.</BODY></HTML>

Older »