Do bit 6 in dictionary flags have any meaning?

heasm66 · February 3, 2022, 3:36pm

I’m experimenting writing a program for reverse ngineering z-machine files (think txd and infoobj). I don’t know if anything will ever come out of it but I wonder a bit about the dictionary format for Inform6 (both grammar versions). Below is and example output (grammar version 2):

2478: 68 A5 14 A5 94 A5 80 00 00 u         <noun>
2481: 6A 68 53 6A DC A5 41 E8 00 uncover   <verb>, Verb#=232
248A: 6A 69 2A E5 94 A5 08 00 00 under     <adj>
2493: 6A 69 50 A5 94 A5 80 00 00 undo      <noun>
249C: 6A 71 51 10 94 A5 C1 D3 00 unlock    <noun>+<verb>, Verb#=211
24A5: 6A 78 22 EA F0 A5 41 D4 00 unscrew   <verb>, Verb#=212
24AE: 6A 7C 5C D5 94 A5 41 E8 00 unwrap    <verb>, Verb#=232
24B7: 6A A5 14 A5 94 A5 88 00 00 up        <noun>+<adj>
24C0: 6A B8 39 2A 94 A5 08 00 00 upside    <adj>
24C9: 6A B8 39 2A A6 9C 08 00 00 upsidedow <adj>
24D2: 6D 57 1E 98 A8 A5 43 AF 00 verbose   <meta>+<verb>, Verb#=175

You may notice that on verbs both bit 6 & bit 1 are set on the flag byte (second to last number on each row). When I look in the Inform Technical Manual, Chapter 8 it has this to say:

These are the so-called "dictionary parameters" dict_par1, 2 and 3.  Inform's
pleasure is to write into them as follows:

    dict_par1: flags
    dict_par2: verb number (counting downwards from 255)
    dict_par3: preposition number (counting downwards from 255) in
               grammar version 1; not used in grammar version 2

The flags are given as follows:

    bit:    7      6   5   4   3     2        1      0
            <noun>             <adj> <plural> <meta> <verb>

The bits <verb>, <noun> and <adj> are set if the word can be used in the
context of a verb, noun and/or preposition (all three can be simultaneously
set).  The <meta> bit indicates that the English verb is "meta", that is,
is a command to the program and not a request in the game.

Do the bit 6 have any meaning, and/or are always bit 6 and bit 1 set in pair (seems so in the files I examined)?
Is there anything I’m missing and I’m a looking at the wrong place in the documentation?

zarf · February 3, 2022, 3:51pm

This is noted in the source:

github.com

DavidKinder/Inform6/blob/e2a204c63b6f406e00221724b450936f4f1fd555/text.c#L1658

      
        
            /*   first being received by these routines) and only alphabetically sorted  */
            /*   by construct_storyfile() (using the array below).                       */
            /* ------------------------------------------------------------------------- */
            /*                                                                           */
            /*   Further notes about the data fields...                                  */
            /*   The flags are currently:                                                */
            /*     bit 0: word is used as a verb (in verb grammar)                       */
            /*     bit 1: word is used as a meta verb                                    */
            /*     bit 2: word is plural (set by '//p')                                  */
            /*     bit 3: word is used as a preposition (in verb grammar)                */
            /*     bit 6: set for all verbs, but not used by the parser?                 */
            /*     bit 7: word is used as a noun (set for every word that appears in     */
            /*       code or in an object property)                                      */
            /*                                                                           */
            /*   In grammar version 2, the third field (adjectivenumber) is unused (and  */
            /*   zero).                                                                  */
            /*                                                                           */
            /*   The compiler generates special constants #dict_par1, #dict_par2,        */
            /*   #dict_par3 to refer to the byte offsets of the three fields. In         */
            /*   Z-code v3, these are 4/5/6; in v4+, they are 6/7/8. In Glulx, they      */
            /*   are $DICT_WORD_SIZE+2/4/6, referring to the *low* bytes of the three    */

…but I don’t remember ever seeing an explanation. It’s probably backwards compatibility for some specific hack that happened in 1994 or so.

heasm66 · February 3, 2022, 3:54pm

Thank you!

As usual I looked at the wrong place…