Getting started with IF: on homemade parsers

[1***@gmail.com]

I recently started writing some OOP classes to represent rooms, items, etc. for IF. I want to try my hand at writing IF, but I have a question.

I would be writing my own parser, at least if I want to use the classes I've written. How difficult is it to create a usable, even if simple, parser? It could be as basic as noun-verb, though I'd prefer something a little bit more sophisticated. I could learn Inform, naturally, but I so far I've enjoyed trying to program this thing myself. Should I write a few simplistic/"naive" parsers first, for practice? What should they do?

I don't have any story for a game yet, but inspiration may come as I program (especially since I have to have some sort of world in place to test parser/world model components as I add them.)

[John W Kennedy]

Post by 1***@gmail.com
I recently started writing some OOP classes to represent rooms, items, etc. for IF. I want to try my hand at writing IF, but I have a question.
I would be writing my own parser, at least if I want to use the classes I've written. How difficult is it to create a usable, even if simple, parser? It could be as basic as noun-verb, though I'd prefer something a little bit more sophisticated. I could learn Inform, naturally, but I so far I've enjoyed trying to program this thing myself. Should I write a few simplistic/"naive" parsers first, for practice? What should they do?
I don't have any story for a game yet, but inspiration may come as I program (especially since I have to have some sort of world in place to test parser/world model components as I add them.)

It’s hard. Basically, you can’t just build a parser, the parser has to
be tightly integrated with the world model. It needs to know that you
shouldn’t name anything you can’t see, or at least that you haven’t
seen, or interact with anything you can’t touch (except that you can
look at it if it’s behind glass doors, or talk to it if it’s behind
bars). I strongly recommend you master both Inform 6 and Inform 7, or
something equivalent, just to give you an understanding of the nature of
the problems involved.

[Peter Wiehe]

Post by 1***@gmail.com
I recently started writing some OOP classes to represent rooms, items, etc. for IF. I want to try my hand at writing IF, but I have a question.
I would be writing my own parser, at least if I want to use the classes I've written. How difficult is it to create a usable, even if simple, parser? It could be as basic as noun-verb, though I'd prefer something a little bit more sophisticated. I could learn Inform, naturally, but I so far I've enjoyed trying to program this thing myself. Should I write a few simplistic/"naive" parsers first, for practice? What should they do?
I don't have any story for a game yet, but inspiration may come as I program (especially since I have to have some sort of world in place to test parser/world model components as I add them.)

In my humble opinion it makes sense to start with a simple verb noun parser. You can improve it later. Of course it depends on your coding mentality. I mean: Perhaps you are a programmer who has more fun when the challenge is greater.

The premium solution would be to separate game data from the general game mechanisms. But that's more difficult. But this way you could publish your game once for all systems and let the player use it on every machine he/she wishes with a system-specific game-runner.
Greetings
Peter

[n***@zzo38computer.org]

Post by 1***@gmail.com
I recently started writing some OOP classes to represent rooms, items, etc.
for IF. I want to try my hand at writing IF, but I have a question.
I would be writing my own parser, at least if I want to use the classes I've
e written. How difficult is it to create a usable, even if simple, parser?
It could be as basic as noun-verb, though I'd prefer something a little bit
more sophisticated. I could learn Inform, naturally, but I so far I've
enjoyed trying to program this thing myself. Should I write a few
simplistic/"naive" parsers first, for practice? What should they do?
I don't have any story for a game yet, but inspiration may come as I program
(especially since I have to have some sort of world in place to test parser/
world model components as I add them.)

What programming language are you using? If you are interested, you may
look at what I implemented for xyzabcde2 (Game of XYZABCDE -- Part II). It
works basically like this:

1. Convert input to lowercase, discard most punctuation, and split the
input into words.

2. Use a binary search in the vocabulary table to convert each word into
the corresponding number. If not found, display an error message. (The
index of the word in the table is used, and it is defined as a compile
time constant, which is then referenced in the rest of the program.)

3. Look up the first word in the verbs table, and call the function listed
there. If not found, display an error message.

4. The verb implementation function will call other functions for parsing
additional words.

5. The noun parse function looks for available objects (currently, far
objects are not implemented), and reads the noun phrases for those objects
to match them with user input.

6. With each noun phrase of an object, it has a priority value associated
with it; there is also a function to determine the extra priority due to
the context.

7. The object with the best priority is used. If there is a tie, then
display an error message, unless the objects have the "indistinguish" flag
set and have the same prototype, in which case, just pick one arbitrarily.

Also, even if you do not use Inform, I would recommend to use a VM such as
Glulx or Z-machine or TAVERN32, if possible. (I do not know of any program
to compile a C code to Glulx, and there is perhaps the difficulty that you
cannot take the address of a local variable in Glulx.)

[Happy MAC XL]

I know you've probably got your heart set on a DIY parser, but can I tempt you to look at ZIL and ZILF? You can edit that existing parser, or other library files, to your heart's content!