I am a Forth novice writing an application with many fixed text strings of widely differing lengths, indexed by two dimensions (with several additional attributes, like title and date.) I am currently using gForth.
I plan to create about 600 words like:
Group1Subgroup7Title ( -- caddr len ) S" A Subgroup Title" ;
But my question is: how can I programmatically execute a word from a string I've generated within another word? It seems like I need to put the word in the input buffer and call parse-word, but I haven't discovered how to do that. For example:
: PrintTitle ( group sub -- )
make-subgroup-title \ consumes group & sub, returns caddr and len or cstring
somehow-invoke-the-word-that-text-represents \ returns caddr len
type ;
I can write make-subgroup-title, but do not understand how to write somehow-invoke-the-word-that-text-represents.
Thanks for your time!

Van Kichline schrieb am Montag, 27. März 2023 um 02:08:07 UTC+2:
> I am a Forth novice writing an application with many fixed text strings of widely differing lengths, indexed by two dimensions (with several additional attributes, like title and date.) I am currently using gForth.
> I plan to create about 600 words like:
> Group1Subgroup7Title ( -- caddr len ) S" A Subgroup Title" ;
> But my question is: how can I programmatically execute a word from a string I've generated within another word? It seems like I need to put the word in the input buffer and call parse-word, but I haven't discovered how to do that. For example:
> : PrintTitle ( group sub -- )
> make-subgroup-title \ consumes group & sub, returns caddr and len or cstring
> somehow-invoke-the-word-that-text-represents \ returns caddr len
> type ;
> I can write make-subgroup-title, but do not understand how to write somehow-invoke-the-word-that-text-represents.
> Thanks for your time!

Try something along PARSE-NAME EVALUATE

minforth schrieb am Montag, 27. März 2023 um 07:44:53 UTC+2:
> Van Kichline schrieb am Montag, 27. März 2023 um 02:08:07 UTC+2:
> > I am a Forth novice writing an application with many fixed text strings of widely differing lengths, indexed by two dimensions (with several additional attributes, like title and date.) I am currently using gForth.
> > I plan to create about 600 words like:
> > Group1Subgroup7Title ( -- caddr len ) S" A Subgroup Title" ;
> > But my question is: how can I programmatically execute a word from a string I've generated within another word? It seems like I need to put the word in the input buffer and call parse-word, but I haven't discovered how to do that. For example:
> > : PrintTitle ( group sub -- )
> > make-subgroup-title \ consumes group & sub, returns caddr and len or cstring
> > somehow-invoke-the-word-that-text-represents \ returns caddr len
> > type ;
> > I can write make-subgroup-title, but do not understand how to write somehow-invoke-the-word-that-text-represents.
> > Thanks for your time!
> Try something along PARSE-NAME EVALUATE

Another helper word to separate names within non-empty(!) strings:

: PARSE-STRING ( a u -- a' u' ap up )
BEGIN over c@ bl = WHILE 1 /string REPEAT
over >r
BEGIN over c@ bl <> WHILE 1 /string REPEAT
over r@ - r> swap ;
\ test
s" fair may 2023 "
parse-string cr type
parse-string cr type
parse-string cr type

On Monday, 27 March 2023 at 01:08:07 UTC+1, Van Kichline wrote:
> I am a Forth novice writing an application with many fixed text strings of widely differing lengths, indexed by two dimensions (with several additional attributes, like title and date.) I am currently using gForth.
> I plan to create about 600 words like:
> Group1Subgroup7Title ( -- caddr len ) S" A Subgroup Title" ;
> But my question is: how can I programmatically execute a word from a string I've generated within another word? It seems like I need to put the word in the input buffer and call parse-word, but I haven't discovered how to do that. For example:
> : PrintTitle ( group sub -- )
> make-subgroup-title \ consumes group & sub, returns caddr and len or cstring
> somehow-invoke-the-word-that-text-represents \ returns caddr len
> type ;
> I can write make-subgroup-title, but do not understand how to write somehow-invoke-the-word-that-text-represents.
> Thanks for your time!

(1) If you use evaluate it will make the string into a new buffer.

(2) You could also ' or ['] a word and execute the token.

(3) ( addr u ) is preferred over (cstring )
{ where cstring means counted string as opposed to a c-string ending in '\0' }

(4) 600 is a lot of words.
If your groups and subgroups are fixed you could do a vectored execution.

On Monday, 27 March 2023 at 07:12:13 UTC+1, minforth wrote:
> minforth schrieb am Montag, 27. März 2023 um 07:44:53 UTC+2:
> > Van Kichline schrieb am Montag, 27. März 2023 um 02:08:07 UTC+2:
> > > I am a Forth novice writing an application with many fixed text strings of widely differing lengths, indexed by two dimensions (with several additional attributes, like title and date.) I am currently using gForth.
> > > I plan to create about 600 words like:
> > > Group1Subgroup7Title ( -- caddr len ) S" A Subgroup Title" ;
> > > But my question is: how can I programmatically execute a word from a string I've generated within another word? It seems like I need to put the word in the input buffer and call parse-word, but I haven't discovered how to do that. For example:
> > > : PrintTitle ( group sub -- )
> > > make-subgroup-title \ consumes group & sub, returns caddr and len or cstring
> > > somehow-invoke-the-word-that-text-represents \ returns caddr len
> > > type ;
> > > I can write make-subgroup-title, but do not understand how to write somehow-invoke-the-word-that-text-represents.
> > > Thanks for your time!
> > Try something along PARSE-NAME EVALUATE
> Another helper word to separate names within non-empty(!) strings:
>
> : PARSE-STRING ( a u -- a' u' ap up )
> BEGIN over c@ bl = WHILE 1 /string REPEAT
> over >r
> BEGIN over c@ bl <> WHILE 1 /string REPEAT
> over r@ - r> swap ;
>
> \ test
> s" fair may 2023 "
> parse-string cr type
> parse-string cr type
> parse-string cr type

two thoughts...
(1) doesnt test for reaching the end of the line.
(2) SKIP and SCAN may not be standard but most forth have them

Van Kichline <vkichline@gmail.com> writes:
>But my question is: how can I programmatically execute a word from a string I've generated within another word?

1) With SEARCH-WORDLIST and EXECUTE

2) With EVALUATE. Note that EVALUATE has some caveats, which may or
may not be relevant in your case
<http://www.complang.tuwien.ac.at/forth/why-evaluate-is-bad>.

- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
New standard: https://forth-standard.org/
EuroForth 2022: https://euro.theforth.net

On Monday, March 27, 2023 at 8:07:56 AM UTC-7, Anton Ertl wrote:
> Van Kichline <vkic...@gmail.com> writes:
> >But my question is: how can I programmatically execute a word from a string I've generated within another word?
> 1) With SEARCH-WORDLIST and EXECUTE
>
> 2) With EVALUATE. Note that EVALUATE has some caveats, which may or
> may not be relevant in your case
> <http://www.complang.tuwien.ac.at/forth/why-evaluate-is-bad>.
>
> - anton
> --
> M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
> comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
> New standard: https://forth-standard.org/
> EuroForth 2022: https://euro.theforth.net
Thanks, all! Evaluate is what I was looking for. 'S" Group2Part3" evaluate type' selects one of many string generators, which I can select by building the word to evaluate.

Rushing ahead yesterday, I built index arrays (tick word comma (repeat)) to look up my words. My total string data comes to ~580 groups and ~80K of text. I believe the indexes add about 5K in 64 bit gForth and save between 3 and 12 evaluations per execution of my program.
The point of this exercise was to be heavy on data and data selection, lighter on calculation. I found two very different approaches; constructing words for querying the dictionary, or constructing array indexes. I imagine I can economize more by using anonymous definitions in the arrays.
I thought leveraging the dictionary for lookups would be a good strategy, but on the order of 1000 items indexes seem simple and efficient. Perhaps if indexes consumed too much of available RAM lookups would be preferable.