In the Basic User Manual doc it states:

“The file RAWDATA/skeleton.top has a bunch of topics already predefined with
keywords but no responders or gambits. If you filled in some of these topics with rules
and hooked the file into filesHarry.txt and rebuilt the data, you’d have a chatbot.”

When you say “hooked the file” am I correct in understanding that just means add the line

RAWDATA/skeleton.top

to the filesHarry.txt in this instance?

Or, alternatively, just dropping it into the HARRY folder does the same thing, does it not?

regards, Richard

You can just drop it in the harry folder, provided it had no topic name or concept name conflicts (given its age, I can’t say).
Or you could cut and paste topics from it into a harry file. Nowadays I recommend a separate file per topic for clarity and maintenance.

I like the separate file per topic, or at least large-ish topics, myself. As I understand the operations of ChatScript, it reads/searches every file in the folder designated in the files*.txt top level file yes? Is there any impact on speed running one file vs many? What exactly sits in memory with the engine when you execute CS?

regards, Richard

Compiling your rawdata source files (:build xxx)  is what reads those files and converts them into internal format which is stored in TOPICS folder. The engine thereafter only reads that TOPICS folder when it starts. There is no speed issue in having source in multiple files. 

The engine loads up the DICT binary files, the TOPIC files, the LIVEDATA files. 
It reads in user data (USERS) each volley (topic state and facts).