From a domain phrase to useful feedback
Under the proposed reuse model, a suitable business-related language would start with the common CoreLang base and add its own domain words. A new CoreLang parser would recognise the source and create a structured record of its meaning. Validation would then check the business rules and return a precise message to the author.
An editor sends document changes to DSLSH. DSLSH keeps the parser up to date and returns structured language information. THE CREXX Edition demonstrates that connection in a working editor and produces parser-backed HTML and TeX renderings. Its browser frontend is a working proof of concept with documented limits, not a production hosted editor.
Execute only when the language needs to run
When an example needs to execute, the proposed first path is to translate its
checked meaning into readable RXAS. The existing rxas tool checks, optimises
and assembles that source. rxlink combines modules, and RXVM runs the
bytecode. Reusing these tools avoids building another runtime and does not make
the source language a REXX dialect—a language variant that follows REXX syntax
and behaviour.
The language’s business meaning remains its own. A descriptive language should not be forced to execute, and a language that does not fit the CREXX execution model should use another path. LLVM remains later backend research after the first CoreLang and RXAS examples work consistently.