NDRC - Next DAAD Reborn Compiler released

NDRC compiles DAAD Adventure Writer source (.DSF) into a DAAD database (.DDB). It is one self-contained executable with no runtime, no interpreter and no companion installation, for Windows and Linux.

It is a from-scratch C port of DRC, Uto's DAAD Reborn Compiler (a FreePascal front end and a PHP back end), verified byte-identical against DRC's own drf+drb pipeline. Uto's DRC is the reference implementation: https://github.com/Utodev/DRC

NDRC was made to streamline the toolchain of NextDAAD and provide a NextDAAD specific target supporting larger databases than other ZX targets. It is included with the authroing kit, so there's no need to download seperately.

The compiler is also compatible with all standard DAAD targets and can be used as a substitute for DRC.

ndrc NEXTDAAD EN game.DSF

This compiles game.DSF for the NEXTDAAD target, English language, and produces game.DDB in the same directory.

ZX, NEXTDAAD, CPC, C64, CP4, CPM, MSX, MSX2, ZX81, PCW, PC, AMIGA, ST, HTML.

ndrc NEXTDAAD EN game.DSF -auto-tokens

NDRC v0.2+ offers enhanced text compression compared to DRC, enabling DAAD DDBs to fit more game text into the hard limit database sizes. This is compatible with existing DAAD interpreters, just requiring an extra CLI flag when compiling your game with NDRC.

DRC compresses text with a fixed token table per language - the same 128 English digraphs for every English game ever compiled. With -auto-tokens, NDRC instead selects up to 128 tokens from the compiling game's own text and encodes the text with an optimal parse over that table. Compressed text measures 8-14% smaller than the builtin table across a corpus of real games, with the biggest wins on prose-heavy games near the 64K database ceiling. Selection is data-driven, so it adapts to any of the five languages - and to your game's actual vocabulary - automatically.

The output DDB is format-identical: it decodes on every DAAD interpreter, on every target, including the original 8-bit interpreters. Without the flag, NDRC's output remains byte-identical to DRC, as always.

Add --tok to also write the selected table as a standard .tok file beside the source. The normal override lookup picks it up on the next run and reproduces the same DDB byte for byte, and the same file works in stock DRC.

Compared with DRT, the existing per-game tokenizer for DRC: on the same corpus of real games, DRT's tables came out roughly even with DRC's builtin English table, because its greedy selection splits the text on each chosen token and locks out the short high-frequency fragments that carry a good table. NDRC scores every candidate against an optimal parse of the whole text instead, which is where the 8-14% comes from - and it runs inside the compile, no separate tool or Python needed. DRT-generated .tok files remain fully supported as overrides, unchanged.