The Science projects help you package your interpreted applications as native executables.
These native executables are called scies (pronounced "skis") and generally include both a native
interpreter and your interpreted application code. Scies are built using a manifest you write that
describes the files to include in the scie as well as the commands to run when executing the scie.
Scies boot with the scie-jump native executable which inspects the environment, the command line
arguments and the manifest contained in the scie to determine which commands to run and which files
they need. The scie-jump then extracts the required files if needed and executes the required
commands.
You likely want to get started with science, the high-level tool for building scies that is
itself a scie built using science:
- Documentation: https://science.scie.app/
- Code: https://github.com/a-scie/lift
You can use the scie-jump directly to build scies, but it's more low-level than science. The
core environment variables and placeholders that power scie commands are all defined and implemented
by the scie-jump though; so its good to become familiar with these in its packaging docs:
- Documentation: https://github.com/a-scie/jump/blob/main/docs/packaging.md
- Code: https://github.com/a-scie/jump
Scies have the ability to be shipped "gouged-out". One or more files can be defined in the manifest
that are not included in the scie but are instead fetched on-demand and verified against a checksum.
The ptex binary does this fetching and fills the gouged-out scie back in:
If you have more questions about using or packaging scies, feel free to drop in to our discord server at https://scie.app/discord.