Autoscalable Programming Language
🔭 Predictable runtime for all computations - A program is represented as DAG(s) where the running time for all computations can be predicted because there is no unbounded recursion or iteration.
⛓ Automatic IO concurrency and parallelism across events and arrays - Alan exploits opportunities for IO concurrency or CPU parallelization across machines in a cluster via arrays and a static event loop without threads, channels, promises, futures, locks, etc.
✅ Almost no runtime errors - No deadlocks, livelocks, undefined variables, divide-by-zero, integer under/overflow, array out-of-bounds access, etc.
👩🚀 Alan is a programming language that does concurrency for you and can thus separate how the software is written from how it runs. To learn more about Alan, take a look at runnable examples or the most Frequently Asked Questions.
For MacOS it is recommended to install Alan via the Homebrew package manager.
brew install alantech/homebrew-core/alan
For Linux and Windows it is recommended to install Alan via the published artifacts. Simply download the zip or tar.gz file for your operating system, and extract the
alanexecutable to somewhere in your
$PATH, make sure it's marked executable (if not on Windows), and you're ready to roll.
wget https://github.com/alantech/alan/releases/latest/download/alan-ubuntu.tar.gz tar -xzf alan-ubuntu.tar.gz sudo mv alan /usr/local/bin/alan
Invoke-WebRequest -OutFile alan-windows.zip -Uri https://github.com/alantech/alan/releases/latest/download/alan-windows.zip Expand-Archive -Path alan-windows.zip -DestinationPath C:\windows
To compile to Alan GraphCode and then run it with the AVM:
You can also compile-and-run a source file with a simple:
Note: To better understand if we are building something people want to use we currently log an event when running an Alan command. Feel free to turn this off by setting the
ALAN_TELEMETRY_OFFenvironment variable to
true, but if you do please let us know how you are using Alan and how often!
If you wish to contribute to Alan, or if your operating system and/or CPU architecture do not match the above, you'll need a development environment to build Alan locally:
Once those are installed, simply:
git clone https://github.com/alantech/alan cd alan make sudo make install
Integration tests are in
/bddand defined using Shellspec. To run all integration tests:
To run a single test file:
make bdd testfile=bdd/spec/001_event_spec.sh
To run a single test group use the line number corresponding to a
make bdd testfile=bdd/spec/001_event_spec.sh:30
The Alan Programming Language is made up of multiple sub-projects housed within this monorepo. Each subdirectory has its own license file and the project as a whole uses two licenses: The Apache 2.0 license and the Affero GPL 3.0 license, with the breakdown as follows:
The AGPL 3.0 license requires that any changes to the code are published and publicly accessible. This is to make sure that any advancements to the compiler and AVM are available to all. The licensing of these tools does not affect the licensing of the code they compile or run. Similarly, the GCC compiler collection is GPL 3.0 licensed, but code compiled by it is not GPL 3.0 licensed.