asdf for Multiple Tool Versions

asdf manages versions of multiple tools.

Install

asdf is a unified version manager for many languages and tools. Where nvm handles Node, pyenv handles Python, rbenv handles Ruby; asdf handles all of them. The discipline is one tool for all version management.

What installation looks like:

• asdf install python 3.11.5.: The command installs a specific Python version. The version is downloaded and built; available for use; the discipline accommodates per-project Python versions.
• asdf install terraform 1.5.0.: Similarly for Terraform. Different Terraform versions for different projects; asdf handles them; the discipline scales beyond languages.
• Per-project versions via .tool-versions file.: Each project has a .tool-versions file. The file specifies which version of each tool the project needs; entering the project directory activates the right versions.
• Many plugins.: asdf supports many tools via plugins. Languages, build tools, infrastructure tools all are available; the team's discipline scales as new tools join.
• Versioned in git.: The .tool-versions file lives in the project's git. Engineers cloning the project get the version requirements; the discipline transfers.

Installation is bounded. The team's investment in asdf scales across all version-managed tools.

Switch

asdf supports global, local, and shell-level version selection. The discipline matches different needs.

• asdf global terraform 1.5.0 sets global default.: The default applies when no other version is specified. Engineers' baseline is the global; project-specific overrides apply on top.
• asdf local sets per-directory.: Within a project directory, asdf local sets the version. The .tool-versions file is updated; subsequent commands in the directory use the local version.
• asdf shell for current shell only.: The shell-level setting is temporary. Useful for testing different versions; the discipline is bounded.
• Hierarchy.: Shell overrides local, which overrides global. The discipline is layered; the most specific applies.
• Tab completion.: asdf supports tab completion. The engineer's productivity is preserved; common operations are fast.

The switching discipline is per-need. Different scopes for different cases.

vs others

asdf replaces multiple per-language version managers. The unified discipline produces operational simplicity.

• nvm for node.: nvm is the popular Node version manager. asdf can replace it; the team uses one tool instead of many; the discipline is simpler.
• pyenv for python.: pyenv handles Python versions. asdf replaces it similarly; the team's tool count drops.
• asdf unifies.: The unified approach reduces the team's tool count. One tool to learn; one tool to configure; one tool to keep working; the discipline scales.
• One tool to rule the version managers.: The team's discipline becomes simpler. New engineers learn asdf; old engineers migrate from individual tools; the consistency improves.
• Migration is gradual.: Teams using nvm/pyenv/etc. can migrate to asdf gradually. New projects use asdf; old projects keep their existing tooling; the migration is bounded.

asdf for tool versioning is one of those development tooling disciplines that pays off in unified version management. Nova AI Ops integrates with development tooling, surfaces patterns, and supports the team's productivity discipline.