You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Every project in this collection is written on the same machine, in the
same terminal, against the same set of tools. This site documents that
layer. Not what to build, but what you build it with.
Pages are written the way you would actually want to read them at the
moment something breaks: the mental model first, then the commands, then
the gotcha that cost an evening. Several are explicitly comparisons,
because the useful question is usually which of two tools rather than
how one works.
For the physical machines see Hardware
Tools, and for containers and
services see IaaS.
The broad tour: what a shell is, file and text manipulation, redirection and pipes, permissions, process and system management, networking, packages, environment variables, and archiving
The longest page here. User types and management, the files behind them, UIDs and GIDs, groups, the permission bits, switching users, and session management
GNU make properly: the core mental model, special variables, phony targets as the number one gotcha, the fact that each recipe line runs in a new shell, the four assignment operators, includes, pattern rules, and the functions you reach for constantly
A command runner that looks like make but is not a build system, because it does not do timestamp-based rebuilds. Recipes, dependencies, arguments, dotenv behaviour, portability, and a direct comparison against make
The platform end to end: what it is, requirements, getting started, storage and partitioning, users, GPU passthrough, cloud-init, device discovery, and maintenance
All ten backends compared, directory through LVM, LVM-thin, ZFS, Ceph, NFS, CIFS, iSCSI, and GlusterFS, with what each is actually for. Worth reading before creating a guest, because the choice is hard to undo
What an AST is and why “abstract”, how one is produced, traversal patterns, where they show up in real tooling, and how they differ from related representations
The non-UML notations that most architecture questions actually want: the four C4 levels, data flow diagrams with a threat model on top, and network topology
Multi-factor authentication end to end: the three factors, methods ranked by security, the attacks against each, how MFA differs from 2FA and passwordless, cloud platform support, and compliance
The library that gives an LLM control of Chromium by serialising the page into indexed interactive elements and looping. Architecture, API surface, model choice, the commercial layer, where it earns its keep, and the gotchas
Not Covered Yet
Git has no page, despite being the tool used most. Version control
is covered from the GitHub side in Web
Development.
No editor configuration page. Vim and Neovim are covered as tools,
but the actual dotfiles and plugin setup are not written down
anywhere.
Nothing on ssh and remote work as a subject, only the key setup
buried in the new-computer checklist.
The two sound testing pages duplicate each other and should be
merged.
Linux/04_Bash_Scritps.ipynb is misspelled in the filename, which is
why the URL looks wrong.
About
Developer tooling notes: Linux administration, tmux, vim and neovim, Proxmox, JupyterLab, Quarto, LaTeX, make and just, regex, C++, PlatformIO and software diagrams.