It would have to be finstem, a simple command-line program I wrote to reduce Finnish words down to their root form.

Finnish is a lot like Latin or Russian in that its words often become lumbering behemoths of rewritten consonants, suffixes upon suffixes, and this makes it hard to look up in a dictionary – that is, until you factor in its very regular orthography and the phenomenal efforts of the Finnish programming industry: finstem is basically a very specialized UI for the OpenOffice spell checker, and I have no shame in admitting that.

<code>finstem.py</code> in action.

cloc puts it at under 200 lines of code altogether, including the tiny fzf wrapper I ship with it to let people stem words as they type them.

An example of the <code>fzf</code> wrapper.

I pretty much consider it feature complete by this point. I use the wrapper dozens of times in an average day while studying the language; there were a few bugs early on my dogfooding revealed, but the nice thing about small CLI programs is there just isn’t that much surface area for issues to crop up.