reflectionsFebruary 16, 20232 min read

A small thing, finished

An idea can get bigger long before it gets useful. A note on keeping the first version small enough to meet.

"A place to save links" sounds like a weekend project.

Then come folders. Shared folders would be nice. Automatic tags could save time. A public profile could turn the whole thing into a way to find people with similar taste. Every addition has a perfectly reasonable story, and together they keep pushing the first useful moment further away.

So I try to keep that moment in one sentence. For a bookmark tool it's this: save a link, close the tab, come back tomorrow and find it again.

First versionno open questions
  1. Paste a link
  2. Save it
  3. Find it tomorrow
    Add something for later
    One useful thing, then the extrasAn imagined bookmark tool. Each extra arrives with a question somebody has to answer.rows enter 200ms, 4px drop · card resizes 240ms · ease (0.22, 1, 0.36, 1)

    Plans never push back

    An unbuilt idea is comfortable because nobody has had to use it yet. The sharing works beautifully in the sketch. The automatic tags never guess wrong. The public profiles are full of interesting people. Next to all that, a page with three saved links and an awkward empty state looks thin.

    The thin page is the one that teaches you something. Maybe the title is enough and folders never matter. Maybe every link needs a note about why you saved it. Maybe you never come back at all. None of that shows up in a more elaborate plan.

    What can wait

    Martin Fowler's note on YAGNI draws a line I use a lot. Capabilities you're only guessing at can wait, while the work that keeps the code easy to change can't. YAGNI is no excuse to skip tests or leave a mess for the next change.

    For the bookmark tool, links that survive a reload are part of the first promise. How a future team shares thousands of collections can wait until somebody actually asks.

    A few decisions really are expensive to reverse, and they deserve thought on day one. "Keep it small" is a way to question the plan. It doesn't cancel what the work needs.

    Give it an ending

    A small project needs a finish line you can reach. Write down the task, then do it start to finish, boring parts included: the empty list, a broken link, a title too long for its row.

    When that works, let the thing exist for a while. Use it. Hand it to someone else. Keep a separate list for the ideas that show up afterwards, so they don't all turn into today's unfinished work.

    It may never become the bigger product you pictured. It may just keep doing the one job you needed, and coming back to something small that already works is one of my favourite parts of this job.