Meet people where they already are

This is the principle that shapes most of OHM's design, and it's worth stating plainly because it explains a lot of otherwise odd decisions.

The alternative approach is to define the correct format and require everyone to adopt it. It's tempting, because it makes the software much simpler. It also doesn't work. Open hardware has no central authority that could compel adoption, and a system that delivers nothing until the world reorganises itself around it will be waiting a long time.

So OHM absorbs the mess instead of legislating against it. That shows up in two places.

Designs: read what people already publish

Open hardware projects describe themselves however suits them — repositories, wikis, PDFs, spreadsheets, structured manifests, prose. OHM's job is to read what exists and translate it into something consistent, rather than to insist on a format before it will help you.

This is most of the work. See the problem.

Workshops: read the directories that already exist

There are already people who have done the difficult, unglamorous work of mapping where workshops are: Maps of Making, and others we intend to add. Rebuilding those from scratch would be wasteful and slightly insulting.

So OHM reads them, and shows a union of what it finds, with the ability to filter by source. A workshop that appears in one of those directories is already in OHM without having done anything or agreed to anything with us.

It's worth being precise about how that data works, because it's easy to assume the worst. In the case of Maps of Making, each space publishes a small file on its own endpoint, which the map reads. The information belongs to the space and is hosted by the space. The map — and OHM — are readers.

So this is not a case of workshops being scraped into someone else's database. It's a case of workshops publishing about themselves, and more than one thing being able to read it. That's the arrangement we want to support and extend, not replace.

The honest cost of this approach

Reading what already exists means inheriting its limits.

If a directory records what a space is interested in rather than what it can produce, OHM inherits that gap and cannot invent the difference. If a design's documentation never says what it's made of, no amount of processing will determine the material.

That is the real constraint on how well OHM works today, and it can't be engineered away — it can only be addressed at the source, by workshops describing themselves more fully and by the directories that carry that information being able to express it.

Which means the useful work is often not in our code at all. It's in improving shared vocabularies, contributing back to the projects whose data we rely on, and making it worth a workshop's time to say more about itself.