A Strava, for boats
The moat around marine hardware and software is disappearing, with individuals and small teams spinning up custom solutions. With a $50 development board and an afternoon with Claude, a motivated boater can now produce a working NMEA 2000 display. Our boats will be increasingly saturated with computers and agents. Hate that? . While APIs are clearly the answer to data sharing, what's not yet settled is where it's stored, and who gates access.
So far, it's fragmented. There are at least a dozen examples of pretty much the same thing — dump vessel and NMEA data into the cloud, save it, graph it, expose it in various ways: Yacht Devices, Actisense, NjordLink, Maretron, SeaSmart, Oria, Garmin OnDeck, SailLogger and plenty more. Xengineering of course has another, for now. But the chance that it, or any of these, survive the next five years is questionable.
Ideal would be an open, manufacturer-neutral, flexible mass database — something to accumulate and efficiently serve as much data as possible in the “yacht intelligence” category. The most interesting potential may be NMEA2k, already the de-facto standard for vessel intranet networks. They've been working on a cloud standard since at least 2024 and should be best positioned to provide this service, but, in their own words:
“This data set will be accessible by NMEA manufacturer, dealer, trade, and boat builder members for use. The data collected from this effort can also serve academia and the marine, conservation and research communities around the world to better protect marine life.”
— National Marine Electronics Association, on the NMEA Cloud
What a fully user-aligned cloud would look like
Not a critique of the list above so much as the standard worth holding any of them to, this one included.
- Free, or priced near the actual cost of storage, bandwidth, and maintenance, with little or no markup.
- Non-commercial — no ads, no products to sell, and no monetization of user data or attention.
- Open for reading as well as writing through the API, including queries across data that owners have chosen to make public.
- Machine-first and self-describing.
- Accurate and transparent about data sources.
- Data sharing controlled by the owner and fully revocable.
- Built according to best practices in data privacy and security.
- Portable and exportable.
- Stored locally first, then synchronized automatically whenever a connection is available.
- Built on public formats, protocols, and enough public code that someone else could stand up a compatible duplicate or replacement.
- Flexible, fast-responding, and efficient.
- Maintained by a responsible, neutral organization or AI with checks and balances.
X Engineering welcomes discussion about Cloud strategy. We will support whatever seems best for Open Source values.
Inspirations
Relevance puts open, free-to-read networks first; the ones that sell the dataset rank last.
Have an opinion on this?
Cloud strategy is not settled and we would rather argue about it in public than decide it quietly. The forum takes posts without an account — a name is the only field.