Listed Is Shutting Down. Cloudline Press Was Built for This.
Standard Notes announced that Listed will shut down Dec. 31, 2026.
For anyone using Listed as a public website or blog, the consequences are straightforward. Public blogs, previous posts and guestbooks will no longer be accessible. The posts themselves will remain privately stored in Standard Notes, but the public publishing layer will disappear.
This creates a problem for one of my projects.
It also provides a remarkably good test of its architecture.
Why I Was Using Listed
I use Standard Notes for writing and have used Listed as a convenient way to publish material from it.
When I began building blackie.black, my personal professional website, I initially considered simply using Standard Notes and Listed for the whole thing.
That would have been much easier.
But I wanted more control over the finished publication than Listed provided, particularly over page structure, navigation, typography and printing. I also wanted the published material to exist independently under my own domain rather than requiring another service every time someone requested a page.
That led to what became Cloudline Press, an Apache-first reference architecture for calm, durable, print-aware web publishing.
blackie.black became its first working implementation and proving ground.
The Important Architectural Decision
Cloudline Press does not serve pages from Listed.
It uses Listed upstream during publication.
The original process looks roughly like this:
Cloudline Press retrieves the public material, normalizes it and builds complete local HTML documents.
Those documents become part of an immutable edition of the website. Apache serves those files directly.
There is no request-time connection to Listed. There is no application server asking Listed for the current version of a page. There is no database, runtime template engine or JavaScript application reconstructing the site in a browser.
Once an edition has been published, it belongs to the local publication.
That distinction suddenly matters quite a bit.
What Happens When Listed Disappears?
On Dec. 31, Listed can disappear and the existing Cloudline Press site should simply keep working.
Apache still has the HTML.
The CSS is still there.
The navigation is still there.
The documents are still there.
The current edition does not know or care whether the upstream service from which some of its source material originally came still exists.
That was an intentional property of the architecture, although I did not expect to receive such a substantial real-world test of it this quickly.
The part that breaks is the next build.
Cloudline Press currently has a Listed-specific synchronization process. Once Listed is gone, it will no longer be possible to retrieve newly published material that way.
That means the publishing boundary has to change.
It does not mean the serving architecture does.
I Don't Think Listed Needs a Replacement
The obvious response would be to find another blogging service and substitute it for Listed.
I don't think that is the right solution.
Doing so would leave essentially the same architecture in place with a different company's name in the middle.
Instead, this is an opportunity to remove that dependency altogether.
Standard Notes can continue doing what it does well: being the private writing environment.
Cloudline Press can take responsibility for publication.
The resulting model should be closer to this:
Exactly how that source boundary should work is now an engineering problem to solve.
The important requirement is that it remain portable and understandable. Cloudline Press should not require a particular hosted blogging service simply to get Markdown from an editor into a build directory.
Listed Becomes Part of the History
I am not planning to pretend Listed never existed.
The Listed synchronization code is part of Cloudline Press's development history, and I intend to preserve it.
Before the shutdown, I also have an obvious archival task: make sure everything I want from Listed has been captured locally, including useful metadata that may not exist once the public service disappears.
Eventually, the old Listed ingestion code can remain with the project as a legacy importer and historical artifact rather than an architectural dependency.
That history is useful because it documents how the system evolved.
Cloudline Press did not begin as an exercise in designing the theoretically pure static publishing system. It grew out of an actual publishing problem.
Listed was part of that problem and part of the original solution.
Now it is going away.
The architecture has to accommodate that fact.
A Useful Failure
One of the principles behind Cloudline Press is that publication and serving should be separate events.
A build can fail without damaging the published site.
An upstream source can become unavailable without taking the published site with it.
A new edition can be rejected without changing the current edition.
An old edition can be restored by moving a filesystem pointer.
Those ideas sound somewhat academic until an upstream service sends an email saying it will cease to exist in a few months.
Then the distinction between content I published and a service displaying my content for me becomes considerably less abstract.
Listed shutting down does not prove that every website needs Cloudline Press. It does demonstrate the particular failure mode Cloudline Press was designed to contain.
The upstream disappeared.
The publication did not.
Toward 2027
There is another reason the timing is interesting.
Cloudline Press takes part of its identity from Henry M. Burt's Among the Clouds, first published from the summit of Mount Washington in 1877. The project is intended as a modern adaptation of that idea: a printing and publishing system operating in a very different kind of cloud.
July 2027 marks the 150th anniversary of Among the Clouds.
I had already intended to use the time leading up to that anniversary to test Cloudline Press, find weaknesses, improve the documentation and harden the reference implementation before giving it a more formal public release.
Listed has now supplied a particularly useful test.
By Dec. 31, 2026, Cloudline Press needs a publication path that no longer depends on it.
By July 2027, the goal is a documented reference implementation at cloudlinepress.amongthe.cloud that another experienced operator can examine, deploy and understand.
The architecture will be better for having lost one of the services that helped create it.
There is something appropriately old-fashioned about that.
A publisher should be able to keep the publication when the press it was made on is gone.