Journals/Essay
Two webs
The World Wide Web is for documents you retrieve. The Real Time Web is for a lamp that is still on. Same Internet. Two jobs.
Thiru · 23 September 2026 · 8 min read
There are two kinds of web on the same Internet. They are not the same job. One is good at a page you ask for. The other is for a lamp that is still on.
The World Wide Web is built for documents. You click. A server answers. You leave. That is right for a news article. It is a poor fit for “tell me the second the coffee maker finishes.”
The Real Time Web is built for live things: sensors, machines, applications, AI agents, and people using those resources. The relationship stays open. Authorized updates can arrive as they happen. The link works both ways.
RTW is not meant to replace the WWW. A site can still publish documents. RTW can carry the live channel beside that site.
A lamp on the network
Take a smart lamp. It has a light and a switch. Treat it as an : the actual network resource. The walkthrough below builds the vocabulary as that lamp participates.
Identity: the ghost
When the lamp can participate, it is paired with a : a shared-memory stand-in for that one resource. What the lamp reports, the ghost can carry. What the ghost is authorized to receive, the lamp can act on.
The ghost is tied to the resource, not to a human avatar. You are not the ghost. You may own the .
The pairing is . One lamp, one ghost. One ghost, that lamp. Never one ghost shared by two endpoints. The ghost provides identity, behavior, and context when an active connection is present. That is different from a stored twin that sits in a database whether the lamp is on or off. RTW does not natively store data in the network.
A home, then a morph
The lamp still needs an application-layer home. That home is a : a service on that allocates ghosts.
When the service is created, the service owner is the primary owner of those ghosts. When you register and confirm a transfer, you own the data layer. The service owner keeps the application layer. They may keep a local copy.
Suppose a scheduler wants the lamp at dusk. That service does not need a custom API into the lamp. The owner the lamp ghost against the secondary service. Two things happen at once. The scheduler gets read access, which the owner can withdraw. The data is reshaped into the scheduler’s local model. Each service only sees its own context.
Who is allowed to see it
Data is consumed at an endpoint when the ghosts at the ends are owned by the same user, or the receiver has approved read access. Same user usually means one person and two of their endpoints, such as their lamp and their display. It does not mean two different people are the same person.
Two different people do not open each other’s data because they share an implementation. One of them morphs, or transfers. Ownership sits on the ghostId. A transfer has to be confirmed. That is a change of ownership, not a casual share.
Underneath is Morphic Architecture Design, introduced in 2008 by Paal Kristian Levang. Four independently owned layers. A change in one layer does not have to redefine the whole relationship.
Today, two live things usually meet through a custom API, a login, and a copied dataset. The architecture’s claim is smaller: keep data with the source, morph it for whoever is authorized, and let ownership stay split.