Explore Transport Proxy

Can we transport a page that will proxy back to the transporter for more pages?

We launch our first image transporter on localhost:4010 and then link to it here with surgery to the page json.

We list several transporters offered by this foreign server. This is a static list maintained while constructing the transport endpoints.

We skip this step by bootstrapping into federated wiki with a remote link to this same page on the minimal foreign home page. localhost github

<p><a id=link target="_blank" href="http://ward.asia.wiki.org/">details</a></p> <script> link.href += location.host + "/welcome-visitors" </script>

# Provenance

Can we keep transporter information in a transported page such that an eventual Future plugin can offer to perform further transports? We experiment with a wikipedia link transporter which we will return to as a proxy. github

Transporter ⇒ Create ⇒ Refresh

Click ⇒ Search ⇒ Future ⇒ Transporter

We handle the first part of this two step process in the transporter itself. The click handling logic in core-javascript handles the second by searching the dom for the page with the clicked link. From this the Create action is found and forwarded to the Future plugin.

The search mimics the logic used to discover proper capitalization which is known to fail on occasion. Can we take into account all of the link paths present in the core javascript. See Render and Follow Links

We've duplicated code from the Transport plugin and more from the Template machinery to prototype an extension to the Future plugin that could (but doesn't yet) exploit information we leave in the Journal. github

To participate as a proxy a transporter would have to handle link titles as well as dropped urls. We've expanded the transporter post payload accordingly. github

We've further extended the proxy payload with the Create action from the previous transport. This can contain arbitrary information that supports complex navigation within a family of transporters.

We see problems when wikipedia returns redirects. We don't have and don't want similar centralized functionality.

# Importer

We experimented briefly with transporting a collection of pages and returning them through the latter half of the export-import mechanism. github

Importer plugins can be saved so long as the page doesn't exceeds the 5MB capacity of a server PUT.

The importer's rendering includes each page's age which is often just milliseconds having been just then constructed.