Proxies and extensible clients
Daniel LaLiberte <[email protected]>
| Newsgroups | gmane.ietf.url |
|---|---|
| Message-ID | <[email protected]> |
Martin J. Duerst writes: > Not necessarily. With the number of URL schemes and URN namespaces > increasing, as well as the complexity of other things the client has > to do, it might be a very good implementation strategy to delegate > all the URL/URN resolution stuff to a *local* server (proxy). Yes, yes! I was going too far. By "client", I sometimes metaphorically mean things in the neighborhood of the client which act as sort of extensions of the client. A local proxy or Java applet that understands a scheme works just as well as, if not better than, explicitly cramming the hardcoded understanding of all schemes into one client. In fact, I am very interested in how a Java applet API can be done to avoid the security concerns (accessing arbitrary new servers). The main problem with new URI schemes is deploying them to millions of clients. Even proxies would have to be educated, so auto-downloadable code that understands new URI schemes and associated protocols would grease the rails on the bridge to the future. dan