Benefits of Uniformity
Jon Udell: These are, of course, best practices for an ecosystem sustained by web standards like URI, HTTP, and XML. But it was wonderful to see those best practices clearly demonstrated in a PDC keynote.
In many ways it feels like we are now about at the point in the evolution of the web where we were two decades ago with respect to GUIs. For that reason, it pleases me greatly to hear that Don Box was on stage demonstrating the value of Hi-Rest. And it pleases me that Atom/AtomPub played a role in this.
On the off chance that anybody is reading this, there is considerable value to be obtained from Lo-Rest. But it comes at a great cost. And the way forward is not to have servicebus.windows.net and user-ctp.windows.net/V0.1/Mesh/MeshObjects using different interfaces. Instead it means adopting HTML as a carrier wave for your signal — which, yes, implies that you need to employ the modern equivalent of screen scraping. And have to deal with cookies. HTML is far from uniform: there are a lot of different tag soup parsers out there at various levels of quality. Having everybody parse HTML consistently would be a huge step forward. At the present time, HTML5 is the most comprehensive and complete approach to this. Having Microsoft more actively participate, and spinning off and fast tracking just those parts of HTML5 necessary to implement this part of the HTML5 vision; well, that would be a huge win.
And there are Hi-Rest analogs to this. I’m confident that the service documents and feeds produced for the PDC demo were all well-formed, and valid. As more people try to emulate this with PHP and templates, this won’t be true quite as much. The first time a “smart quote” or a non-Latin1 character gets into the stream, it will fall over. Having everybody recover in the same manner would also be a big win, though not quite as huge as it is for HTML. Anne calls his explorations into this topic XML5.