Re: Identification of documents in Web applications

Hi Noah,

On Mar 8, 2011, at 11:11 PM, Noah Mendelsohn wrote:

> I've posted some thoughts at [1] on the identification of documents in Web applications.

With regard to your two main points:

> Use of AJAX implementation technology is not a sufficient excuse for failing to provide first class URI identification for documents on the Web

As a Web developer, I don't want an "excuse", so much as I want to do the right thing to make my Web application be successful. I can (as a Web developer) surely be forgiven for following the latest hot software trends as long as my customers like the results, no?

What are the benefits to my application of modeling the application resources this way? I think you've mentioned or hinted at some in your blog post, but there are probably some more - here's a partial list that I can think of (and I think I'm only roughly restating in different language things that are said already in both your post and Jeni's)

i) The application will be more usable by more people. The idea of "progressive enhancement" is probably relevant here.
ii) Because you will make fewer assumptions about "other people's code", your Web resources are less likely to depend on other people's code, thus your application is generally more robust in the face of unexpected failure that may be out of your own control. 
iii) Your application will be more open to "network effects" - use, that you have not explicitly intended, of parts of your application which are more decoupled from each other. 

> Where practical, model your application as a collection of documents, each with its own URI

There are benefits to this model, and I think they should be clearly described and "evangelized". I think there are probably the same benefits even in the simulator case if the simulator is consuming Web resources which have their own URIs, and also in a case where the Web application can expose relevant URIs for its own (client-side) resources.

Finally, with regard to '?' vs. '#', to me '?' provides arguments that are interpreted relative to the server-side namespace - the names are proposed and interpreted by the server and sent by the client. In the '#' case, the names are provided by the document (type) creator (which may not be the server) and are interpreted by the client. I think there is a case for using both, in different circumstances. Under what circumstances should I use '#' parameters in a URI? What is needed to make this work with media types in a reasonable way?


- John

> Noah
> [1]
> [2]
> [3]

Received on Wednesday, 9 March 2011 16:09:05 UTC