On Wed, May 18, 2011 at 3:32 PM, Alejandro Imass <[email protected]> wrote: > On Wed, May 18, 2011 at 2:50 PM, John M. Dlugosz <[email protected]> > wrote: >>
[...] > Yeah, let me re-phrase. The REST URLs must look more like paths on a > drive rather than an API. The URL parameters can be used freely (and > must be used in many cases) as extra information for the GET request. > For example, for filtering a collection, for conditioning the result, > the presentation and in general for _conditioning_ the GET. > I want to point out that the OP only referred to the GET URL. Furthermore, let's remember that in HTML 4 forms are a usually representation of a resource (the representation is actually HTML but the form usually presents data to the user). In other words, forms in HTML 4 are the only way to serialize data, or to represent data in the HTML doc. XHTML offers more, but in the end most interactions in Web 1.0 and Web 1.5 (just to give a name to Web 1.0 + AJAX), are done with the html form artifact. This limits you to GET and POST for the action method of the form. anything you want to do RESTful (PUT, HEAD, DELETE) will probably require the use of JavaScript and some other serialization scheme (XML or JSON, I prefer the latter for many reasons). So then your resources as such would be invisible to your browser URL, maybe or maybe not depending on your far REST you want to go. For full Web 2.0 REST your URL in the browser would probably point to the application URL, that is an HTML Shell with a lot of JavaScript. The interactions from there on will be AJAX from JS and your URL in the browser would probably not change that much and your resources will be XML or JSON. Beleive it or not this actually more "Web"/REST that what we have today! Best, -- Alejandro Imass _______________________________________________ List: [email protected] Listinfo: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/[email protected]/ Dev site: http://dev.catalyst.perl.org/
