|
Wow, so much confusion in this article, it's hard to know where to start. Let's see...
1. The whole thing about WebDAV is confusing. Why would we want to try to be compatible with a protocol nobody uses?
2. The API will not be deferring authentication mechanisms to arbitrary other standards. The supported authentication mechanism(s) are currently under discussion. WS-Security is one of many proposals.
3. The whole confusion with the API being "tightly coupled" with the syndication format. WHO CARES if the site has a syndicated feed or not? The API defines a way for authors to add/edit/delete/search content on their site. Your questions about "well what if I don't want a syndicated feed" just don't make any sense at all. The conceptual data model is similar, and so the serialization is similar. That's all.
4. "Since the entries can be considered a single XML document" is just laughable on its face, and is another sign of category error. Just because we use XML in the API doesn't mean that we're mandating it for a storage format. XPath makes no sense at all for this application. All the personal CMSes I'm aware of either store content in a series of flat files, or in a low-end database. Neither supports XPath. Mandating an XPath-based query interface would put this waaay out of the realm of implementability for 99% of the target audience.
5. As Joe pointed out, RSD is being dropped in the next draft.
6. I think you are also confused in thinking that a REST approach mandates *separate* URLs for each action. This is a common misconception. The URLs for new post, edit post, delete post, and get post can all be the same URL of a single script (such as a CGI), with no different parameters or anything. The CGI script dispatches based on the HTTP verb used. See http://bitworking.org/news/Carrot_Versus_Orange for an example.
7. Your entire argument in favor of SOAP seems to be based on the level of indirection introduced by RSD (soon to be nixed) and several misconceptions, such as that you need separate URLs for each action (not true).
In short, this entire article is poorly argued. Where it is not poorly argued, it is dangerously misleading, and where it is not misleading, it is simply wrong.
|
Previous Message
