- From: Jon Radoff <jradoff@novalink.com>
- Date: Fri, 02 May 1997 09:06:14 -0700
- To: Larry Masinter <masinter@parc.xerox.com>
- CC: w3c-dist-auth@w3.org
> WEBDAV is trying to solve a different problem: a protocol > level interface to sufficent capabilities for new versioning-based > authoring applications to have direct access. Since we're > looking for interoperability for a new set of functionality > at the protocol level, we need standards for those new functions > if we're really going to have interoperability. As far as I can see, there is nothing within WEBDAV that can't be done already within the context of the HTTP server infrastructure. Therefore, I am at a loss as to why it is to anyone's advantage to produce a protocol level standard to provide the functionality. As you indicated, there are "dozens" of products that provide the functionality of remote document editing via the Web. Shouldn't we be concerned with producing standards for modularizing these capabilities and providing standard methods for making requests in to them while preserving the existing investment in underlying technology? I think any standard that simultaneously proposes to extend HTTP, require a rollout of a whole new generation of Web servers, and creates a new class of client application for such a narrow application as WEBDAV will fail as vendors continue to produce proprietary extensions in the form of server-API plug-ins (CGI, et al) and Java applets; customers are not willing to blindly leap to "standards" if it requires a significant change to the underlying infrastructure. In a very practical sense I think extensions to HTTP methods should be reserved for those things that are absolutely necessary - in other words, those things that just can't be done any other way. This could ultimately lead to having zillions of new HTTP method "standards" not just from WEBDAV but from all sorts of other standards efforts as well; some day you might have to go down a checklist of which HTTP methods your Web vendor supports. And of course, some vendors will start to make up their own methods. It just seems like there is an opportunity here to standardize on some robust, usable methods within the existing infrastructure and this talk of extending HTTP to support a whole new range of client apps and tools is like killing an ant with a sledgehammer. Some of the early discussion in this group appears to have been highly influenced by the Web server implementors who may see an advantage in doing it this way; I think for this to be a successful endeavor some of the ideas so far will need to be reevaluated with the input of other types of vendors as well as potential users. Jon
Received on Friday, 2 May 1997 09:05:31 UTC