- From: Mark Nottingham <mnot@mnot.net>
- Date: Tue, 16 Sep 2008 06:42:10 +1000
- To: Mike Schinkel <mikeschinkel@gmail.com>
- Cc: "'URI'" <uri@w3.org>, "'Joe Gregorio'" <joe@bitworking.org>, "'David Orchard'" <orchard@pacificspirit.com>, "'Marc Hadley'" <Marc.Hadley@Sun.COM>, <fielding@gbiv.com>
Most of my use cases are for doing things like putting them into headers and content so that people can build new/dynamic protocols for machines with them. Making them more friendly for humans to read is a side effect; what I'm interested in is making them easier/more intuitive for humans to *mint*, because they'll usually be the ones writing them (just as with HTML, although I think the collective crowd of authors will be a bit more technical, but only a bit). My observation is that for those protocol-building cases, it's usually a light integration; someone needs to create a URI with a particular piece of data in it, and they don't want to constrain its form, so they need a template. They don't need list data or complex operators, or if they do, they can specify some pre-processing. I don't disagree that there may be a place for "whiteboardable" templates, but it's a secondary use case for me. Cheers, On 16/09/2008, at 4:26 AM, Mike Schinkel wrote: > Roy gave me his comments in response to my fervent advocacy for URI > Templates to be optimized for humans. Mark it sounds like you have a > similar > position to mine which is there is a strong needs for a very human > friendly > URI Template specification. I envision many uses for URI Templates in > Content Management Systems and Application Frameworks but I think Roy > envisions using them in Apache, routers, proxies, etc. > Unfortunately, I > don't think any URI template specification will come to a resolution > as long > as there are those competing and evidently non-compatible objectives. -- Mark Nottingham http://www.mnot.net/
Received on Monday, 15 September 2008 20:42:54 UTC