- From: Paul Prescod <paul@prescod.net>
- Date: Fri, 8 Jan 2010 07:22:58 -0800
- To: Graham Klyne <GK-lists@ninebynine.org>
- Cc: Mike Brown <mike@skew.org>, uri@w3.org
On Fri, Jan 8, 2010 at 3:50 AM, Graham Klyne <GK-lists@ninebynine.org> wrote: > Mike Brown wrote: >> >> I'm not convinced 'file' really needs an RFC at all. The lack of specific, >> prescriptive guidelines aren't really impeding implementers, and 'file' URI >> interoperability in browsers isn't something I hear much clamoring for. > > That's not my sense. A lot of web applications, other than browsers, use > file: URIs as a way of accessing local resources alongside global ones. In > my past experience as a developer, a little more consistency in the various > ways in which file: URIs are constructed and interpreted would be really > helpful (e.g. handling drive names on Windows systems; exactly what > constitutes an "authority", etc.). I agree with both key parts. file:// is used by a lot more than just browsers. Browsers are actually the least interesting consumers of those URIs. If it were more reliable it could become a more general replacement for "raw file paths" which would in turn mean that URIs are allowed in many places that only file paths are. (e.g. configuration files) One could imagine a universe in which it is just as easy in standard APIs to open a URL as to open a file, and to use a file:// URL as to use a file path. But the current mess (on Windows in particular) does not allow that. If one were to take on this problem, you'd probably make a standard for how to interpret file:/// URLs on Windows and how to interpret them on posix-like systems, and that would cover the vast majority of users in the world. > The trouble is, it's a Hard Problem made complex by the heavy hand of > history. I agree. It is true that there is no single constituency who would care enough to push for vendors to update their handling to be consistent. Paul Prescod
Received on Friday, 8 January 2010 15:23:30 UTC