W3C home > Mailing lists > Public > uri@w3.org > May 2005

Re: Converting filenames to file: URIs

From: Graham Klyne <GK@ninebynine.org>
Date: Fri, 13 May 2005 12:55:37 +0100
Message-Id: <>
To: Frank Ellermann <nobody@xyzzy.claranet.de>, uri@w3.org

At 13:31 13/05/05 +0200, Frank Ellermann wrote:
> > What's missing here?
>My browsers also eat backslashes and do the mapping
>on their own.  Lynx may treat file:/// as ftp:///,
>so to be sure I always use file://localhost/

I think you touch here an important issue that I'd like to tease out, if I 
can.  My postings on this topic, both recently and from some time ago, have 
all related to writing *applications* that access files and/or web 
resources.  In these cases, the issue of what a (presumed interactive) 
*browser* does with the URI doesn't arise.

To the extent that any discussion of filename/URI correspondence impacts 
browsers, is it fair to say that it is most usefully directed to 
recommending how browsers *should* behave, rather than how to cope with 
broken browser behaviour?  (I would argue that converting file:/// to 
ftp:/// in the absence of specific knowledge about the host system is broken.)

Are there any particular common use-cases you have in which it is important 
to convert a filename into a URI that works in a dominant majority of 
browsers?  And if so, is this not something that should be dealt with 
separately from non-browser-application handling of filenames and URIs?


Graham Klyne
For email:
Received on Friday, 13 May 2005 12:39:42 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:09 UTC