Re: comment on vicinties vs URIs
felix winkelmann 10 Jan 2005 07:51 UTC
On Sun, 09 Jan 2005 23:30:03 -0800, Per Bothner <xxxxxx@bothner.com> wrote:
>
> Vicinities that aren't general enough to handle URIs don't get the job
> done. There's no point bothering to define an API for them.
Just a short remark before this discussion will go out of hand:
- This SRFI doesn't look like a general pathname facility to me, if I understand
it correct, it merely is intended to simplify accessing common resource-
locations relative to the executing program and/or implementation
("pathname->vicinty" being the one exception)
- It does so in a reasonably simple manner, any attempt to blow this up
into a "the network is the filesystem" be-all-and-end-all device
might be buzzword compliant and trendy, but really is unneccessary.
We still live with our local filesystem and will so for a good deal of time
to come.
- It should be apparent that generalizing this all to URIs brings with it
some security issues
- Even if an all-is-an-URI solution is to be found, I don't see why it couldn't
be compatible with the current (non-URI) draft.
cheers,
felix