- From: Michael A. Dolan <miked@ncd.com>
- Date: Wed, 15 Feb 1995 14:33:07 -0800
- To: ietf-lists@proper.com (Paul Hoffman), uri@bunyip.com
At 09:00 AM 2/15/95 -0700, Paul Hoffman wrote:
>>If the host is required, then why make yet another syntax that must parsed ?
>>I favor the more canonical, finger://host[:port]/blah
>
>I think we're overusing "host" here. For those of us with shell accounts on
>a Unix system, the <query> part of "finger:<query>" is the same as the
>argument we would give to the "finger" command at the shell. If this is
>unclear from the examples in the I-D, I'd love to have better ones.
>
>>Also, in the other scheme, how would one specify a non-standard port ?
>
>You don't. As I said in the I-D, RFC 1288 specifies only one port, and so
>does this URL.
Seems unnecessarily restrictive - local host fingerd, and ( to a lesser
extent) standard ports only.
Since most of the OS's in the world haven't a clue what
a fingerd is, why make an OS-dependent URL syntax that can't be directed
to someplace that has a fingerd running ?
In otherwords, from my Windows 3.1 system, how do I send a finger request
to host, "foo", for "user@bar" running a fingerd on port 9372 ? Maybe
I'm missing something in the syntax ?
Whatever advantages of simpler syntax for UNIX-centric systems would seem
to be outweighed by the need for more general usage.
Mike
-----------------------------------------------
Michael A. Dolan <mailto:miked@ncd.com>
Vice President, Strategic Planning
Network Computing Devices, Inc. (619) 445-9070
FAX -8864
Received on Wednesday, 15 February 1995 17:35:52 UTC