To: Mitra <mitra@pandora.sf.ca.us>
Subject: Re: FTP URL mapping
In-Reply-To: Your message of Thu, 03 Mar 1994 22:28:57 +0000.
<CM414A.Iy6@pandora.sf.ca.us>
Date: Thu, 03 Mar 1994 20:44:11 -0500
From: John Curran <jcurran@nic.near.net>
Message-Id: <9403040144.aa05625@nic.near.net>
--------
] From: Mitra <mitra@pandora.sf.ca.us>
] Subject: Re: FTP URL mapping
] Date: Thu, 3 Mar 1994 22:28:57 GMT
]
] Keith is of course correct, that a precise spec for retrieving the file
] requires a step by step process. However - there is another
] consideration, these URL's are generated by humans as often as by
] machines, and most humans dont know how to retrieve files from their own
] ftp servers - they just know where the file is.
Mitra,
I believe in trying to interoperate with _almost_ everything.
However, I'm certain that are not FTP servers _so_ broken that
interoperability is not a worthwhile goal.
] If anyone believes this is simple, I suggest looking at the gopher->ftp
] gateway, which has to special case for about 10 different ftp
] implementations.
Just offhand, how many of these implementations would work without any
special handling if the access process was a two-step "change directory"
then "retrieve file" sequence? I'm of course presuming that the two
components were clearly identified in the FTP URL...
/John