On Tue, 1 May 2007 16:50:00 +0530, Pradeep_Vasudevan wrote:
>
> I tried to FTP from my Windows PC to Mainframe with the help
>of command prompt. It did work. Could any one suggest my if the FTP from
>a PC to a Mainframe sequential file is possible with 'Wise FTP' or any
>other such FTP software.
>
The FTP interface specified in RFC 959 was never designed for the
sort of automated hierarchy traversal performed by GUI FTP clients.
It's remarkable not that such products work well, but that they
work at all.
RFC 959 provide (among others) the following commands:
NLST, which returns a list of file names, one per line. The reply
from NLST does not distinguish between directory names and base file
names.
LIST, which returns a display, intended to be human readable, of file
names with unspecified descriptive information. Implementors generally
cause LIST to return the directory listing characteristic of some
command prompt option familiar to users of the host system.
SYST, which returns a description of the host operating system. The
z/OS FTP server nowadays returns:
215 UNIX is the operating system of this server. FTP Server is running on
z/OS.
... Smart GUI FTP clients can use this reply to select a scheme to parse
the output of the LIST command to infer information not provided by NLST.
But here z/OS provides a rude surprise for the smart client: the format
returned by LIST depends strongly on the last CWD performed. If the user
does CWD to a UNIX directory, the format returned is that of the UNIX
"ls -l" command; after CWD to a Classic prefix, the format is that of
(I'm guessing) that returned by the TSO LISTDS command. So the smart GUI
client can't rely on the value in the "215 UNIX ..." response, but must
further guess the format to parse based on content. And this depends on
the client's considering z/OS prevalent enough to be worthy of the support
resource.
It's remarkable that they work at all.
HTTP is a protocol more suited to automated traversal: HTML is better formally
specified than the response to FTP LIST. Regardless, before the ascendancy
of the WWW and its search engine crawlers, there were analogous crawlers
such as Archie that traversed the FTP (and NNTP?) universe.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
