On Sun, 24 Nov 2013 09:32:57 -0600, John McKown wrote:
>The z/OS catalog is indeed a wonderful thing. Of course, one thing that
>makes it easier is that a data set name can be a maximum of 44 characters
>long, rather than about 255 for UNIX and Windows path+file name. ...
>
Generally much more than 255 on current UNIX systems, even z/OS UNIX.
The 255 we know is an arbitrary, sometimes constrictive, restriction
imposed by z/OS allocation when a pathname is associated with a DDNAME.
I don't understand why, given that a DYNALLOC TU is described by a 16-bit
length field. MVC legacy?
And the longer path names may have higher mnemonic value.
>On Sun, Nov 24, 2013 at 9:05 AM, Ze'ev Atlas wrote:
>
>> In Unix and Windows you need to know where the heck is your file, like:
>> aaa/bbb/ccc/ddd/eee/fff/myfile
>>
In z/OS, you need to know what the heck is the fully qualified name of
your data set, like:
DSN=AAA.BBB.CCC.DDD.EEE.FFF(MYFILE)
... I see little practical difference. And the UNIX/Windows hierarchial
file system with the current working directory notion makes it easier
to associate a family of related files. And easier to create prototype
environments,
The strength of the catalog is in its ability to locate data sets on
offline media. But as both UNIX and z/OS move to a single-level
storage model assisted gy HSM and similar products, that advantage
is shrinking. I'm accustomed to seeing messages on Solaris
analogous to the HSM recall messages I see on TSO.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
