Sam, I like z/OS as much as the next guy, but I have to disagree with most of your comparisons.
On Fri, Mar 26, 2010 at 4:20 PM, Sam Siegel <s...@pscsi.net> wrote: > > Then there are things like checkpoint restart. On unix, that is "something > the database does". There is now OS level facility that lets you restart > where you left off. > So, you can checkpoint your position in a sequential dataset and synchronize this with a DB2 commit for convenient restart processing without buying a third party product? z/OS checkpoint/restart is pretty archaic IMO. It is much easier to do this on *nux, since a file pointer is just a number. > There are also file system issues. It seems lately that there have been > many comments regards ckd disk, etc. Consider the flip side. Unix does not > have anything close to vsam. If you want keyed file access, you either use > a database, ryo or purchase a third party product. Vsam goes a long way to > solve many problems. Most unix developers are very surprised at > vsam's capabilities. > There are several popular *free* databases, like MySQL. Not to mention free b-tree implementations. Maybe the problem is that *nix developers just don't understand how cool VSAM is.... Just explain the details of VSAM SHAREOPTIONS to them, that will convince them :-) > Tapes and tape management are woefully lacking on unix. There are still > many cases, outside of backup, were tape plays an important role. With a > unix system, using tape is difficult at best. > Seems to me like you have to buy a vendor product on any OS to do tape management well. Virtual tape subsystems are the future anyway, and that technology is more or less platform neutral IMO. > Take a look at JCL. Many unix developers don't see the true facility of > JCL. That is indirect specification of resources. JCL (svc 99, etc.) > allows a program to address different input and output as well as printers > and other external devices without the underlying program knowing which > specific resource is being used. On Unix you must pass very specific > information to fopen or open to obtain the resource. A "smart" program can > read this information from a configuration file or > via environment variables. However, there is no standardized way of doing > this. This makes every "production" job on a unix machine that much > more difficult to manage. > It is simply done differently on Unix... on Unix you use shell scripts to parameterize file names as shell variables which are used as arguments to programs. If anything, shell scripts are *more* flexible than JCL. (But on z/OS, you can do both, even in a batch job!) You didn't really mention the things that I see as the biggest advantages of z/OS of Unix: - massive multiprocessing with great resource management (WLM) - JES (but I guess you did cover this) - stability, backwards compatibility, reliability... FWIW, did IBM sell more net-new z engines last year for z/OS or for Linux? Kirk Wolf Dovetailed Technologies http://dovetail.com > Regards, > Sam > >> ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
