Hi,
We are working on migrating an Application from VAX/VMS to Linux environment.
I have 2 queries in the context of Linux.
1. We need to migrate some of the command files (set of VMS DCL commands) to linux shell script. Is there any tool or knowledge base available for this? Currently we are migrating these files command by command. Can any body suggest better way of doing this work?

2. We are migrating Proc from VAX/VMS to Linux. Can any body point me to any tool or knowledge base available to support this activity?

Best Regards
Arpan

I have a lot of experience with VAX/VMS. It is the best d****d operating system ever created. VAX/VMS has never been surpassed in terms of job control, file locking, and other resource management. Linux is weak in all of these areas.

I recommend that, if you think that you have to migrate to a Unix environment, you go with a true Unix. Specifically, I recommend that you use Sun OS/Solaris. It is a mature operating system with much better job control and memory management than Linux. Of course you can run Solaris on x86 but the SPARC platform is more comparable to the VAX/VMS platform as far as being enterprise grade. I happen to think that Sun equipment is way overpriced but at least it is a more appropriate alternative to VAX/VMS than any x86 platform.

http://www.sun.com/service/migration/index.xml

IMO.

But, to answer your questions, there is no easy migration between VAX/VMS and any Unix(-like) environment. They are like deer and fishes. There is simply no comparison. I once saw a book with a name something like "Unix for VMS Administrators". I was not impressed with it's insights. Any Bourne shell manual would be equally useful.

Can you tell your question hit a nerve? I miss my VMS!!!

I have administered many platforms in my career. 1985-CP/M, MS-DOS, Microsoft Xenix, 1986-VMS, 1995-Linux, 2001-Tru 64 and Solaris, and all of the Windows versions from v3.1 through W2003. Of all of those the only one that I thought was well written was VMS.

Amen brother. I am a huge fan of Linux, but VMS was AFAIC the best, *real* operating system I've ever used, and it's what I used for years after I got my degree. Y2K problem? Nope, 4 digit years were standard. User-specific limits on CPU time, disk space, and even which hours during the day when someone could log in? Totally configurable. Multiple versions of a single file*, with limits on how many versions of a particular file a given user could have? Yup. Queue and threading controls? Built-in. RWX? Nope - try RWED. (If you've never used VMS, that's Read, Write, Execute, and Delete, and there definitely is a difference between having permission to edit/write a file vs. having permission to delete a file). Email and real-time, interactive messaging? Absolutely - see the MAIL and PHONE commands.

It's been many years since then, but I still miss it, big time. The day my company switched from VAX/VMS to deploying Windows NT client/server was a really sad day for me. Due to concerns over data security, the drives were physically dismantled, and I saved one of the platters. IIRC, there were 6 per drive, about 8" in diameter, with 200Mg storage capacity each, for a total storage capacity in each hard drive unit (roughly the size of a shoebox) of approx 1.2G. Back in the early 90's, that was huge. I left that company not too much longer after that move from VMS to NT. Today, that company no longer exists.

* This automatic, by-design feature created a new file each time an explicit save command was issued, and it was invaluable in being able to recover a previous version of a file. Say you're working on a subroutine that does [whatever], and you're interested in trying a couple of different approaches. In VMS, you could modify the code, then save it any number of times, and know that each save was stored in a different file. Example:
If you had made (say) 8 or 9 different tweaks to a routine over the course of the day, and realized that your latest attempts were useless, you could easily go back to the better version, which might have been 7 versions ago. Just purge the later, useless files, and you've lost nothing. In Linux, you only get [file] and [file]~, so if you want to save 'work in progress' versions, you have to manually specify a different name. Don't even get me started on Windows. The fact that there are entire companies whose business model amounts to nothing more than "let's allow the user to iteratively save copies of a given file" is just plain astonishing to me. With VMS, "version control" simply never entered the equation -- it was automatic, and useful to an extreme degree.

I recently (2001 - 2003) worked for a mutual fund company that was switching from VMS + Solaris to all NT on the servers. Talk about poor judgment! Happily the company was bought by a very big bank that used Solaris so the mutual fund end users only had some of their data on NT machines. I don't know what happened to their VMS and Tru64 machines. I was laid off because the merged company had more tech support than the bank wanted to keep and they let everyone in the mutual fund company go.

We can't begin to list all of the excellent features that are unique to VMS. I still can't believe that Unix file systems don't record the creation date/time of a file! That can be very frustrating. And the fine control on file permissions, as you pointed out, is very useful. Another thing I can't believe about Unix is that a normal user account can change the ownership and permissions on any file in his/her home directory. In VMS I would create a login.com file, set it to be owned by the system administrator, and make it readable by the end user. That way they couldn't change the login.com. You can't do that in Unix. The end user can always change the ownership and permission of any file in their home directory such as the .bashrc file, for example, and modify it. That's really stupid.

User account privileges. Enough said. If you know then no more needs to be said; if you don't know then no short description can fully explain it. It's very handy. I know that Unix can be compiled with "CAPABILITIES" built into the kernel, but it never is and it doesn't come close to VMS user account privilege assignment. I think back to the numerous times that software developers would say "Just give me set priv privilege." and we'd laugh until we cried.

Then there is system tuning. I was a g** d***** guru regarding VMS system tuning. The things that you could do vs. Unix were amazing. I"ve purchased, AND READ , several books about Unix performance tuning. What a joke! When I think of sysgen and of user account tuning in VMS it makes Unix look about as tunable as MS-DOS.

And don't talk to me about the Linux job scheduler. Uggh. A 300MHz VAX with 96MB RAM running VMS could support twenty people logged in running X software without any delays. The VMS job scheduler was unbelievably excellent.

And what about sharing disks? Unix and Linux STILL can't do that as well as VMS did it in 1986. The cluster wide lock manager was and still is unique in computer operating systems.

Don't get me started. I love VMS.

It's a little bit discouraging when the person who posted the original question doesn't bother to respond to the replies. I guess it's a question of basic manners and courtesy to others.

I apologize for not responding you in time. But to be honest, most of it was not making any sense to me.

I can understand you preferences for one or another OS, but what you state here is untrue. I just created a file in my home dir as root, and assigned it 644 permissions. There is no way I can change the ownership.


jlinkels@jlinkels_lt:~$ ll we*
-rw-r--r-- 1 root root 8 2007-01-22 07:17 weg.weg
jlinkels@jlinkels_lt:~$ chown jlinkels:jlinkels weg
chown: cannot access `weg': No such file or directory
jlinkels@jlinkels_lt:~$ chown jlinkels:jlinkels weg.weg
chown: changing ownership of `weg.weg': Operation not permitted


Are there any people out there who speak IBM's MVS?

jlinkels