If OpenVMS is profitable, where's the money going?

[Brad McCusker]

Steve -

"Is your team looking at this kind of thing please Brad?"

As Martin said, being able to run configure is a primary goal of the UNIX Portability (UP) effort. The UP effort wants _programmers_ to be able to sit at a VMS machine and program just as easily as they do at their UNIX machine. Lofty goal? Yes. But, I think we can get pretty close.

Certainly configure is a big piece of that. For may applications, it works today. I did a hands on session at the boot camp (and will do others in May) that provided the students the opportunity to do just that. The students took a few applications, expanded them, configure, make, and run! Well, OK, in some cases, it is necessary to do some edits to either configure scripts, makefile or code, but, in many cases those edits are minimal.

I'd be happy to show you some time. Will you be at the bootcamp?

Brad

[Willem Grooters]

Brad,

I would like to propose a slight modification:

The UP effort wants _programmers_ to be able to sit at a VMS machine and program just as easily as they do at their UNIX machine.


Read the first line like this:

... wants _UNIX_programmers_ to be able ...

I'm not sure whether these programmers are willing to adhere to different way of thinking, but from what I've seen, heard and read, quite a lot won't, and stick to their (by some called "pathetic") alternatives to what is standard on VMS and missing on Unix.

The initiative is great, don't get me wrong. But porting a Unix program to VMS is something more than just get it compiled, linked and running. It will also include changes to match, where applicable, the VMS way of doing things; if it only were to adhere to security, stablity and co-operation standards.

Willem

[Ian Miller.]

truely this is the "the thread that would not die !" :-)

Willem's point above is one of his key comments. However I think the key point of the UP initiative is to increase the number of applications available on VMS and to reduce the time taken between the release of a new version of an existing application on say sun Unix and then on VMS.

Teaching unix programmers to program properly is a laudable aim but outside of hp's goals.

[Willem Grooters]

However I think the key point of the UP initiative is to increase the number of applications available on VMS and to reduce the time taken between the release of a new version of an existing application on say sun Unix and then on VMS.


That's Ok.
But as stated: compiling, linking and let it run is one thing. What I need is co-operation of these programs with other, VMS-NATIVE programs. There are quite some issues there in communication, signalling, coordination and data exchange, that ARE causing trouble and should be taken care about. At least part of the money should go there (to keep it ON the topic...).

Willem

[Brad McCusker]

However I think the key point of the UP initiative is to increase the number of applications available on VMS and to reduce the time taken between the release of a new version of an existing application on say sun Unix and then on VMS.


Yes, this is a key motivating factor - it is one of the answers to "why" we are doing it. My earlier statement was addressing the "what" question rather than the "why".


What I need is co-operation of these programs with other, VMS-NATIVE programs.


This is an EXCELLENT point! Those who have seen me speak I hope will recall me discussing this exact point. Yes, it is not enough to just get the programs to run, they absolutely need to be able to interact with the traditional VMS applications on the server.

And this one of the things that makes the project so difficult. If all I needed to do was provide an XPG compliant library, we could have done that by now. Or, if that is all someone really wants, then, run HP-UX or T-64, they already have these libraries. VMS needs those libraries, but, we also need them to work on and with VMS.

Brad

Clusters is another important point. X/Open specs make no mention of clusters, and, we could easily ignore them. But that would be ignoring one of the strongest VMS selling points. So, as much as possible, whenever possible, we do take clusters into consideration.

[Willem Grooters]

Brad,

We need to talk this over anyway. The current state is unusable for me - as I pointed out in an earlier talk on this (TUD Netherlands).


Willem

[Uwe Zessin]

"Teaching unix programmers to program properly ..."

Ian, not that I want to defend all Unix programmers, but I have seen quite bad programs on VMS, too. Programs that are used in manufacturing. Programs that don't even check return codes. Programs that don't use an IOSB - oh, yes, I know it is optional ;-) Programs that use the asynch. call, but don't have proper synchronization implemented. Programs that screw up the original status code. Detached processes that try to use screen routines for the output of error messages. And so on...

It is really amazing how bad a program can be and still work on VMS !!

[Willem Grooters]

VERY true.

Even worse: if a badly written program isn't documented properly - if at all. That's another dicipline a lot of programmers are missing. It's part of the education - at least, ought to be....

Willem

[Jan van den Ende]

Willem noted several point of 'standard' VMS programming that tends to fail or be absent in ported Unix code, Brad added Cluster aspects.
One issue I have found on a number of occasions is the exit code.

In production environments you should always be aware of the possibility that not everything always goes as intended.

Constructs as
$ ON ERROR THEN ...
and its cousins SEVERE_ERROR and WARNING, or
$ IF .NOT. $STATUS THEN ...
(or $SEVERITY)
are VERY important!
And then suddenly they have NO meaning if the exit statusses are not set correctly!

Jan

[Ian Miller.]

to sort of get back to the original point it would be interesting to know if VMS as a seperate business (inc sales, service etc)is profitable and does that profit get invested in VMS (VMS engineering and VMS layered product engineering appear short handed at present).

[Willem Grooters]

Ian,

You should get a reward for the longest threads in the VMS forum, in terms of duration, number of replies and real discussion! (most, I guess, slightly off topic ;-))

To get back to the topic:

Given the fact that VMS-engineering seems to be able to be so VERY profitable - regardless lack of (official) support, spreaded FUD, moving attentions, you name it - I stick to my first reaction, for a great part: to lessen the loss of the ESG department. All initiatives: Unix portability, Itanium porting, new and further developemnts cost money, but if it's true that the VMS _gain_ is exceeding it's costs quite a lot (is $800M revenue a true number?), I think the BIG lot gets into other (that is: not VMS-related) business.
I bet, it has always been that way, perhaps even in Digital times (and surely in Compaq's days).

Willem

[Jan van den Ende]

Willem,

that bet you have won, at least for part of the Digital timescale.
I remember Decus Barcelona 1996, about the darkest time of VMS inside Digital.
It soon became a sport to try, but NOBODY succeeded in getting a single management or sales person to say "VMS" or "OpenVMS" during this WHOLE week!
And it was very difficult to find a smile on the face of any VMS engeneering related person.
One of them (after we had to promise never to give away his name, and nowadays I am not sure anymore who it was) in a session gave vent to their collective sorrow:
VMS and VMS related business revenue the previous year was just over DOUBLE the company numbers, or, more then half of the net VMS result was spent on (in order) PC's, Windows, and Unix. But: VMS was ( ... Legacy ... ) and that was worse then the worst 4-letter word!

Jan

[Keith Parris]

> it would be interesting to know if VMS as a seperate business (inc sales, service etc)is profitable <

According to Mark Gorham, VP for OpenVMS, at the HP user group meeting in Amsterdam last summar, HP's total annual OpenVMS revenues (Hardware, Software, Storage, Services, everything) is somewhere between US$3500 million and US$3000 million, with US$500 million in profits.

(To put this revenue figure into proper perspective, HP is the largest seller in the (much-hyped) Linux marketplace, and HP recently announced it has reached US$2500 million in total revenues for Linux.)

> and does that profit get invested in VMS <

HP continues to make significant financial investments in OpenVMS and layered products, as the Rolling Roadmap will attest (see http://h71000.www7.hp.com/openvms/roadmap/openvms_roadmaps.htm).

But the simple fact that there exists a profit over and above the expenses means that not all that money is being spent on OpenVMS.

Like the profits from the Imaging and Printing division, the extra profits of OpenVMS contribute to the overall success of HP.

While those of us in this forum would prefer that more money be spent in areas like OpenVMS advertising, and we firmly believe that such actions would result in even higher revenues and profits for HP, we don't get to make the decisions on how the money is spent.

And it is important to OpenVMS that its owner be a strong and profitable company, so contributing to HP's overall success contributes in turn to the success and longevity of OpenVMS as well. And OpenVMS' significant profits also make it highly unlikely that HP would do anything which might disrupt that profit stream.

[Ian Miller.]

Keith, that's just the sort of info I was looking for. I agree that VMS has to contribute profit to hp and it a good thing it does. I would like to see more people employed by VMS Engineering and VMS layered products engineering teams so they can do an even better job. What is being done already is wonderful but I do see comments indicating more people would be usefully employed.

[labadie_1]

May be some money could be spent to have more software available on Vms, like Sap, Jonas and others. Today when the customer looks for some soft, if it is not available on Vms, he will just pick up another OS.

[John Eerenberg]

Keith,

> "While those of us in this forum would prefer that more money be spent . . . we don't get to make the decisions on how the money is spent."
And most likely, spending more money wouldn't change VMS's position in the trade rags and in most CIO/CFO/CEO's minds. I have already addressed/posted my thoughts on the cultural problems HP has regarding VMS. I believe fixing that does not take money.

> "And OpenVMS' significant profits also make it highly unlikely that HP would do anything which might disrupt that profit stream."
This sounds good, but I really can't agree with you here. Not in the long term. Yes, it will stay that way for 1 or 2 or 3 years, but what about 4 or 5 years from now? It is hard to say what will happen, but I look at the market segments that are transitioning away from VMS such as the financial market. VMS no longer does wires transfers for many, many financial institutions. Has anyone ever done a wire transfer from, or to, their bank account? In 2003, roughly 60% of these transfers were done by a VMS system. Today it is better than half that much. By the end of 2005 I'm not sure there will be any VMS systems transfering your money. Other application vendors are taking note and another VMS application that dominates in this market has started a transition . . . These are highly visable applications running at many, many financial institutions. but alas, I rather not repeat myself too much.
My point is that VMS is declining (the new VMS based exchanges overseas do not make up for the losses noted above). It is naive to believe that VMS is declining only in financial apps. From where I stand, I am looking at a mountain covered in snow and I see a snow ball rolling down hill gathering more snow . . . If it gets big enough, that just might "disrupt that profit stream." :-(

john

[Ian Miller.]

hopefully it is this sort of thing that will get some serious attention at hp executive level (Carly are you listening?). I hope a big drop in vms revenue is not what is needed before hp sit up and take notice.

At present I see VMS engineering doing a excellent job with limited resources. I accept that in the real world there will never be resources available for everything but feel that at present there are not enough people working on VMS, TCPIP, and other key layered products. The people have many good ideas but there are not enough of them to implement them. The attitude of hp to VMS and VMS Marketing is a another problem which also needs to be addressed but my key point here is that more resources working on VMS and key technologies are needed to make VMS even more wonderful than it already is. Don't get me wrong I'm very pleased with the working being done on the UP initiative and eBusiness technologies and core VMS enhancements such as shadowing mini-merge but feel more could be done and the money coming into to hp for VMS should be used to do it.

With proper marketing VMS could become THE platform for serious applications.