[Info-vax] Happy new Year !

[Michael Kraemer]

Phillip Helbig---undress to reply schrieb:

> 
> When I started with VMS in 1992, I heard from many people that it was 
> already obsolete and wouldn't be around in 5 years, much less 10 or 15.

and that prediction became true in many (most?) places indeed.

> Queue famous Mark Twain quote.  But...

In the meantime he passed away.
So even *that* prediction has come true.

> 
> ...I have to agree that this is a good summary.
> 
> What will happen?
> 
> There are still paying customers who have a large investment in VMS with 
> code which would be very difficult to port.  (And it is the PAYING 
> customers who decide what will happen.)  I see four possibilities:
> 
> 1) Oracle will buy VMS to keep Rdb customers.  Most Rdb customers, if 
> forced to move off VMS, would probably move off Rdb as well. 

To Oracle, most probably. So discontinuing VMS would be
a win for Oracle. Only one DB to maintain and develop.

> This might 
> not be as absurd as it might sound at first.  Remember Porsche wanted to 
> buy Volkswagen, but what came out in the end is the Volkswagen bought 
> Porsche.  Compaq paid $ 9 billion for DEC.  Say VMS is worth 1 or 2 
> billion.  Shouldn't be a problem for Larry.

A bit unrealistic, I'd say. Those $9Bn represented roughly
DEC's annual revenue back then.
What's the annual revenue VMS generates these days?
Last time I checked HP's BCS generated less than $4Bn
(with a profitability below HP's average),
so even if one assumes an optimistic 10% share for VMS,
it's "value" would be in the millions rather than billions.

> 2) Big companies with a lot invested in VMS will get together and buy
> VMS so that they can call the shots, after realising that this would be
> cheaper than porting all their code, not to mention the decreased 
> reliability, availability etc.

Of course I can't speak for big companies,
but somehow common sense tells me, that if one
has legacy apps on a legacy platform with doubtful
future, the probability (and pressure) is very high
not to port it 1:1, but just leave the past behind
and switch to new apps with similar (or even better)
functionality on new platforms.

> 3) HP will decide to port VMS to REAL industry-standard hardware.  I 
> think the technical reasons for not doing so are no longer valid.  After 
> the ports to ALPHA and then Itanium, this should be relatively easy.

And most probably will end up in another lineup of proprietary hardware,
so what would one win?

> 4) HP will abandon VMS, all customers will have to jump ship and in a
> few years HP will sell VMS for just a few million, realising that if it
> waits any longer it won't be able to sell it at all, or only for a token
> price.  Could VMS survive as some sort of open-source project? 
> Probably, but it is not clear a) on what level and b) if this would be 
> enough to keep big customers interested enough to fund support (which 
> would make it possible for some folks to work on VMS full time who 
> otherwise would have to earn their money otherwise).  Of course, HP is 
> not obliged to open-source it, but they might.

This would depend on the fraction of third party's IP
and the difficulty to disentangle it from HP's IP.
IBM for example has been asked twice to release OS/2
as open source, but refused to do so, probably because
they don't own all of the IP.

> What would I like to happen, best first?  3 1 2 4.  What do I think will 
> happen?  1 3 4 2.

5) HP just discontinues VMS and (almost) nobody cares.
So better keep your legacy VAXen and Alphas in good shape,
they might turn out to be the only native VMS boxes in the future.