[Info-vax] Building for Customers, Revenue (was: Re: Bliss was Re: Learning VMS application programming)

[Kerry Main]

> -----Original Message-----
> From: Info-vax [mailto:info-vax-bounces at info-vax.com] On Behalf Of
> Stephen Hoffman
> Sent: 13-Sep-14 12:08 PM
> To: info-vax at info-vax.com
> Subject: Re: [New Info-vax] Building for Customers, Revenue (was: Re:
> Bliss was Re: Learning VMS application programming)
> 
> On 2014-09-13 14:31:04 +0000, Phillip Helbig---undress to reply said:
> 
> > In article <lv1jr9$ecd$1 at dont-email.me>, Stephen Hoffman
> > <seaohveh at hoffmanlabs.invalid> writes:
> >
> >> Best case and assuming a massive budget for development, that's
> most of
> >> a decade of very intensive work to bring the server features forward
> to
> >> what will then be current, and probably twice that effort would need
> to
> >> be invested to create a viable desktop and probably a desktop that
> >> too few folks would want.
> >
> > How would the effort compare to the effort of porting to x86?
> 
> The x86-64 port is probably ~four years end-to-end, maybe longer — this
> from initial designs to a viable and supportable environment with the
> major third-party products ready — as a guess, and you're far less able
> to update the operating system and its features while the port is
> underway.   (Data: Starting with an existing and established
> engineering team, the last port occurred between June, 2001 and
> January, 2005 for a subset of Itanium systems.)
> 

This may be true, but let's not forget that this was done by a company
who's priorities were not OpenVMS and there were constant budget
cut backs and "refocussing" of key resources to other areas of the 
company.

Let also not forget that 2-4 years is not that unrealistic. Microsoft
announced "Cairo" as the NT follow-on back in 1991 timeframe, then
after delays, "re-announced" NT5 and then after delays re-announced
Windows 2000. It was early to mid-2001 before W2k actually shipped.

Almost 10 years from announcement to shipping product.

Good read of OS development positioning:
http://tinyurl.com/OS-Backgrounds

> The port is much smaller than the updates involved in updating the
> operating system to be competitive, in my estimation.
> 
> OS X Server is so massively past VMS now in most ways.  While I'm
> "unfairly" comparing VMS to OS X Server now, in five or ten years, OS X
> Server will be closer to where everybody else will probably be, or
> expect.  Windows Server is getting both easier and more capable as
> well, and Exchange Server is one of the best available platforms for
> business computing.
> 

OS X Server does not, to the best of my knowledge, have much, if any
presence in financial, lottery, stock exchange, manufacturing, and
other mission critical environments. While its market share has
decreased over the years, OpenVMS has an established reputation
as a rock solid computing platform. The reason for this decrease
in share was largely based on high pricing, poor marketing and the 
perception that its parent company had better things to do with its
dollars... all of which are now fixable at some future point in time.
Yes, of course, there are areas of OpenVMS that need enhancements

> The OS is increasingly a hardware abstraction layer and a bag of
> drivers for the higher-level tools and applications most folks care
> about.  Where that abstraction layer isn't already a virtual machine.
> 
While there is certainly a place for them, VM's are simply a means to 
improve server utilization. The challenge with commodity OS's use of
VM's is that their culture feels the best deployment model is one bus 
app per VM.  There are huge costs associated with licensing, support 
and security of this model.

While this might work ok in a distributed model, imho, the one bus 
App per VM is not the best model for implementation of heavily
centralized private/ hybrid cloud computing models. 

Btw, Exchange server is not a good example of HA servers. It is 
actually coyote ugly from a back end server perspective. Just ask any 
SysAdmin who has had to implement a multi-site Exchange Server
environment and then implement fail-over and fail-back. To recover
a lost email requires the SysAdmin to recover an entire mailbox to 
a test server and then extract the email to meet the end user request.

What has driven the proliferation of Exchange Servers is the wide 
acceptance of the Outlook client as a defacto enterprise email client.

[snip..]

Regards,

Kerry Main
Back to the Future IT Inc.
 .. Learning from the past to plan the future

Kerry dot main at backtothefutureit dot com