On 11/3/2022 2:37 PM, Dan Cross wrote:
> In article <tjujnk$18a1d$1@dont-email.me>,
> Dave Froble <davef@tsoft-inc.com> wrote:
>> On 11/2/2022 3:10 PM, Bill Gunshannon wrote:
>>> On 11/2/22 14:11, Dave Froble wrote:
>>>> On 11/2/2022 1:11 PM, Dan Cross wrote:
>>>>> In article <tjrk05$qlu5$3@dont-email.me>,
>>>>> Dave Froble <davef@tsoft-inc.com> wrote:
>>>>>> The best tool for the job is the overriding criteria.
>>>>>
>>>>> And under what circumstances is VMS the right tool
>>>>> for the job for a new installation? I can't think
>>>>> of any.
>>>>
>>>> You misunderstand. The "right tool" would be the apps that the user needs.
>>>> OS doesn't matter, unless it is required for the "right tool".
>>>
>>> You just threw gasoline on Dan's fire. What "tool", right or otherwise
>>> is available on VMS and not on any of the more common systems?
>>
>> If I can mention one, does that prove my point? Or do you need some higher number?
>>
>> The Codis ERP package that small power equipment distributors all over North
>> America use runs on VMS, written mostly in Basic.
>>
>> 1) It isn't likely to be ported anywhere else.
>>
>> 2) It has no ready replacements.
>>
>> That is a tool that is vital to it's users.
>
> Note the question was _for a new installation_.
>
> I can't imagine many new installations are reaching
> for a VMS-only ERP package written in Basic.

The end user would not care and usually doesn't even know
what programming language the application is written in.

But they do care about OS. VMS today would be a hard sell,
but the premise for all this discussion is that VSI get
VMS made some sort of competitive.

Arne

On 11/3/2022 2:35 PM, Dan Cross wrote:
> In article <tjubrr$17lrh$1@dont-email.me>,
> Dave Froble <davef@tsoft-inc.com> wrote:
>> On 11/2/2022 1:11 PM, Dan Cross wrote:
>>> In article <tjrk05$qlu5$3@dont-email.me>,
>>> Dave Froble <davef@tsoft-inc.com> wrote:
>>>> [snip]
>>>> The best tool for the job is the overriding criteria.
>>>
>>> And under what circumstances is VMS the right tool
>>> for the job for a new installation? I can't think
>>> of any.
>>
>> You misunderstand.
>
> Maybe, or did you ignore the question?
>
>> The "right tool" would be the apps that the user needs.
>
> Ok. So under what circumstances are "the apps" that
> the user needs for a new installation available under
> VMS and nothing else?
>
>> OS doesn't matter, unless it is required for the "right tool".
>
> Bluntly, I don't believe you. There's a lot of OS
> dependency baked into IT management. For example,
> how do we do backups for bare-metal DR?

Obviously there are requirements for OS and HW.

Everybody will want:
- some level of reliability
- some level of security
- some level of support
- a reasonable cost
- backup capability
- etc.

Most will have several additional requirements:
- must run in VMWare ESXi
- must run in AWS/Azure/GCP
- must work with some enterprise operation monitoring solution
- must work with some SIEM system
- etc.

But these are all pass/fail.

There are a number of requirements that need a pass grade.

But there are no extra points for coming in above pass.

This is the reason why major DCL enhancements are not likely
to happen. Sure DCL could use some enhancements, but they
will not sell a single extra VMS license.

Arne

On 2022-11-03, Dave Froble <davef@tsoft-inc.com> wrote:
> On 11/3/2022 9:20 AM, Simon Clubley wrote:
>> On 2022-11-02, Dave Froble <davef@tsoft-inc.com> wrote:
>>>
>>> Software that solves specific issues for a particular vertical market many times
>>> will not exist in other software packages. The list can be rather long. Just a
>>> few examples:
>>>
>>> Handling of cores
>>> Superscession of part numbers
>>>
>>> The lack of either of those examples is a total show stopper.
>>>
>>> This is applications that have over 40 years of specific development for this
>>> particular vertical market. Some obsolete. Some quite recent.
>>>
>>> The customers have looked hard for those "alternate ERP packages". The search
>>> has failed.
>>>
>>> One customer is trying a cloud solution. Looks like they are on course for a
>>> future crater.
>>>
>>> You can claim all you want about alternatives. Doesn't mean any exist.
>>>
>>> The most common response from potential new vendors is "we can't do that".
>>>
>>
>> What is it precisely about supersessions that new vendors can't do ?
>
> The biggest issue with potential new vendors is that they seem to have never
> heard of the concept, and don't want to hear about it.
>

I'm finding that rather hard to understand. Parts management, regardless
of industry, has a number of universal factors associated with it, and
supersession management is one of the core ones.

I could understand them having issues with the more complex scenarios
(which depends on how much effort they put into their software), but basic
supersessions, where exactly one existing part number is renamed to exactly
one new part number, are easy to implement.

Who exactly are these vendors ?

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

In article <tk1lng$1i5c$2@gioia.aioe.org>,
Arne Vajhøj <arne@vajhoej.dk> wrote:
>On 11/3/2022 2:37 PM, Dan Cross wrote:
>> Note the question was _for a new installation_.
>>
>> I can't imagine many new installations are reaching
>> for a VMS-only ERP package written in Basic.
>
>The end user would not care and usually doesn't even know
>what programming language the application is written in.

The IT managers better know what they're signing
themselves and their staff up for. That's not a hard
concept.

>But they do care about OS. VMS today would be a hard sell,
>but the premise for all this discussion is that VSI get
>VMS made some sort of competitive.

I thought the premise of this discussion was that there
are things that VSI is going to have to do to make VMS
competitive that they don't seem to be doing (and in some
cases they seem to be doing the opposite), and a lot
of people are responding with arguments that haven't
made much sense in 30 years.

- Dan C.

On 11/4/2022 9:52 AM, Simon Clubley wrote:
> On 2022-11-03, Dave Froble <davef@tsoft-inc.com> wrote:
>> On 11/3/2022 9:20 AM, Simon Clubley wrote:
>>> On 2022-11-02, Dave Froble <davef@tsoft-inc.com> wrote:
>>>>
>>>> Software that solves specific issues for a particular vertical market many times
>>>> will not exist in other software packages. The list can be rather long. Just a
>>>> few examples:
>>>>
>>>> Handling of cores
>>>> Superscession of part numbers
>>>>
>>>> The lack of either of those examples is a total show stopper.
>>>>
>>>> This is applications that have over 40 years of specific development for this
>>>> particular vertical market. Some obsolete. Some quite recent.
>>>>
>>>> The customers have looked hard for those "alternate ERP packages". The search
>>>> has failed.
>>>>
>>>> One customer is trying a cloud solution. Looks like they are on course for a
>>>> future crater.
>>>>
>>>> You can claim all you want about alternatives. Doesn't mean any exist.
>>>>
>>>> The most common response from potential new vendors is "we can't do that".
>>>>
>>>
>>> What is it precisely about supersessions that new vendors can't do ?
>>
>> The biggest issue with potential new vendors is that they seem to have never
>> heard of the concept, and don't want to hear about it.
>>
>
> I'm finding that rather hard to understand. Parts management, regardless
> of industry, has a number of universal factors associated with it, and
> supersession management is one of the core ones.
>
> I could understand them having issues with the more complex scenarios
> (which depends on how much effort they put into their software), but basic
> supersessions, where exactly one existing part number is renamed to exactly
> one new part number, are easy to implement.

Consider simple things such as gaskets and such. Sourced from multiple mfgs,
each with their own part numbers, which can change quite often. A superscession
loop of more than 20 part numbers is quite normal.

> Who exactly are these vendors ?
>
> Simon.
>

Small outdoor power equipment distributors.

Software vendors in the auto parts business have attempted to enter this market
multiple times, and have failed every time. Talking about some rather large
software vendors.

It is strange, as one would think that auto parts software should work. Has not.

Believe it or not, there are instances where things need to be specific, and one
size fits all fails miserably.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

> -----Original Message-----
> From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of Arne Vajhøj
> via Info-vax
> Sent: Thursday, November 03, 2022 9:28 PM
> To: info-vax@rbnsn.com
> Cc: Arne Vajhøj <arne@vajhoej.dk>
> Subject: Re: [Info-vax] The real problem that needs solving to grow VMS
>
> On 11/3/2022 2:35 PM, Dan Cross wrote:
> > In article <tjubrr$17lrh$1@dont-email.me>, Dave Froble
> > <davef@tsoft-inc.com> wrote:
> >> On 11/2/2022 1:11 PM, Dan Cross wrote:
> >>> In article <tjrk05$qlu5$3@dont-email.me>, Dave Froble
> >>> <davef@tsoft-inc.com> wrote:
> >>>> [snip]
> >>>> The best tool for the job is the overriding criteria.
> >>>
> >>> And under what circumstances is VMS the right tool for the job for a
> >>> new installation? I can't think of any.
> >>
> >> You misunderstand.
> >
> > Maybe, or did you ignore the question?
> >
> >> The "right tool" would be the apps that the user needs.
> >
> > Ok. So under what circumstances are "the apps" that the user needs for
> > a new installation available under VMS and nothing else?
> >
> >> OS doesn't matter, unless it is required for the "right tool".
> >
> > Bluntly, I don't believe you. There's a lot of OS dependency baked
> > into IT management. For example, how do we do backups for bare-metal
> > DR?
>
> Obviously there are requirements for OS and HW.
>
> Everybody will want:
> - some level of reliability
> - some level of security
> - some level of support
> - a reasonable cost
> - backup capability
> - etc.
>
> Most will have several additional requirements:
> - must run in VMWare ESXi
> - must run in AWS/Azure/GCP
> - must work with some enterprise operation monitoring solution
> - must work with some SIEM system
> - etc.
>
> But these are all pass/fail.
>
> There are a number of requirements that need a pass grade.
>
> But there are no extra points for coming in above pass.
>
> This is the reason why major DCL enhancements are not likely to happen.
> Sure DCL could use some enhancements, but they will not sell a single
extra
> VMS license.
>
> Arne

While the last 15-20 years was about reducing HW costs, imho, the next 15-20
years will be about reducing SW costs.

Hence, the very high costs for traditional products like VMware, Oracle, MS
SQL etc. are now under strategic reviews by many large companies.

I also think the traditional model of requiring a different server for each
business application and/or multiple tiers is going to change as well, but
that is a different discussion.

One previous DC migration engagement I did about 4 years back involved a
large Cust who had a traditional Oracle DB/middleware environment (on Oracle
Solaris btw) for their mission critical environment. While our engagement
was restricted to transition, not transformation, we did learn that their
next gen platform they were developing on was alternate Linux/much cheaper
DB/middleware products.

In the words of one of their senior app architects "we just could no longer
continue with the old Oracle platform where every new feature or add-on
ended with the question "how many cores will that require?""

For those not familiar with Oracle, their DB core licensing is likely the
most expensive in the IT industry. List pricing (ok, few pay list prices and
have varying degrees of discounts) is $47K per server core.

My point is that just because a product is well established today, like all
past changes in the IT industry, it does not mean their market position is
not going to change in the future.

Regards,

Kerry Main
Kerry dot main at starkgaming dot com

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On 11/6/2022 11:08 AM, kemain.nospam@gmail.com wrote:
> While the last 15-20 years was about reducing HW costs, imho, the next 15-20
> years will be about reducing SW costs.
>
> Hence, the very high costs for traditional products like VMware, Oracle, MS
> SQL etc. are now under strategic reviews by many large companies.

> One previous DC migration engagement I did about 4 years back involved a
> large Cust who had a traditional Oracle DB/middleware environment (on
Oracle
> Solaris btw) for their mission critical environment. While our engagement
> was restricted to transition, not transformation, we did learn that their
> next gen platform they were developing on was alternate Linux/much
cheaper
> DB/middleware products.

> For those not familiar with Oracle, their DB core licensing is likely the
> most expensive in the IT industry. List pricing (ok, few pay list
prices and
> have varying degrees of discounts) is $47K per server core.

PostgreSQL and MySQL are definitely doing well in the database market
(against Oracle DB and MS SQLServer).

And Spring Boot and Quarkus are similar doing well in the Java
application market (against IBM WebSphere AS and Oracle WebLogic).

And if one is not into Java then PHP, node, Django etc. are
also low cost solutions.

> I also think the traditional model of requiring a different server for each
> business application and/or multiple tiers is going to change as well, but
> that is a different discussion.

That is happening with the introduction of containers.

> My point is that just because a product is well established today, like all
> past changes in the IT industry, it does not mean their market position is
> not going to change in the future.

It most likely will change.

It has changed from decade to decade in the past. It seems highly
likely that it will continue to do so in the future.

Arne

On 11/2/2022 9:30 AM, Simon Clubley wrote:
> When you finally answer my question about why people would be persuaded
> to try VMS in the first place,

I actually did.

You apparently just did not recognize it as an answer.

> could you also answer why people would be
> willing to put the future of their company at risk by buying a product
> from a small vendor (VSI) that will stop working if that vendor goes bust ?

If VSI goes bust - which there is no indication that it will - and
if there is noone picking up VMS - which seems unlikely if there
are companies willing to pay - and it falls unlucky at the
end of a license period and if the server is critical and if
they did not negotiate and got a forever license from VSI, then
it would be a problem.

But that is a lot of if's.

Arne

On 11/2/2022 9:23 AM, Simon Clubley wrote:
> On 2022-11-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> On 11/1/2022 10:02 AM, Simon Clubley wrote:
>>>
>>> What is the argument you would make that is strong enough to persuade them
>>> to go with VMS in spite of the problems involved with doing so ?
>>>
>>> Until you can answer that question, you are not going to be selling VMS
>>> into brand new sites.
>>
>> No - I am thinking like someone that does not get his
>> understanding of IT decision making from reading Dilbert.
>>
>
> In that case, you should be able to understand that people are not
> interested in getting fired and losing their pension and salary for
> trying something new that comes with major risks of its own without
> a VERY good reason for trying that something new.
>
> You correctly talk about only small numbers being required to try VMS
> compared to the industry as a whole.

We all know the "Nobody ever got fired for choosing IBM", "Nobody
ever got fired for choosing Microsoft", "Nobody ever got fired for
choosing Java".

And it is very funny. But it has very little to do with the real world.

For individual contributors it may be a safe choice to stay
mainstream - everybody else is doing it so you can't fire me for
bad performance.

But at the decision maker aka CIO level only results counts -
excuses does not matter.

The CEO-CIO discussion would go like:

CEO: I am not happy that project X failed
CIO: Not my fault I choose the same platform as everybody else
CEO: I don't care - it failed and it does matter what everybody else do
CEO: Out of curiosity - did it fail other places and you still picked it
or is everybody else except you able to get it working?
CIO: I will go clean my desk

And it should be pretty obvious from history that the decision makers
are not always making the same decision as everybody else. If they
did then the IT landscape would change very little. But the IT landscape
is constantly changing - 1990 was different from 1980, 2000 was
different from 1990, 2010 was different from 2000 and 2020 was
different from 2010. That is only possible because decision makers
make other decisions than what everybody else does.

> You have not yet talked about _why_ even those small numbers would be
> persuaded to try VMS. IOW, you have not yet answered the question I asked
> at the start of this thread.

There has been examples - you just did not recognize it.

You need to think of the market not the individual decision.

For the individual decision there will be a specific reason to
make a specific decision.

But if you look at maybe 500000 decision makers making decision
for 1 million solutions involving 5 million servers, then
they are not all going to apply the same analysis and logic
and end up with the same result. They will have different
background, different priorities, different experiences and
will end up with different decisions.

Some decisions will be more likely than other. There will
be a lot more picking Linux than VMS. But some will pick
VMS (the premise is still that VMS clears all the "must haves").
The uncertainty in such a large number of decision processes
will ensure that.

Of course each decision will still have something driving
that decision. But it does not need to be a particular important
reason or for that matter a particular good reason.

It could be purely technical based: we like the deep integration
of cluster in VMS, we like the batch system in VMS, we like VMS
calling convention etc..

It could be risk based: we like the fact that VMS code base
is small, we like the fact that VMS code base is very stable,
we like to use different OS'es so all systems it not at risk
when a vulnerability is found so we pick VMS for some tasks.

It could be totally non-technical: the last project used
Linux and failed so it has to be something else let us randomly
pick VMS, the CIO used VMS in his first job 30 years ago, the
CIO played golf with Johan Gedda recently.

None of these will apply to huge part of decisions.

But if each of them apply to 0.05% of decisions and we use
the numbers above then it adds up to 4500 picks of VMS
for 22500 servers.

My guess is that VSI would be very happy even with a
much smaller number.

Arne

On 11/4/2022 10:50 AM, Dan Cross wrote:
> In article <tk1lng$1i5c$2@gioia.aioe.org>,
> Arne Vajhøj <arne@vajhoej.dk> wrote:
>> On 11/3/2022 2:37 PM, Dan Cross wrote:
>>> Note the question was _for a new installation_.
>>>
>>> I can't imagine many new installations are reaching
>>> for a VMS-only ERP package written in Basic.
>>
>> The end user would not care and usually doesn't even know
>> what programming language the application is written in.
>
> The IT managers better know what they're signing
> themselves and their staff up for. That's not a hard
> concept.
>
>> But they do care about OS. VMS today would be a hard sell,
>> but the premise for all this discussion is that VSI get
>> VMS made some sort of competitive.
>
> I thought the premise of this discussion was that there
> are things that VSI is going to have to do to make VMS
> competitive that they don't seem to be doing (and in some
> cases they seem to be doing the opposite), and a lot
> of people are responding with arguments that haven't
> made much sense in 30 years.

I think everybody agrees that VMS needs to have a significant
number of improvements before it becomes some sort of
competitive in the market for new customers.

But my point is that when VMS pass all the basic checks,
then some will pick it - I do not consider it necessary
to have a super killer feature that only VMS have to make it
sell.

Arne

On 11/6/2022 8:16 PM, Arne Vajhøj wrote:
> For the individual decision there will be a specific reason to
> make a specific decision.
>
> But if you look at maybe 500000 decision makers making decision
> for 1 million solutions involving 5 million servers, then
> they are not all going to apply the same analysis and logic
> and end up with the same result. They will have different
> background, different priorities, different experiences and
> will end up with different decisions.
>
> Some decisions will be more likely than other. There will
> be a lot more picking Linux than VMS. But some will pick
> VMS (the premise is still that VMS clears all the "must haves").
> The uncertainty in such a large number of decision processes
> will ensure that.

Note that the premise is still that VMS is at a state where
it has everything to make it a viable solution - it makes
the shortlist as I called it 50 posts or so up.

Also note that it is obviously a requirement for VMS success
that when it gets picked it actually works. But I am not so
concerned about that. Why wouldn't it work.

Arne

On 11/2/2022 9:23 AM, Simon Clubley wrote:
> On 2022-11-01, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> On 11/1/2022 10:02 AM, Simon Clubley wrote:
>>>
>>> What is the argument you would make that is strong enough to persuade them
>>> to go with VMS in spite of the problems involved with doing so ?
>>>
>>> Until you can answer that question, you are not going to be selling VMS
>>> into brand new sites.
>>
>> No - I am thinking like someone that does not get his
>> understanding of IT decision making from reading Dilbert.
>>
>
> In that case, you should be able to understand that people are not
> interested in getting fired and losing their pension and salary for
> trying something new that comes with major risks of its own without
> a VERY good reason for trying that something new.

I am of the opinion that a reasonably competent person would most likely make
his choices for a system, OS, and such work out Ok. After all, his choices
would reflect his competence, right? And if (s)he isn't competent, perhaps
firing is a good course to take.

> You correctly talk about only small numbers being required to try VMS
> compared to the industry as a whole.
>
> You have not yet talked about _why_ even those small numbers would be
> persuaded to try VMS. IOW, you have not yet answered the question I asked
> at the start of this thread.

VMS is not new. It's capabilities are well known and documented. If a person
is competent, and VMS best fits the needs, then choose VMS. If VMS is the best
fit, and not chosen, that again goes back to "competent", right?

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On 2022-11-06, Arne Vajhøj <arne@vajhoej.dk> wrote:
>
> Note that the premise is still that VMS is at a state where
> it has everything to make it a viable solution - it makes
> the shortlist as I called it 50 posts or so up.
>
> Also note that it is obviously a requirement for VMS success
> that when it gets picked it actually works. But I am not so
> concerned about that. Why wouldn't it work.
>

One obvious potential issue is some network stack incompatibility,
either within the network stack implementation or due to changing
external factors.

Two examples of the latter would be stronger encryption protocols
implemented elsewhere (this is currently a real issue within VMS),
or a sudden additional security requirement such as 2FA support
required for access to VMS systems.

Outside of security, there's also a possibility that some parts of
VMS may work just fine in testing, but have performance issues during
really heavy loads due to parts of VMS being much heavier weight than
in other comparable operating systems.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 2022-11-06, Dave Froble <davef@tsoft-inc.com> wrote:
> On 11/2/2022 9:23 AM, Simon Clubley wrote:
>> You correctly talk about only small numbers being required to try VMS
>> compared to the industry as a whole.
>>
>> You have not yet talked about _why_ even those small numbers would be
>> persuaded to try VMS. IOW, you have not yet answered the question I asked
>> at the start of this thread.
>
> VMS is not new. It's capabilities are well known and documented. If a person
> is competent, and VMS best fits the needs, then choose VMS. If VMS is the best
> fit, and not chosen, that again goes back to "competent", right?
>

And now we are back to what is essentially my original question: In what
situation would VMS be considered "the best fit" for a new installation
these days ?

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 11/7/2022 8:41 AM, Simon Clubley wrote:
> On 2022-11-06, Dave Froble <davef@tsoft-inc.com> wrote:
>> On 11/2/2022 9:23 AM, Simon Clubley wrote:
>>> You correctly talk about only small numbers being required to try VMS
>>> compared to the industry as a whole.
>>>
>>> You have not yet talked about _why_ even those small numbers would be
>>> persuaded to try VMS. IOW, you have not yet answered the question I asked
>>> at the start of this thread.
>>
>> VMS is not new. It's capabilities are well known and documented. If a person
>> is competent, and VMS best fits the needs, then choose VMS. If VMS is the best
>> fit, and not chosen, that again goes back to "competent", right?
>>
>
> And now we are back to what is essentially my original question: In what
> situation would VMS be considered "the best fit" for a new installation
> these days ?
>
> Simon.
>

I don't know any answer to that question, and not just for VMS. What would be
the situation for any platform, OS, language, and such? Any such answer would
belong to specific requirements for a task.

Do you have any answers for any specific OS?

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On 11/7/22 09:53, John Dallman wrote:
> In article <tk9oe1$3csgi$1@dont-email.me>, davef@tsoft-inc.com (Dave
> Froble) wrote:
>
>> VMS is not new.It's capabilities are well known and documented.
>
> Its capabilities are not well-known to most people working in IT this
> year, and actually reading documentation is deeply unfashionable.
>
>> If a person is competent, and VMS best fits the needs, then choose
>> VMS. If VMS is the best fit, and not chosen, that again goes back
>> to "competent", right?
>
> I think you might be disturbed by the level of competence shown by many
> IT decision makers these days. I have been on calls where product
> architects demanded the ability to run x86-64 Linux Docker containers on
> SPARC Solaris, and didn't know what the architecture problem was with
> that.

I think you might be disturbed to find out that this is nothing
new. :-)

bill

On 2022-11-07 02:16, Arne Vajhøj wrote:
>
> We all know the "Nobody ever got fired for choosing IBM", "Nobody
> ever got fired for choosing Microsoft", "Nobody ever got fired for
> choosing Java".
>
> And it is very funny. But it has very little to do with the real world.
>
> For individual contributors it may be a safe choice to stay
> mainstream - everybody else is doing it so you can't fire me for
> bad performance.
>
> But at the decision maker aka CIO level only results counts -
> excuses does not matter.
>
> The CEO-CIO discussion would go like:
>
> CEO: I am not happy that project X failed
> CIO: Not my fault I choose the same platform as everybody else
> CEO: I don't care - it failed and it does matter what everybody else do
> CEO: Out of curiosity - did it fail other places and you still picked it
> or is everybody else except you able to get it working?
> CIO: I will go clean my desk

I think you're leaving out one important step here.

Someone at some point (should) always ask: Why did it fail.
And then there comes the analysis of the why, and in there will be the
question of the choice of platform, and then it will be recognized that
it was the same platform everyone else chose, so that was/is not the
problem. Something else is.

If someone gets fired, it is in the end not because of the choice of
platform in this case then.

However, if they picked VMS, and it fails for whatever reason, the
retrospect will reach the point of the decision of platform, and
everyone will recognized it was not the same platform as everyone else
chose, and the digging will stop there, and the guy who made the call
for VMS is let go.

Even if you think it should be otherwise.

Johnny

On 07/11/2022 13:52, Dave Froble wrote:
> On 11/7/2022 8:41 AM, Simon Clubley wrote:
>> On 2022-11-06, Dave Froble <davef@tsoft-inc.com> wrote:
>>> On 11/2/2022 9:23 AM, Simon Clubley wrote:
>>>> You correctly talk about only small numbers being required to try VMS
>>>> compared to the industry as a whole.
>>>>
>>>> You have not yet talked about _why_ even those small numbers would be
>>>> persuaded to try VMS. IOW, you have not yet answered the question I
>>>> asked
>>>> at the start of this thread.
>>>
>>> VMS is not new.  It's capabilities are well known and documented.  If
>>> a person
>>> is competent, and VMS best fits the needs, then choose VMS.  If VMS
>>> is the best
>>> fit, and not chosen, that again goes back to "competent", right?
>>>
>>
>> And now we are back to what is essentially my original question: In what
>> situation would VMS be considered "the best fit" for a new installation
>> these days ?
>>
>> Simon.
>>
>
> I don't know any answer to that question, and not just for VMS.  What
> would be the situation for any platform, OS, language, and such?  Any
> such answer would belong to specific requirements for a task.
>
> Do you have any answers for any specific OS?
>
Certainly in UK Local government the rules to connect to the Public
Sector Network (PSN) require that you use an up to date operating system
and have a support contract.

This makes many of the free Linux systems, where per-incident support is
not available, very expensive compared to windows. I think this is
probably current:-

https://ubuntu.com/pricing/pro

so we had 2500 desktops which needed to be on a support contract, $25
per desktop per year, comes to $50,000/year. With Microsoft we had per
incident support...

... and before you all tell me windows is horrid and unreliable, my
experience of working in a 200 server/2000 desktop environment would
suggest otherwise.

Dave

On 2022-11-07, Dave Froble <davef@tsoft-inc.com> wrote:
> On 11/7/2022 8:41 AM, Simon Clubley wrote:
>> On 2022-11-06, Dave Froble <davef@tsoft-inc.com> wrote:
>>> On 11/2/2022 9:23 AM, Simon Clubley wrote:
>>>> You correctly talk about only small numbers being required to try VMS
>>>> compared to the industry as a whole.
>>>>
>>>> You have not yet talked about _why_ even those small numbers would be
>>>> persuaded to try VMS. IOW, you have not yet answered the question I asked
>>>> at the start of this thread.
>>>
>>> VMS is not new. It's capabilities are well known and documented. If a person
>>> is competent, and VMS best fits the needs, then choose VMS. If VMS is the best
>>> fit, and not chosen, that again goes back to "competent", right?
>>>
>>
>> And now we are back to what is essentially my original question: In what
>> situation would VMS be considered "the best fit" for a new installation
>> these days ?
>>
>
> I don't know any answer to that question, and not just for VMS. What would be
> the situation for any platform, OS, language, and such? Any such answer would
> belong to specific requirements for a task.
>
> Do you have any answers for any specific OS?
>

You can argue for Linux for many server-based applications and that includes
some of the same reasons that used to hold true for VMS at one time:

1) Massive installation base. This isn't something you have to worry
about going away in 5 years.

2) Huge trained professional base. This isn't something you have to
worry about finding trained people for in 5 years.

3) An actively probed and maintained OS, which means security issues
are found and fixed and related issues also found and fixed. This ends
up making the OS much more robust, especially as new functionality is
also added to the OS to counter emerging security weaknesses.

For example, Linux now has MAC security whereas VMS is still in the DAC world.

4) A huge ecosystem of software, especially for server applications, with
all the advantages that brings. This is the same position DEC used to be
in decades ago. This includes a large range of programming languages and
development environments, as well as full major application packages.

5) You can have a contract support level that matches your business
requirements and _you_ are in control of how much support you need
(or don't need).

When you combine all of the above, then Linux (for a combination of
technical and business reasons) can easily be argued to be the best
fit for many applications.

You might find something that's more technically advanced in some areas
than Linux, but until you match all the non-technical advantages above
that Linux has, then Linux is still the best fit in the real world for
those applications, even if some alternatives have some more advanced
features technically.

You want VMS to be used instead of Linux in a new environment ?
The above is a subset of the issues you need to address before that can
even become a possibility.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 2022-11-07, Johnny Billquist <bqt@softjar.se> wrote:
>
> I think you're leaving out one important step here.
>
> Someone at some point (should) always ask: Why did it fail.
> And then there comes the analysis of the why, and in there will be the
> question of the choice of platform, and then it will be recognized that
> it was the same platform everyone else chose, so that was/is not the
> problem. Something else is.
>
> If someone gets fired, it is in the end not because of the choice of
> platform in this case then.
>
> However, if they picked VMS, and it fails for whatever reason, the
> retrospect will reach the point of the decision of platform, and
> everyone will recognized it was not the same platform as everyone else
> chose, and the digging will stop there, and the guy who made the call
> for VMS is let go.
>
> Even if you think it should be otherwise.
>

And even better, the person whose responsibility it is to choose between
(say) Linux and VMS, knows all of the above in _advance_ of making this
decision and will be thinking of this while they are making their decision.

That's the barrier in the mind of the person making this decision you need
to smash through before they will even _consider_ VMS instead of "good old
and known" Linux.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

On 11/7/2022 8:37 AM, Simon Clubley wrote:
> On 2022-11-06, Arne Vajhøj <arne@vajhoej.dk> wrote:
>> Note that the premise is still that VMS is at a state where
>> it has everything to make it a viable solution - it makes
>> the shortlist as I called it 50 posts or so up.
>>
>> Also note that it is obviously a requirement for VMS success
>> that when it gets picked it actually works. But I am not so
>> concerned about that. Why wouldn't it work.
>
> One obvious potential issue is some network stack incompatibility,
> either within the network stack implementation or due to changing
> external factors.
>
> Two examples of the latter would be stronger encryption protocols
> implemented elsewhere (this is currently a real issue within VMS),
> or a sudden additional security requirement such as 2FA support
> required for access to VMS systems.

There are some things that need to be done to make the shortlist.

Including some security stuff like 2FA.

> Outside of security, there's also a possibility that some parts of
> VMS may work just fine in testing, but have performance issues during
> really heavy loads

Test includes performance test.

> due to parts of VMS being much heavier weight than
> in other comparable operating systems.

I do not consider VMS heavy weight. When I see how little
that is running on VMS compared to Linux or Windows, then I
consider VMS lightweight. Try count number of processes and
memory usage.

VMS IO may be a pain point. First SSD and a bit of optimized
config. Later some deeper changes.

Arne

On 11/7/2022 7:56 PM, Arne Vajhøj wrote:

> VMS IO may be a pain point. First SSD and a bit of optimized
> config. Later some deeper changes.

When pretty much your whole database is in cache, you won't worry about disk
I/O, and SSDs are icing on top of the cake.

--
David Froble Tel: 724-529-0450
Dave Froble Enterprises, Inc. E-Mail: davef@tsoft-inc.com
DFE Ultralights, Inc.
170 Grimplin Road
Vanderbilt, PA 15486

On 2022-11-07, Arne Vajhøj <arne@vajhoej.dk> wrote:
> On 11/7/2022 8:37 AM, Simon Clubley wrote:
>
>> due to parts of VMS being much heavier weight than
>> in other comparable operating systems.
>
> I do not consider VMS heavy weight. When I see how little
> that is running on VMS compared to Linux or Windows, then I
> consider VMS lightweight. Try count number of processes and
> memory usage.
>

I/O model, process creation, etc are the kinds of things I am referring
to when I describe VMS as heavy weight compared to (say) Linux.

> VMS IO may be a pain point. First SSD and a bit of optimized
> config. Later some deeper changes.
>

That may not be anywhere near enough. For example, what about network
I/O performance on VMS ?

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

> -----Original Message-----
> From: Info-vax <info-vax-bounces@rbnsn.com> On Behalf Of Simon Clubley
> via Info-vax
> Sent: Monday, November 07, 2022 9:41 AM
> To: info-vax@rbnsn.com
> Cc: Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP>
> Subject: Re: [Info-vax] The real problem that needs solving to grow VMS
>
> On 2022-11-06, Dave Froble <davef@tsoft-inc.com> wrote:
> > On 11/2/2022 9:23 AM, Simon Clubley wrote:
> >> You correctly talk about only small numbers being required to try VMS
> >> compared to the industry as a whole.
> >>
> >> You have not yet talked about _why_ even those small numbers would
> be
> >> persuaded to try VMS. IOW, you have not yet answered the question I
> >> asked at the start of this thread.
> >
> > VMS is not new. It's capabilities are well known and documented. If
> > a person is competent, and VMS best fits the needs, then choose VMS.
> > If VMS is the best fit, and not chosen, that again goes back to
"competent",
> right?
> >
>
> And now we are back to what is essentially my original question: In what
> situation would VMS be considered "the best fit" for a new installation
these
> days ?
>
> Simon.
>

How about an application environment whereby the nature of the application
environment required a well established shared disk strategy rather than a
shared nothing strategy?

Reference: Pro's and Con's of shared disk vs. shared nothing
<http://www.benstopford.com/2009/11/24/understanding-the-shared-nothing-arch
itecture/>

Regards,

Kerry Main
Kerry dot main at starkgaming dot com

--
This email has been checked for viruses by AVG antivirus software.
www.avg.com

On 2022-11-11, <kemain.nospam@gmail.com> <kemain.nospam@gmail.com> wrote:
>
> How about an application environment whereby the nature of the application
> environment required a well established shared disk strategy rather than a
> shared nothing strategy?
>
> Reference: Pro's and Con's of shared disk vs. shared nothing
><http://www.benstopford.com/2009/11/24/understanding-the-shared-nothing-arch
> itecture/>
>

Some reading for you Kerry:

https://en.wikipedia.org/wiki/GFS2

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.