[Info-vax] What Will Drive More OpenVMS Adoption?

[DeanW]

On Sat, Dec 4, 2021 at 4:20 PM Arne Vajhøj via Info-vax <info-vax at rbnsn.com>
wrote:

> On 12/4/2021 4:34 PM, DeanW wrote:
> > On Thu, Dec 2, 2021 at 8:50 PM Mike K. via Info-vax <info-vax at rbnsn.com>
> > wrote:
> > * WASD
> If you need something VMSish then WASD would be a fine choice.
>
> But if you need something industry standard then it is Apache or nginx.
>

nginx would be OK with me if it ran decently; I'm not a fan of Apache, from
how it was first named to how crap it runs on VMS. (Maybe "ran"- I tried it
once a long time ago, switched to WASD and never looked back.)

> > * OpenJDK.current-ish, Python, Ruby?
>
> Hopefully VSI will get a OpenJDK 17 out soon as 8 is getting old (and I
> do not see much point in 11 at this point in time).
>

There were some weirdnesses moving a project from 8 to 11, and at least
Java SE 8 is (supposed to be) supported until 2035 or some ridiculous
number; longer than the more current LTS version even. I suspect the paying
installed base (of which my employer is a member) is huge and ossified.

> * Postgres
>
> Promised.
>

I can run with that... ;-)

> > Maybe the killer VMS platform is a rock-solid K8S / VM hosting platform.
>
> Most want bare metal hypervisor for VM's in production.
>

Yeah, I can't afford ESXi so I run Proxmox (which is built on Debian) for
my home lab. I hope someday to host VMS on that cluster.


> And k8s very Linux centric. Microsoft support k8s for Windows. VSI could
> support k8s for VMS. But Linux market share is sky high.


Pretty sure you can run (appropriately designed) Linux OCI containers on
Windows Docker; haven't tried K8S itself so I'm not sure how K8s stacks up
in the M$ world, though I am working on a project that will go that way at
some point.

I don't want to run VMS containers- I want to orchestrate Linux containers
on a VMS cluster; the OCI layer should be pretty well documented and VMS
could probably offer better partitioning and resource control (constraints)
than Linux can. But running OCI containers directly on the hypervisor
(VMWare Tanzu) is a thing now, so an opportunity to get to market and win
with that is probably gone.

I (and the audience here) can think of reasons why VMS is a valuable
environment, but to people who have never been exposed to that (or have
forgotten what they learned in college 30+ years ago), the advantages are
too far lost, and the move from what is now the norm is too risky.  At this
stage, I don't see people flocking to VMS because it's VMS- it needs a
killer app not available elsewhere, or a clear advantage over what can be
done in other ways that are now more "mainstream".

Decades ago Robert Heinlein wrote "Methuselah's Children"- SciFi about a
secretive organization of naturally long lived humans who flee Earth to
escape persecution for not giving up the (non-existent except for genetics)
secret to long life. They come back some time later to find that scientists
who went looking for the secret found other [medical] means to live longer.
This is kind of how I feel about VMS and Linux- I see new Linux things pop
up to do awkward work-arounds for what VMS could do naturally a long time
ago. Kubernetes is one of those. ;-)

For the life of me I do not understand the Linux mindset. I'm trying to set
up a smattering of flavors- Debian, Ubuntu, CentOS, RHEL, Oracle Linux,
SUSE. Even Debian and Ubuntu, which are related, how do you set up TCPIP?
dhcpcd (even for a static IP!), or netplan- and which back end? GAH. I'd
prefer a BSD which (when I dabbled there) was at least consistent, but now
there are so many flavors it probably suffers the same disease. Some days
I'm not sure if I want to chew through the straps another 15 years, or
practice peeing in a bottle and go work at the nearby AMZN warehouse.


-- 
Dean Woodward    =o&o
dean.woodward at gmail.com