[Info-vax] GNV (was: Re: September 6, 2016 - new Roadmap and State of the Port updates now on VSI website)

[Stephen Hoffman]

On 2016-09-11 13:31:42 +0000, IanD said:

> The issue is we are trying to bring OpenVMS forward after years of neglect
> 
> Linux has won, it's as simple as that. MS are second fiddle in the 
> server space, OpenVMS is way down the ranks
> 
> I keep hearing about gnu and I use some of the nix tools like awk and 
> when I can be bothered, I resort to the likes of Python to get things 
> down but what I see is OpenVMS still insisting on being OpenVMS and 
> nothing more

Ayup.  GNV is useful for some stuff and is fundamental for porting some 
open-source software to OpenVMS.   For other work, Python or maybe Lua 
or other tools — even DCL — might be a better choice.

> Is gnu released as part of OpenVMS now?

Nope.

> Isn't it a separate download?

GNV is entirely separate, and the current GNV environment pieced 
together from a variety of PCSI kits and updates acquired from various 
locations.

Haven't looked for a BSD-licensed shell recently to see what's 
available there, but versions of common tools are available with 
BSD/MIT licenses.

If anything akin to GNV is incorporated into the base OpenVMS distro, 
it would probably have to start with the BSD tools for copyright 
reasons.

> Why?

Most of the GNV software is GPL.   That constrains how the software can 
be integrated and distributed.    Further, HPE hasn't re-spun a GNV kit 
in a while, but some folks have done and are doing more than a little 
work to update parts of the kit — which is why the bits of the current 
GNV environment are somewhat scattered around the 'net.

> Why isn't it part of the standard OpenVMS distribution and install?

It'll likely remain separate, due to the software copyrights.

> If the majority of the world is linux now, then why are we not doing 
> everything possible to make OpenVMS inviting for the majority to give 
> OpenVMS a go?

VSI is aimed squarely at the installed base.   For the next five years 
minimally, they have more than a little work on OpenVMS itself.

> If tcp/ip wasn't treated like a second class citizen it too might have 
> been better integrated into OpenVMS over the years instead of the 
> piecemeal implementation it is today

IPv4 and IPv6, PKE, LDAP, there's a bunch of bits and pieces that are 
ad-hoc or add-on.    What I've been grumping about for a while, too.

> It starts with an embracing attitude and then flows from there.
> As long as people want OpenVMS to remain OpenVMS from days gone by then 
> it's going to fade away into further obscurity. I'm not advocating 
> OpenVMS becoming linux, but why not embrace the best linux has to 
> offer? What's the harm? More functionality surely equates to better 
> business solutions in quicker time-frames?

Use the best ideas, and improve on them.

> I'd like to see all the good nix tools installed as default in OpenVMS 
> and added to the documentation. If I had known about things like awk 
> when I started out with OpenVMS I would have used it a hell of a lot 
> more over the years
> 
> Having spent a little bit of time learning some of the aspects of awk 
> has eliminated the need to write some pita dcl scripts. Awk is pretty 
> dam quick too. For data that in column format awk can be very handy

Tools such as awk and python can easily exceed what DCL is capable of, yes.

> It's about being productive and some of those nix tools are pretty dam 
> helpful in knocking out solutions and should be part of every OpenVMS 
> users repitior but are often not because they are not even part of a 
> standard OpenVMS install and you will not see them in the documentation 
> either - which is really silly IMO, OpenVMS should ship with as many 
> helpful tools as it can

That adds support costs, and software update effort and costs.   
There's also the trade-off between rolling individual solutions for 
each tool — all the RMS databases underneath various stuff and with the 
endemic issues involved there — versus spending some extra time and 
effort getting a database or other API or tool available across all of 
those and also available for end-user developers, and making for a 
generally better platform.    There's also the other end of this whole 
discussion, where OpenVMS has never been particularly good at 
deprecating and removing old pieces.

TANSTAAFL.

> I didn't realise the nix diff tool was superior to OpenVMS's. Time to 
> have a look I think

I keep suggesting folks go learn some new-to-them tools, and for just 
this reason.

> Food for thought. MS are racing like crazy to add nix related aspects 
> to windows, including the bash shell of all things. Even MS with it's 
> monopolistic attitude of old has turned over and is working with linux 
> rather than trying to go against it, time for OpenVMS to do the same 
> isn't it (or perish)...

More than a few folks here seem to think that what they knew of 
Microsoft circa 2000 is still applicable.   Or of decade-past Apple and 
macOS or iOS, or of Linux, for that matter.

Or more succinctly:

... "Software is "done" in the same sense that a lawn is "mowed"."

... "Any successful technology creates problems only it can solve."



-- 
Pure Personal Opinion | HoffmanLabs LLC