[Info-vax] VMS and package managers, was: Re: Micro Focus Merger with Hewlett Packard Enterprise's Software Business Segment

[Stephen Hoffman]

On 2016-10-04 12:40:46 +0000, Simon Clubley said:

> ...
> For example, the HP C compiler requires a manual change to 
> systartup_vms to start HP C _after_ startnet.com as shown here:
> 
> http://h41379.www4.hpe.com/commercial/c/docs/ig_bvbv_001.html#systartup

Looking at the DECC$STARTUP code, there doesn't seem to be a reason for 
that sequencing requirement.   Not unless the network tips over, if 
that starts up after C.    (If the network startup tips over, why?  
That seems rather like a bug that got papered over, if the network does 
tip over or does undo this startup.)

There is the usual INSTALL stupidity in the C startup, around an 
inconsequential UI distinction between ADD and REPLACE making it into 
the final syntax, and around some shuffling secondary to determining 
whether there are enough global pages and global sections.   Which 
means the installation and startup itself didn't work; that there's no 
reservations, etc.   (That's good, but can and should be done 
automatically across all installed products.  Not in bespoke uniquely 
crafted from artisan-selected DCL added into each product startup.)

The image install is probably best moved over into VMSIMAGES too, as 
why are we chasing around rinky-dink physical memory configurations and 
extra set-up steps in 2016, and headed for 2021?

Then there's that even have to deal with a separate and 
manually-sequenced network startup.

Without a look at how the pieces work and fit together from a few 
thousand meters of altitude, the familiar and comfortable old ways and 
the old messes do perpetuate, mutate, complexify and inevitably 
eventually strangling...    OpenVMS doesn't need it to be convoluted, 
arcane, complex, manually-managed, hard to use, and requiring specific 
expertise for even the simple tasks.   Like installing a C compiler.


-- 
Pure Personal Opinion | HoffmanLabs LLC