[Info-vax] IT shops call on HP and Oracle to kiss and make up

[John Reagan]

"SeanOBanion" <sean at obanion.us> wrote in message 
news:a0318e69-38dc-4b3c-9914-bab2c73e19eb at q26g2000vby.googlegroups.com...
On Oct 6, 9:56 am, Keith Cayemberg <keith.cayemb... at arcor.de> wrote:
> IT shops call on HP and Oracle to kiss and make 
> uphttp://tinyurl.com/6docx5r

>From TFA:
"OpenVMS, in particular, would be a messy operating system to port,
Eunice said. “I call it a ‘guts-exposed’ operating system; it doesn’t
have much of an abstraction layer to simplify the movement to a new
chipset." Plus, OpenVMS has been ported several times already over the
course of its 30-plus-year history, he pointed out. It started out
life on the VAX, then moved to the DEC Alpha chip, and finally to
Itanium. "The prospect of another port is unpalatable,” Eunice said."


Somewhere I remember a comment about the port to Itainum resulting in
less hardware dependence, though not a flat out HAL (hardware
abstraction layer).
Or is that a hopeful fantasy on my part?


I agree.  I don't buy the argument either.  Going from VAX to Alpha, I'd 
guess almost every file in OpenVMS kernel had to be touched in one way or 
the other.  From Alpha to Itanium, less than 10% if I'd have to guess.  Most 
of the architecture specific code is indeed isolated in macros or support 
routines like SWIS.  There would be some very tricky problems to solve with 
regards to calling conventions, memory mgmt, and as we've pounded on lately, 
Macro-32 and register assumptions.  I don't think "guts-exposed" is fair.

John