[Info-vax] VSI: "Official 8.4-1H1 Launch"

[Dirk Munk]

clairgrant71 at gmail.com wrote:
> On Thursday, June 4, 2015 at 4:29:45 AM UTC-4, Jan-Erik Soderholm wrote:
>> IanD skrev den 2015-06-04 01:28:
>>> On Wednesday, June 3, 2015 at 11:17:53 AM UTC+10, Craig A. Berry wrote:
>>>> On 6/2/15 6:16 PM, IanD wrote:
>>>>
>>>>> I guess to start with it will be a port to Intel with as few
>>>>> functional changes as necessary just to get the OS across - more
>>>>> major works and future based changes will come along later
>>>>
>>>> Or you could read the roadmap and not guess quite as much:
>>>>
>>>> http://www.vmssoftware.com/pdfs/VMS_Software_Roadmap_20150224.pdf
>>>
>>> My comments were more high level guess by myself as to the likely
>>> strategy of moving OpenVMS to Intel X86
>>>
>>> The roadmap doesn't spell out the details of what aspects of the OS will
>>> make it over to Intel initially. I know the roadmap states 'Architecture
>>> Common' and that may mean everything existing moved forward or it may
>>> mean lowest common denominator - it's hard to know
>>>
>>> For example, will all clustering topologies make it in the first release
>>> or will only the most common make it first round?
>>
>> I do not know what "all" means here, but I expect to see those
>> "clustering topologies" that makes sense on the actual X86 systems.
>> Then maybe "upgrade only support" (or whatever it it called) for
>> those topologies that might be needed for an move to x86.
>>
>>
>>
>>
>>
>>    Without the specifics,
>>> which the roadmap does not give (and I didn't expect it to give), I
>>> merely guessed at the likely philosophy behind how the port might be
>>> done
>>>
>>> It will be interesting (actually exciting) to see an updated roadmap in
>>> say 6 or  even 12 months time to see the progress being made and
>>> hopefully to see more specific details :-)
>>>
>
> "All" means everything unless there is some very specific reason why something in VMS doesn't make sense on the future platform and that would be really rare. I had a slide once that depicted 5% of VMS knows about the platform and the other 95% does not. Once you get the compilers and the 5% working you get the rest for free. For example, there is absolutely nothing to do to make clustering work on a new platform; it is part of the 95%. There may be some restrictions as to what previous versions of VMS can be clustered with the new version but we have always tried to keep those to a minimum.
>
> For first boot (remember the Itanium Boot Contest?) we will leave out a lot of stuff not necessary to boot, login, and do a DIRECTORY command, but once we move to building the entire system you get everything.
>
I remember that HP had the first Itanium boot in January, and was 
shipping CD Roms to third party suppliers in June.