[Info-vax] problem with LSE installation

[David Froble]

Phillip Helbig (undress to reply) wrote:
> In article <0e123128-f1d4-437d-8f04-b617c9f46682 at googlegroups.com>, Hans
> Vlems <hvlems at freenet.de> writes: 
> 
>>> Forget VAX.  It's not part of the common code base.
>>>
>>>> Assume that the modifications are non-compatable with earlier versions, 
>>>> and Alphas could not remain in a cluster without the new stuff ....
>>> Let's be happy that the first release of VMS on x86 will allow an ALPHA 
>>> in the same cluster.  Who cares if ALPHA development stops after 8.X and 
>>> VMS 12.3 on X86 can't run with ALPHA in a cluster.
>> If I were a commercial user then I'd agree with that. As a hobbyist I'd hate
>> it if ALpha would be dropped within a year or two. 
> 
> Same here.
> 
>> I have only one Itanium,
> 
> I have none.
> 
>> That said, the way I read Rob's line "We like Alpha; really, we do" is that
>> Alpha plays an important if not essential part in building VMS for I64.
>> If there's such a dependency then Alpha very likely will be kept at the
>> same development level as I64. Whether the same will eventually hold for
>> x86_64 is anyone's guess but I doubt that. By that time mainting Alpha's
>> will be as expensive as mainting VAXes today (in a commercial shop).
>> Just my two cents (as in hfl 0,01)
> 
> I really don't think there will be any new development for ALPHA.  If 
> something comes along for the ride, fine.  If I can finally get 8.4 
> working, I'm happy to stay at that until VSI comes out with VMS on x86.
> My only requirements are that the hobbyist programme continues, for both 
> ALPHA and x86, and that when x86 is available, I can upgrade to the 
> version of VMS on ALPHA which can cluster with it.
> 

VSI has mentioned a new TCP/IP stack.  So, if they develop one, and it's 
much better than the current one, wouldn't you want to use it?  Even on 
your Alphas?  And what if it required something new in VMS?  Yeah, maybe 
a patch might be made available.  Or maybe not.

It appears to me that keeping a common code base, and test builds, are 
one thing.  Distribution of new builds is quite another.  I'm thinking 
this might be a "sticking point".