> Methinks it all depends on what the meaning of the word "seamless" is. Note
> that Tandem did a CISC-to-RISC transition with minimal disruption to the
> customer base. VAX to Alpha was more disruptive, but it was a reasonably
> successful transition nonetheless. (And would have been more successful had
> DEC managed the marketing, positioning, and phase-in more coherently.)
VAX to ALPHA wasn't seen as such a dismal failure because it was overshadowned
by Digital dropping a large portion of the software portfolio that had made
VMS popular, so any porting failures were in fact blamed on DEC abandonning products.
Digital used the Alpha port as an excuse to drop its leadership position in
messaging for instance, by not porting Message Router and all the Digital and
3rd party gateways that existed, not providing an upgrade path to its newer
Mailbus 400 product which never got off the ground, and forcing messaging
customers to keep some vaxes in their network/cluster to handle the messaging
portion that couldn't run on Alpha. At the very least, Digital could have
bowed out of the MTA business and made PMDF the official product for messaging backbone.
This, of course, was the prelude which gave some paper ammunition to Bobby GQ
Palmer to dump email alltogether and force all Digital to start using
Microsoft products for corporate email etc etc.
Had the VMS engineers accomplished the same level of porting as the Apple
engineers had done, Digital could not have used the VAX->ALPHA migration as an
excuse to drop so many products since VAX executables could have been run on
Alpha transparently by customers. Digital could have had all its products
automatically ported to Alpha, leaving the marginal products in emulation mode.
There is much less software available on VMS now, so I guess the port to
Itanium won't result in so many software apps not being ported. However, I
suspect that much of the 3rd party freeware applications won't make it to
IA64. If customers don't move to IA64, the apps they maintain won't move to
Compaq/HP has an uphill battle to convince VMS customers to want to move to
IA64. It isn't enough to provide the means to port (the engineer's job), you
also have to make customers want to migrate (marketing's job).