Charles P. Bowles (cpbowles@gte.net)
Sat, 25 Mar 2000 17:46:14 -0600
Eric, as a member of the user community, representing some end-user
customers, I appreciate the use of NDN/PNDN as an important feature of
messaging, whether it's through a VPIM-based server or on an IVM-based
desktop. If I receive a message from you with several parts, but my system
can't handle all of it, I'd like to know that I didn't get all of it, but
I'd like to receive the parts my system could handle. As the sender, I'd
like to know that only part of my message arrived and was handled properly.
> Unlike a VPIM system, an IVM system can receive an arbitrary message from
> the Internet.
But I disagree with this statement. The TMIA is, in fact, talking actively
about using the internet as a transport, so a VPIM system WILL receive
arbitrary messages from the internet at some point. Shouldn't we assume
this will happen?
>
> The point of IVM isn't that the desktop can play a message from a legacy
> voice mail system. The point of IVM is that the legacy voice mail system
is
> more likely to play a message from the desktop. I have a lot of
flexibility
> at my desktop. I can even figure out how to play raw G.726, even on a
Mac.
> However, I don't have much flexibility at my legacy voice mail system.
It's true that legacy VMs will not have as much flexibility as desktop
systems, but I forsee that the eventual outcome of Wide Area Messaging is
that messages will need to fly from server to desktop and back with ease, so
that the end-users will not know that one of their recipients is on a
desktop and another is on a server. It seems that making VPIM systems and
IVM systems categorically different will make it impossible to reach that
dream of messaging from anywhere to anywhere. Yes, they have different
needs, capabilities, and issues, not the least of which is the reluctance to
upgrade the legacy servers with the new flexibilities, but we WILL be
upgrading them to handle VPIMv2 at a minimum.
>
> On IVM -> VPIM back-off: My premise is that IVM is something different
from
> VPIM. If IVM was still VPIM v3, then it would make sense to say that a v3
> system should fall-back to v2 protocols when interacting with a v2 system.
> However, I don't see legacy voice mail systems using IVM for inter-system
> communication. This is because all of the things that are important to
> voice mail systems, like privacy, guaranteed delivery, return-receipt when
> the subscriber actually listens to a message, and so on ARE NOT PART of
> IVM. These things break in the real Internet mail world.
>
> VPIM v2 and future vXen define how voice messaging systems interact with
> each other over the Internet. In this case, the far-end voice messaging
> system is a specialized, profiled, and limited Internet mail end-point.
We
> break some rules of Internet mail, because the market requires us to
> provide voice mail semantics to inter-system messaging.
>
> IVM defines how voice messaging systems interact with generic Internet
mail
> end-points. We cannot break any rules of Internet mail because, as
> general-purpose end-points in the Internet mail system, we have no control
> on the far-end end-point.
>
>
>
> [snip]
>
> > I propose we state that if a voice
> > messaging system wishes to receive arbitrary messages from Internet
> > clients, it must implement IVM. This renders the IVM back-off to VPIM
> > issue
> > moot.
> >
> I do not see how this makes it moot.
>
> First, an IVM system may not be able to receive an arbitrary message from
> the Internet. What if it is in some special codec that you can't deal
> with?
> What IVM is doing (I hope) is allowing more clients the ability to play a
> message and also the possibility to create one. The IVM profile is
> unlikely
> to be the default setup on all desktops. Still an IVM system is more
> likely
> to be able to play a 'voice email' than a VPIM system.
>
> Second, the IVM to VPIM back-off is intended to define another piece that
> needs to be standardized. It is always useful to suggest an appropriate
> fallback mechanism.
>
> Finally, it does not make sense to have a new voice mail system that
cannot
> communicate with the old voice mail system. That is, an IVM VM system
that
> cannot 'back off' to communicate with a VPIM VM system. Despite the fact
> that everyone will not implement this, I think it is still an important
> feature to standardise. Else we those of us who want to implement this
> will
> come up with different ways of doing it.
>
>
>
>
This archive was generated by hypermail 2.0b3 on Sun Mar 26 2000 - 02:48:05 IDT