Richard Town wrote the following to Craig Ford, and I quote (in part):
-=> Note: Copied from HS_MODEMS by WIMM/2 1.31
    RT-> How can existing vested commercial interests be maintained? 
  CF> As it has always been maintained.... by adding more features than the
  CF> other vendors, or by offering more attractive pricing for equivialent
  CF> merchandise.
 RT> Let's hope it's not x2>v90 vs. K56flex>V90
In time, the proprietay protocols will fade unless the provide a decided 
performance advantage of the standard. There are still instances where 
proprietary protocols such as PEP and HST have an edge.
   CF> They do not prescribe the techniques of building modems, rather 
   CF> they detail the behavior that modems conforming to the 
   CF> recommendations shall have. A great deal of the implementation 
   CF> of a recommendation is left to descretion and ingenuity, 
   CF> and some aspects are entirely optional.
 RT> Yes.  But that does not mean that it's permissable for a V*
 RT> complient product to be unable to connect with another V* complient
 RT> product. If so then ITU(t) is fair game for constructive criticism.
As long as aspect of the protool are optional it is entirely possible for two 
compliant devices to fail to interoperate. The recommendations are 
intentionally non-specific in some aspects to allow for differentiation 
between competitors.
  CF> The purpose of interop teting is to iron out incompatabilities between
  CF> the various implementaions of a recommendation.
 RT> That's what I'd always thought.  However posts to USR_MODEMS (and
 RT> the UK equivalent echoes) seems not to achieve this. One wonders
 RT> whether USR beta testers ever even _talk_ to others...  FYI UK
 RT> Courier's .7778 was a result of co-operation.   Marketing-wise it
 RT> showed up some modem makes for not offering 244 byte or greater
 RT> LAP-M block size, and SREJ; MNPn and extended V42, etc
Interop is a moving target. As vendors refine their implementations to 
account for circumstances that don't show up in beta testing, or just 
refining the state of the art relative to implementation, things break. 
Putting together a device that works is not as trival as some would like to 
believe.
Regards....
Craig
aka: cford@ix.netcom.com
   : craig.ford@2001.conchbbs.com
--- timEd/2 1.10+
---------------