B. Clark says that a vxWorks implementation on the TriCore architecture
seems to be an illusion.  There is a Wind River press release from 1999
saying that they have an agreement with Siemens to write one, but no
signs that they actually have done so.

B. Sahr is working at comparing Nucleus+ and ThreadX operating systems,
and finds them very similar in capabilities, footprints, and costs.  The
ether stacks are purchasable from separate vendors, and the two stack
vendors are again very similar in capabilities, footprints, and costs.
Vendors may sell a light version of the stack, without things we may
not want (eg not clear that we want multicast).  There will probably 
be an NRE charge for a BSP for the TC-11 IB, from both vendors, but not 
clear how much.  It looks like a system with the kernal and stack will
fit in about 100 kBytes of code, and perhaps another 50 kBytes of system
buffers.  (This is significantly smaller than we are used to for a vxWorks
system.)  BS continues working on the comparison, and intends to look at 
the next software layer up, to see what's available (in particular a JVM 
for java).  He hopes to have the information in hand suffient to decide 
between these two vendors by next week.

K. Ryan suggests that it might be worthwhile to do interference testing
on a monoboard system (of which many are available at a cost lower than
budgeted), to compare with the in-house designed MIB.  The probability 
that one could find one lower in RFI generation than the MIB is quite low, 
but if such could be found, there would be a very large saving in software 
effort.

B. Sahr and W. Koski note that the delivery of the TC-11 IB evaluation 
board is delayed by three weeks to April 15.  WK considers this delay
benign, but would begin to worry if it slips again, or if the planned 
production run slips.