Notes from the meeting of March 26.

W. Koski has found, somewhere in the bowels of the Wind River web pages,
that they do claim to support the Tricore architecture.

B. Clark says he is rather disappointed by the visit from the Wind River
sales team, in that the company seems mostly interested in systems 
larger and fancier than we are currently interested in.  One suspects
that might lead to the company eventually abandoning the system that
will fit in the MIB.  He has looked at the current VLBA station 
software to get an idea of the size of the vxWorks footprint (ignoring
any difference arising from differences between the 68000 and Tricore
architectures).  We have a fairly fat system there.  The system image is 
588 kBytes, and the system, at startup, mallocs an additional 900 kBytes.  
The VLBA station software the requires an additional 1200 kBytes of code 
and malloc'ed data areas.  He will generate a more spartan system, but one
that we might actually use, and see what the equivalent numbers are.
B. Rowen and K. Ryan both state that they have made minimum systems
with system images of order 300 kBytes.

B. Clark asks if GNU supports a compiler for Tricore.  B. Rowen says 
Greenhills sells support for the ThreadX on Tricore, probably implying
that they use the GNU tools.

K. Ryan is playing with Jbed in his borrowed monoboard computer.  He has
a demonstation java program up and running in a 389 kByte footprint.
He will pursue an actual http server for next week.

B. Sahr is pursuing details about ThreadX and NucleusPlus operating systems
and should have significantly more information by next week.

B. Sahr notes that Rational Rose is now installed on windd.  Use of UML
will be supported, but by no means required.

J. Robnett notes that in April he and others will be making decisions about
space allocations in the Control Building for the monitor/control stuff.
B. Clark complains that there are too many uncertainties in the transition
plans to be able to predict the amount of space needed.  JR says he is 
forced to proceed with worst case assumptions.

Project plan now has a date of April 30 to chose the MIB RTOS, but this
should not inhibit us from making a decision earlier, which would be 
extremely desirable, considering the number of other activities waiting
on this decision.  B. Rowen and K. Sowinski will be in Penticton studying
the correlator for the first two weeks in April, so they may have to make
special efforts if they want to register opinions on this matter.