Level 1 Calo Software Meeting at RAL on Monday 15 February 1999.
----------------------------------------------------------------

Present: Scott, Kithsiri, Bill, Tony (part time), Tara, Eric, Murrough.
Apologies: Norman, Reg.


The agenda (as rearranged at the meeting) was:

(1) Progress with CVS,AFS & StP................Tara
(2) Requirements document......................Tara,Reg,Murrough
(3) Hardware Testing and Workplan
    (a) Plan for backplane tests...............Scott
    (b) Plans for 1999
        (i)   Hardware schedule................Tony
        (ii)  DSS..............................Tara,Kithsiri
        (iii) ROD..............................Bill?
(4) Hardware infrastructure plans..............Murrough (for Norman)
(5) Progress with DAQ prototype -1 code........Murrough
(6) Next steps.................................All
(7) Preparations for joint meeting at Bham.....All


Progress with CVS,AFS & StP
---------------------------

Tara presented the progress he has made with setting up CVS on the Rutherford
AFS server. He has established a CVS repository for the three main pieces of
our existing software: the diagnostics, the daq and the analysis program.
The current status is described on the web via
      http://hepwww.rl.ac.uk/Atlas-L1/Software/cvs_repository.html

The software is world readable; write access requires an AFS account at RAL.
This is presently granted for Tara, Norman and Murrough. Birmingham colleagues
will follow when AFS is installed there. Access can also be granted to non-UK
colleagues.

So far this setup only deals with the source code. The binaries are built
using the existing Makefiles after installing the source on the local system.
We should consider the use of SRT in future.

Tara also gave a (lack of) progress report on Software through Pictures
(StP) the CASE tool which has been bought for use throughout ATLAS. He has
tried to use this on the RAL HP system, but has been stalled for far waiting
for a CERN AFS account in order to access the SYBASE system which StP uses
as its database.

Tara has been reading some of the introductory documentation and hopes to be
able to try the reverse engineering tool, eg to produce some class diagrams
from the existing analysis program. Discussions with Steve Fisher suggest
that - at least in the offline community - the reverse engineering tool has
been the most useful. Murrough mentioned that the DAQ group seem to have
made extensive use of the design tools though.


Requirements document
---------------------

Murrough has yet to install the version updated in his absence. [Now done].
Tara and Reg have read through and talked about the calibration issues, but
no concrete progress has been made. We have also not made much progress on
the overall workplan and timescales. Despite the immediate necessity to
concentrate on the imminent DSS and ROD modules we should not lose sight of
the longer term issues....


Hardware Testing and Workplan
-----------------------------

(a) Plan for Backplane Tests
----------------------------

Scott stated that no software was required for the backplane test cards
being developed at Birmingham.


(b) Plans for 1999: DSS and ROD Modules
---------------------------------------

Tony presented the current schedule for the DSS and ROD modules. The design
reviews have (in principle) concluded. The feedback will be considered by
Viraj and Tony later this week. The DSS motherboards are expected to be
available for testing at the start of April, with the LVDS daughtercards
about two weeks later. The ROD will follow in mid May.

Several people have read the DSS specification. There are a number of issues
related to software control of the module which need to be clarified. These
include: the procedure for starting and stopping tests; whether the registers
and memories may be accessed by computer during a test run; whether the card
can be configured to produce an interrupt to the DAQ system when an error
occurs; some details of the memory map.
Since the spec was produced a feature to load the FPGAs via VME has been added.

Both DSS and ROD can be clocked and controlled by the TTC system (although
the old TCM may also be used). We have a TTC source on order. We must find
out how to use it!

More detail of how the DSS module can be used by software is required. Tony
will discuss this at his meeting with Viraj.

Given the short timescales, we will (at least to start with) add these
modules to the existing diagnostics and DAQ software. We agree that Scott
and Bill (with Normans help when he returns from holiday) will handle the
DAQ side. Kithsiri (with Murroughs help) will incorporate the DSS into the
diagnostic software.

Norman had drafted a test plan for the DSS module with LVDS daughtercards.
This assumes that the diagnostic type tests on registers and memories have
been successfully completed. He then suggests using the DAQ to:
- test one module connected to itself; varying the cable lengths
- test with several modules still with each connected to itself
- interconnect several modules, all interleaved
- repeat with the modules in different crates with heavy VME activity
- investigate stability of the timing window
- complete tests with Glink and Slink daughter cards
This plan needs to be fleshed out once the details of how the card can be
controlled by software are available.


Hardware infrastructure plans
-----------------------------

Murrough reported a conversation with Norman. The RAL group plans to replace
the existing VAXstation used as an Xterminal with a couple of PCs running NT.
Initially these would just be used as Xterminals to the LynxOS system, but
in the longer term we should consider running our software on NT.
Both Tara and Murrough wondered whether Linux might be preferable to NT.

[It was not discussed at the meeting, but there are also plans to try to run
the diagnostics on a separate platform from LynxOS, using the network to access
VME on the LynxOS system].


Progress with DAQ prototype -1 code
-----------------------------------

Murrough reported that he has made no progress with this since returning from
holiday. In the short term, work on the new modules will take precedence.


Next steps
----------

Work on the DSS and ROD is urgent. In the longer term Scott and Bill should
attend OO design and C++ courses. We should have a clearer period for serious
work on new software after the next few months.

Norman has suggested that instead of trying to coax software out of the DAQ
group, we should consider using the Level 2 "lightweight" DAQ software which
is freely available and which is based on the DAQ groups designs. The feeling
at the meeting was that unless we hit major difficulties in working with the
DAQ group it would be preferable if we only had to made one transition to new
software.

[After the meeting though, in conversation with Fred Wickens, it appears
that since the last public presentation of the Level 2 software at Chamonix,
considerable effort has been made to unify the Level 2 and DAQ group software
even at the level of the calling sequences, so that any transition might be
fairly painless].


Next Meeting
------------

Murrough has to produce an agenda for the software meeting at Birmingham.
This should include reports on progress with the new UK modules; something
from Cornelius describing the software he has written for the Heidelberg
test system; the status of work on software tools; progress with the DAQ
group; consideration of overlap with future offline code (we have had a
long email on this topic from Traudl).