At line 0 added 34 lines. |
+ !!! Avoid Trouble with SICS |
+ |
+ The underlying assumption is that SICS is usally quite mature. Serious bugs are rare. |
+ There are some conditions when problems with SICS tend to occur: |
+ |
+ # When the software and hardware has changed after the startup of SINQ. Normally staff is available at this |
+ time to sort things out. |
+ # When you do something new: new hardware, new routines etc. |
+ # When the hardware breaks wile SINQ is running |
+ |
+ There are some things you can do to minimze issues and have them resolved quickly: |
+ |
+ # Know your instrument well enough to be able to switch it on or off. A surpring number of problems have to do |
+ with this. |
+ # Be aware of the flags in the inst.tcl file which control optional hardware. |
+ # If you plan larger changes to the hardware which might require new drivers consult with LDM computing early. |
+ # At SINQ startup, check your instrument thoroughly: try as many options as feasible, run a known sample to |
+ verify that everything is OK. Perhaps you write a batch file for this which you share with LDM computing? |
+ # Schedule additions to your instrument or trying out new features into times when computing support is available. |
+ # When there is trouble with SICS, do not only work around it: report it by email. Only then SICS can be |
+ improved for everyone. |
+ |
+ |
+ Be aware of the fact that testing DAQ software is a problem: we have automatic tests for many features of |
+ SICS which can be tested offline. But some issues only occurr when running with real hardware and samples. |
+ And LDM computing does not even have a hardware setup for tests anymore. If we can have two weeks of beam |
+ time at SINQ startup to test with samples and everything we might be able to deliver a perfectly working SICS. |
+ As no one is willing to waist so much beamtime for software, the only alternative is what we do today: wait |
+ until the problem occurrs and try to fix it quickly. |
+ |
+ |
+ |
+ |
+ |