Minutes of Meeting, June 19, 2006 10:00 CDT
Present: Rob, Jim P., Wally
By Phone: Claude, Suzanne
Minutes written by Rob.
Corrections and comments are welcome.
[ My editorial comments in square brackets. ]
Our next meeting is Friday June 30, 2006 at 10:00 PM CDT
In the Bullpen, WH7XE.
Rob Reported on his discussions with Shekhar and Margaret.
I reported to them that we had lots of candidates for
a logbook but only one for a notebook. The notebook
candidate is HepBook (KBook) for which there are
issues with costs and access to source.
They replied that we should descope to just logbooks
and that it is important to get a logbook going soon.
If we need to spend non-salary money
( such as with HepBook) we need to have a case built from
specific user requirements.
Shekhar: right now the only way they have to store
scope pictures is with an elog. There are
presently about 12 test areas, including
Cornell and JLAB. No one likes the present data
communication model. The only way to view these
pictures is to have access to all of the separate
elogs at each site.
Shekhar believes that Cornell will buy into a common
solution. Margaret thinks that ANL and SLAC will buy in.
Drag and drop would be "highly desirable".
Its OK for some features like this to work at the lab
but not on your home pc ( where you are presumably mostly
just browsing or making text entries ).
[ The context of this discussion was a sideline on
making it easy to enter images. The main line of
the discussion was the issue about unified access
to the scope images. ]
Shekhar: cavity tests will be done here, at Cornell,
at JLAB and maybe later at SNS. Its important to be a
ble to make entries from any site and to be able to
search across all sites from a single entry point.
[ This could mean one instance accessable at all sites
or separate instances that are cross searchable, or even
separate logbook technologies so long as they are cross
searchable. ] It would be nice if the same GUI was
present at all sites.
Shekhar wants explicit support for travellers.
( Note that Camille and Andy disagree with this ).
I should get in contact with Peter Kasper who is
doing a similar exercise for databases.
Margaret asked if we need any sort of data integrity
checks. [ See below ].
Notes on our meeting:
1) Suzanne: At D0, the process logger can add images
automatically.
2) A good word to use is "capture" to include
adding images, other attachments, drag and drop ...
3) Rob has a talk scheduled with the Fermi ILC group.
[ was schedule for June 22; now moved to June 29 ]
- Need to use this to get some names of testers.
4) We should do some sort of interface tests with
products on our shortlist. The testers should
be people who will use the products on a day to day
basis. We do not have the time or resources to
do a proper scientific tests.
- Given them a few tasks and no more direction.
- Observe them.
- Have them fill out a comment form.
Maybe we can use an elog to hold the comment forms?
We should expect that many differences among products
will be subtle.
Should do: TD/AD/TTF/CRL
5) Shekhar wants to use the elog as a datalogger
and/or data catalog. We think that both are
bad ideas.
6) One of the options at a local installation is that
users can make their own little custom sql scripts
and run them against the database. Is this
something that's useful? dangerous?
7) For test stands the shift model is not a natural fit.
We need to able to search by cavity or magnet ID.
- Wally mentioned that one might be able to use
the chapter feature of the AD log for this purpose?
[ I am not sure how good a fit this is? ]
- Suzanne mentioned that MINOS has something like
this with forms for each crate, using the CRL.
8) Those present agreed that the JLAB elog does not
offer enough new functionality over the local products.
So it is off the list. [ I spoke with Janice
afterwards and she is comfortable with this. ]
9) We reaffirmed the decision that the SNS elog is
off the list:
- it requires license fees for WebObjects
- We do not have source access for WebObjects.
10) About the complexity of XLST. Jim P says that
it is about as complicated as Tex Macros. XLST is used by
the TTF elog - this is an acceptable level of complexity.
There are minutes attached to this event.
Show them.