Meeting Notes for 26 November 2008
Present: Mihael, Tom J., Liz,, Tom, L., Dan K.
Absent: Eric, Bob, Mike W., Randy, Kris, Joao
New OGRE pages
Dan K. introduced us to a new version of the
OGRE website that he has been working on to get feedback. He developed it to provide an alternative to the current OGRE interface that uses popups for the data selection and has all the choices on one page. Eric had suggested that maybe he could lead the user step-by-step as Eric does in Bluestone. Cosmic also has the student choose the data first, then goes to a page to choose the input parameters including the plot parameters and then generates the plots. On the new OGRE site are three main buttons on the page: Choose Data, Select Quantities and Plot It!.
Initially some of us had trouble seeing his interface because it required us to set the text size to the largest size. Dan has used a drag and drop approach to the interface to make choices. It needs more contextual help for someone to know how to use it, but he can add that. The concept of choosing triggers also requires some background information.
Dan can fill in the minutes here if he would like to provide more information.
Mihael expressed concern that Dan was developing yet another version on his local machine and not putting the software into SVN as he had agreed to do. Mihael converted an earlier version of OGRE to jsp to get it to work with the Grid. Each time Dan makes a new version of OGRE on his local machine without jsp, Mihael is further from completing his work on getting OGRE to work on the Grid.
There seemed to be two main issues regarding the interface:
- To provide an interface that leads the students through the steps so that they are not confronted with one page with lots of items to choose.
- To provide an interface that uses common interface practices.
Various people voiced concern that he had not succeeded with the second. It was not clear why he had to use drag and drop to set up preferences. Normally checkboxes, pull-downs or radio buttons are used for these. Certainly that is what has been used in the other e-Labs and often in operating systems like the Mac. Dragging and dropping a file on an icon can cause something to happen in some interfaces so there are other uses for it. Tom J and Liz were concerned that the interface as it is is too difficult to understand. [Tom L: bear in mind that Dan indicated that the contextual help was to be added later.] Issue 1 could be addressed using the more standard approach with pull-downs, etc. Tom L. felt that the interface worked and said they had done some testing with one student. Liz was concerned about having continuity between interfaces for the e-Labs. [Tom L pointed out that Bluestone did not follow the cosmic "look and feel" re/ analysis tools. My recollection is that we have discussed this area and left the design of the analysis tools up to the particular experiments--I'll review our discussions and come back to this point. This amounts to a comment on the second starred point, above.]
We may not understand the complexities of the choices that need to be made when doing the analyses and the requirements they make on developing an interface for the testbeam as well as the CMS data site. [Tom L also pointed out that this redesign of the interface, while implemented at the test beam level initially, was designed to handle the complexities of the MC and run data version still being developed.]
We all agreed that we really need a face-to-face meeting to address this. Marge suggests that this can happen in January.
-- Main.LizQuigg - 02 Dec 2008