Date: 1-Feb-2000
Attendees: C. Burkhardt, R. Douglas, K. Peterson, D. Shaw, H. Shukla
Most AIs from the last meeting were completed. The question of vulnerabilities of the DM to unanticipated requirements (or changes to requirements that are presently thought to hold) was revisited. One new concern is the question of how the DM can be verified and validated. Once a solid set of use cases is defined it should be possible to address this issue, but for now we will have to keep V&V in mind.
The class diagrams for RPS2 and for SpecView are on-line, accessible from the APT-DM Web page. Evidently the temporary license for the "Together" software has expired, so it is not possible in the near term to generate UML diagrams for the ETC and VTT componenets of the SEA. But as Chris pointed out, most UML tools can read Java class files and generate the corresponding UML, so once the UML tool for APT is selected, the diagrams for the ETC and VTT will be straightforward to generate. In the mean time the JavaDoc for the SEA is available.
| All | Continue thinking about key vulnerabilities of the DM efforts to volatile requirements, and post new items to the apt-dm distribution. |
| Dick | Collect the DM vulnerabilities on the APT Web site. |
Chris evaluated a number of IDEs for potential use on the APT effort, and summarized the results in a document, which will be discussed at the APT Project meeting later. Chris had evaluated a number of other IDEs, some of which have been used on other projects (e.g., Java Workshop, NetBeans, Simplicity), but they were not considered viable candidates for one reason or another. We suggested that the evaluation document be expanded to include the non-viable candidates, so that future projects can benefit from this research.
One area of concern is that of compatibility of the data files between various candidate UML tools. It is possible that we will need to capture our data models in UML before the selected UML and IDE tools are configured for use. However, we expect that the data files will be compatible, or compatible enough, such that data modelling can begin with one or more available UML tools, and the results can be ingested by whatever UML tool is eventually adopted.
| Chris | Include in the IDE Assessment document the systems with which we have some experience, but consider to be inferior. |
| Chris | Make one or more of the IDEs available to the APT group for evaluation and initial work. |
We have reached stage 2 of the high-level development plan (see the previous meeting summary for details), where the initial project infrastructure is nearly in place, the data models for most other relevant systems has been collected, and high-level modelling for the APT can begin. The TransVERSE project has generated a useful data model which can serve as an initial starting point for APT. We will begin with the high-level concepts (utility, observatory, proposal, etc., classes and the activities model, and progress to greater specificity. Some areas, such as the VTT and the ETCs, will need greater scrutiny (and iteration) since these facilities have not heretofore been formally included in the Phase 1 processing system. The results of the first round of high-level modelling will be used to divide up the problem into areas that we will develop individually or in groups of two.
| Rob | Construct or capture relevant UML diagrams from the TransVERSE project for use as a starting point for our work. |
| Dick & Hemant |
Begin white-boarding essential elements of the instrument and astroModel packages, particularly as they relate to the ETCs. |
The next meeting (Feb. 8) will focus on white-boarding the high-level
description of the data model.
Copyright© 2000 The Association of Universities for Research in Astronomy, Inc. All Rights Reserved.