Software Tools Minutes 6 June 1995 ====================== Present: Pushpa Bhat, John Hobbs, Al Jonchkeere, Stan Krzywdzinski, Qizhong Li-Demarteau, Laura Paterno, Serban Protopopescu, Harrison Prosper, Rajendran Raja, Suemee Shim Scribe: Laura Paterno Harrison called our first meeting to order with a brief agenda to discuss and refine our charge and to organize our efforts to fulfill that charge. Before getting to the Agenda, our the chairman had a list of questions based on our charge which he posed to the group. Below is the list of said questions and the various responses they evoked from all present: Questions to Consider: 1) How specific do we have to be? Everyone agreed that we want to be as specific as possible whenever possible even down to selecting a specific tool. The one exception to this was when it came to COMPILERS/DEBUGGERS. Here it was agreed that we would specify certain minimum requirements that it must meet and we would give examples of tools which we know meet those minimum requirements. 2) Our we obliged to target specific Operating Systems? As the Computing Division has decreed that we WILL (whether we want to or not) switch to UNIX we all agreed that for the short term we should make sure we specifically target our efforts here. However, we also all agreed that for the long term we should not exclude other systems from our discussions as who knows the future of any Operating System. 3) Is VMS out of the picture? It's hard to say. We have been told to switch to UNIX but for the short term it is still going to be here and Code Migration and training isn't going to happen overnight. The answer is: we don't know. The rest of the questions were tabled until after we discussed the charge of our group. I list them below for completeness and your general reflection for the next meeting. 4) Should we look to CERN for salvation? 5) Should we look to Industry? 6) Should we talk to CDF and other experiments? 7) Can our current tools simply be mapped to the new Operating System or do we need to rethink how we build, for example, D0RECO, D0GEANT, D0USER,...? 8) How much of the Object Oriented Paradigm should we anticipate using, if any? 9) Is ZEBRA dying? Will it be dead by 2000 or 2005? 10) Code Management: can it be platform independent. We now return to the scheduled agenda... Below is another reminder (just in case you have forgotten it whilest celebrating over the Fourth of July) of what that charge is and who the members of the committee are (not that you don't already know who you are) SOFTWARE TOOLS WORKING GROUP (A) MEMBERSHIP Drew Baden Pushpa Bhat Krzystof Genser John Hobbs Alan Jonchkeere Stan Krzywdzinski Qizhong Li-Demarteau Laura Paterno Harrison Prosper Serban Protopopescu Rajendran Raja Ransom Stephens (B) CHARGE This group is charged with recommending solutions, including both tools and a general description of the manner in which they are employed, satisfying the following needs: a) Document/News Management b) Text editing c) Code management d) Compiling/debugging e) Graphical User Interface building/editing f) Database management (relational and/or object) (C) TIMESCALE OCTOBER 1, 1995 items a) and b) JANUARY 1, 1996 items c) and d) APRIL 1, 1996 items e) and f) We briefly reviewed our charge again and then looked into whether the charges were appropriate and the deadlines were okay. Since all other groups are scheduled to be in before us it was felt that enough time was available to meet said deadlines. Some discussions then followed about the various charges. The Code Management issue brought up quite a number of questions. 1) Is there a good UNIX equivalent to CMS? 2) Do we want a centralized library or do we put things out in a central place when they are ready? 3) In moving from VMS to UNIX we lose the history we have of code. If we develop on UNIX we have to do a conversion from VMS to UNIX. How do we keep CMS up to date on the VMS side? Do we need to write tools to auto update back to VMS from UNIX? 4) Is there something commercially available or something that CERN is working on? The Computing Division is addressing this and other UNIX issues. Should we get someone to come to our meeting to speak about what they are doing? The contact person for this would be Paul Lebrun. The other topics were also briefly discussed before it was decided that we would concentrate on items a) & b) from our charge since they were to be completed on the timescale of October 1. Harrison then proceeded to present to us all a long list of tools that are currently used by D0. Below is said list... CURRENT DZERO SOFTWARE TOOLS ============================ Editors - EVE, LSE Convertors - DVIPS (LATEX/TEX->PS), A2PS (ASCII->PS), D0SOURCES (out of date) Viewers - XDVI, GHOSTVIEW, CDA_VIEWER, DBANK/ZBANK, DZSURV/DZSUM Text formatting tools - LATEX/TEX, Microsoft Word Code Management tools - USERLIB, BETA_UTIL(MMS,CMS), LIBTEST/LIBBETA, RELEASE_LIBRARY GUI builders - COMPACK, TCL/TK, MOTIF Code builders - PROGRAM BUILDER(PBD) Code Checkers - D0CHECK Code PreCompiler- D0FLAVOR(machine dependency stuff) Utilities - D0SETUP Applications - D0X Other - SRCP/EZBANK We went briefly through this list and came up with a few more thought provoking questions... 1) Are we going to continue building our code with PROGRAM BUILDER? 2) Is anyone discussing modularity issues for code development? 3) Do we want to have something that will force a coding standard on people? After this we briefly discussed some examples of tools that our chairman found in Industry... BRISTOL TECHNOLOGY - has a WINDOWS to UNIX portability toolkit which allows you to run WINDOWS applications on UNIX. 3D LABS & ARGONAUT SOFTWARE - have a chip which improves the speed of OpenGL called GLINT. COMMON DESKTOP ENVIRONMENT for UNIX with look and feel like MAC or PC. Al Jonchkeere mentioned that there was a Computing Division group called FUE (Fermilab Unix Environment) which was discussing a lot of these issues. Do we want to buy into? The meeting adjourned after a brief discussion of schedules for various members in terms of when the next meeting would be.