- See
http://www.npac.syr.edu/users/gcf/petcandt/index.html for relevant web links which are not repeated below.
- A quiet revolution is recent opening of internal interfaces to all Microsoft technologies including Excel Word and PowerPoint. This is needed for COM to work properly -- all of these systems must be built out of documented components. This allows you to build shared (say with Tango) Word editors; Java viewers for any Microsoft application, a database version of PowerPoint which assembles PowerPoint presentations dynamically from individual foils.
- We can classify "Synchronous Collaboration" activities into the following broad areas
- Support of Synchronous Teaching as in Syracuse-Jackson State Experiment
- Support of Synchronous Collaboration for scientific research and administration. This is support for what Bill Wulf calls a collaboratory. Note this somewhat harder than synchronous teaching in that the interactions are less structured and this implies the infrastructure must be more robust and support universal deployment
- Support of sophisticated systems such as those needed in Crisis Management (Command and Control) or modern constructivist views of education (see
http://www.cilt.org). Here Tango can be viewed as software infrastructure and "support" corresponds to porting of sophisticated applications to Tango. These include mapping and situation assessment tools for command and control; systems like the Biology workbench or Belvedere in education; shared computational steering or program development support in collaboratories. In education case, the physics simulations were ported illustrate this. In collaboratory case, we can illustrate with the current shared Emacs and potential shared visualization tools. We developed with XII (http://www.niusr.org/), a shared 2D mapper with links to database of situation reports.
- We can classify "Asynchronous Collaboration" activities in the following ways. Note that as asynchronous training is based on students browsing web pages and so sophisticated web environments are included here.
- "Training or University Server" which would typically be a web-linked database supporting
- Personal and Performance data (PAPI) including grades
- Educational Objects including HTML pages, applets and links to sophisticated servers for simulations
- Storage of teacher and student (homework and notes) material
- Specification of the "digital library-like" information which includes the so-called metadata defining Universities, training programs, courses, lectures etc. in somewhat sort of hierarchical fashion.
- Multimedia information including archival audio (and possibly video) linked to lectures
- Resources such as glossaries and other support material
- "Specialized Servers" such as "Virtual Programming Laboratories" as developed by NPAC and Cornell or generalized web servers as those developed by NCSA for education. Lotus Notes as in Alliance Intranet illustrates a more general shared document resource. This provides a richer storage capabilities than a typical University Server "Homework support".
- Note that web-linked database technologies such as Cold Fusion and JDBC are relevant to either "Training/University" or "Specialized" servers
- Finally we have special asynchronous support tools
- Threaded Discussion Lists or just ordinary email
- Shared Event calendars and Scheduling Tools
- Tools for shared document preparation in asynchronous fashion where individuals check out material.
- We note that we view "shared browser" capability of a synchronous collaboratory as allows one to conduct "guided tours" synchronously of the same back-end resource needed for asynchronous work. Thus we have three basic categories of technologies/material
- Back end Content in Databases
- Asynchronous Tools
- Synchronous Tools
- Where is Tango? So Tango has demonstrated all the needed capabilities to support all styles of synchronous collaboration. Currently major effort is to increase robustness so that can support all modes with "real users". The major difficulties are variable Internet quality of Service and erratic changing Netscape browser support. The Internet can be addressed effectively by using DREN but Netscape still has many bugs and uncertainties noted below, make it hard to have a very strategic plan.
- They do not have full support for Java 1.2 and the latest AWT including the excellent "swing" classes and AWT1.1 event model.
- Maybe Netscape will not stay in browser business
- There is an important set of HPC core training content, which may not be gotten without a focussed effort. This includes
- HPC Specific general documents such as HPF and MPI definitions, review articles and like material. The CEWES CDROM project aims to collect this material and so does NHSE (National HPCC Software Exchange)
- HPC Glossaries. Note there are several ones on web for general computing such as
http://www.matisse.net/files/glossary.html or http://www.packardbell.com/glossary/ or http://www.pcwebopaedia.com/
- We also need raw material such current MPP product definitions as in NHSE (NPAC collected this 4 years ago but needs to be updated) or even Yahoo at
http://www.yahoo.com/Computers_and_Internet/Supercomputing_and_Parallel_Computing/
http://nhse.npac.syr.edu/ contains current versions of NPAC HPCC System surveys ("roadmaps") and glossaries.
- Structured data such as glossary items or HPC System definitions should be stored in a database or using XML with custom tags. Currently we use HTML to define a glossary but XML allows us to separate the display and content of the data. One can easily map relational database tables into XML or vice versa. One can view XML as a convenient file based alternative to a web-linked database for this type of information.
- Note vision includes Education at HBCU/MI Institutions as a goal and so support of education (as well as training is uncontroversial. The need for education for DoD users is less clear although there is some need which is naturally associated with Universities local to MSRC's (or user location). This needs support for virtual collaborative universities, as curricula material is likely to come from other sites (such as PET Universities).
- In ASC May 6 1998 meeting we identified 4 forms of training material which could be put on the web
- Basic PowerPoint or HTML Summary material (foils)
- Summary material with elaboration from recording (audio and perhaps video) of associated lecture(s). This is called "Audio/Video over foils" and it is not clear if editing of multimedia material important
- Cornell Virtual Workshop enhanced material with multiple levels of material and associated resources such as glossaries, programming laboratory, illustrative applets etc.
- Classic instructional CD-ROM (which can of course be a web site) with rich professionally produced multi media material
- Note that care should be taken in producing any sophisticated material NOT to use HTML, as this is very difficult to integrate into databases or with systems like Tango. This is related to remarks in 7) above. Sophisticated applications require an understanding of the structure of the information. This can sometimes be guessed from an HTML version but only with an error prone process. XML with carefully designed customized tags is an approach that is much easier to keep up to date and to integrate material with latest tools.
- So the PET project to produce high quality curricula material should design the XML or equivalent database structure first.