Comments on Gateway Materials 1) Organization – this needs to be made simpler and smaller with a distinction between advisory and decision making roles; between management and technical design decisions. The flow of requirements to technical decisions to implementation planning and deployment needs to be described. The total “cost of management” should be less than around 20% of total. I have separately suggested a different plan from that described. Namely: Project leader -- Perhaps Fox Project Manager -- Asbury 3 Major Teams Back End -- ASC and OSC Middle Tier NPAC Front End lead OSC (Ken) Minor Teams Database and Object structure of documents/entities Tennessee NetSolve Individual PSE's Minor Teams report to parts of major hierarchy Technical Leads of teams can be called design team but don't think concept needed. 2) Front End Organization: I believe there was discussion of and agreement that we focus on a single initial PSE (namely CCM) with Ken being lead of the PSE front end activity. We should involve the other PSE teams in evaluation but postpone detailed implementation until we get early experience from CCM and most importantly a design template on which we should build multiple PSE’s. This template (design framework) should be led by Ken with participation from other PSE’s. 3) Database: This is an area dear to my heart as it relates to my experience with training material. Namely here we need a database to collect *) Type of material already in your Oracle database describing users *) Specifications of backend systems and services needed by middle tier and front end user interfaces *) Components of documentation and related material accessible from front end. I recommended being careful NOT to write such material in HTML but rather use the formal document object model (perhaps with XML) so that one can capture key components which will allow better customized displays which adjust as web technologies continue to evolve. Remember HTML is not designed to define structure of a document but rather its display *) Archiving of sessions *) Scientific data I believe that this is a significant activity, which needs a technically strong team, which combines technical expertise of NPAC in web-linked databases, W3C DOM and XML as well as OSC front end and ASC backend experience. We should consider this database as an important resource to be managed by middle tier 4) There is a set of front-end displays, which are generic and potentially very complex – I refer to batch queue status and similar displays. There is a proposed middle tier interface in these areas. Whose responsibility is front-end display for these? 5) Back-end Testbed: It seems wise that system be tested in multiple places as well as at ASC. Otherwise there may be some errors, which are not detected as only show up in some setups. This seems controversial as suggests OSC testbed helpful and for some reason that appears to be a problem for Nicholls. 6) Level of Effort: I found the overall level as alarmingly large. I think that a more modest effort would be appropriate to start with and that as described for PSE activity it would be advisable to use a staged approach. In looking at level of effort estimate, I was unclear on two further points – do we count students as 1 LOE or as funded – namely 0.5 LOE? Secondly the efforts at NPAC and I am sure also OSC require substantial technical support to succeed. Is this part of LOE or some additional implied overhead? 7) Haupt Material I: I didn’t understand comment that security on “submit a job” would be a significant performance overhead 8) Haupt Material II: I would have thought we should support multiple CA’s from the start. Otherwise it will be very difficult to test and interoperate systems. 9) I think we should study carefully before detailed implementation the possible relevance of JINI and whether we should involve this technology and perhaps a collaboration with Sun Microsystems.