This memo is meant to help discussion of future of NPAC. There are two major issues 1) NPAC cannot preserve its intellectual property as this is owned by some unclear combination of the University and those associated with NPAC activity in question. Portland Group, Carrier, Careweb, VDCE, Financial modeling illustrate the difficulty in different ways 2) NPAC cannot naturally fund its “overhead” or infrastructure. One needs to address both these questions in any plan for NPAC. We believe that both problems are so hard that there is no evolutionary road to solution and suggest that if one wishes to preserve important parts of NPAC activities, one needs to formally close the organization – declare success – and start a new organization or multiple organizations with enterprise plans and groundrules that address problems 1) and 2). We give a few details on basic issues and then discuss some scenarios/options. In general we stress that NPAC alone cannot solve these difficulties and cooperation is required from the University – if only in clarifying that a certain approach is nonviable. I will refer to a generic future center as newNPAC. 1)Intellectual Property Issues There are many possible enterprise (business) models for “centers”. One is typified by a (small) business with a group of people working together with a clear understanding that business itself (not the director or programmer who wrote code nor manager of project) owns any artifacts produced by the enterprise. The enterprise could owe a fraction of revenues on any product to an outside entity such as the sponsoring venture capital group. A second model is that of the University department where largely independent faculty do their own thing and own (after University takes an agreed fraction) their own thing. The CASE center is organized accordingly to the second model. NPAC’s funding model is that of a small business as it is largely responsible for supplying its own funding based on concepts and artifacts that it invents (Web computing ideas in VDCE) develops as investments (e.g. database technology used in Carrier, Careweb, or financial modeling) or from a previous contract (Compiler technology in Portland Group). However the IPR (Intellectual Property Rights) rules are applied as though NPAC was a department. This mismatch between funding and IPR enterprise models is a fundamental problem for NPAC. Indeed much of ECS faculty and others (e.g. Dean of Nursing) concerns and dislike of NPAC stem from their perception of it as a department without understanding that it’s funding model prevents it from operating as a department. I believe that one can only run a center if IPR and funding model largely agree. If a future newNPAC is to be of the general size of NPAC, I believe it must operate with the rules of a small business. Employees of newNPAC must assign their IPR rights to the organization. One can split NPAC into components with different rules – a largely University funded (via tuition/base budget) distance/continuing education unit can operate with traditional University department rules. Finally we note that the situation is not totally lost. Various horses such as Careweb have fled the stable under current IPR rules. However NPAC’s crown jewels remain including WebWisdom (generic nifty Web education technology), TANGO collaboration and related artifacts from Rome Lab CIV project, Web based Neat Tools from work of Lipson and Warner. These will be lost (to either NPAC or newNPAC) through current leaky IPR sieve unless we change mode of operation very soon! 2)Overhead Issues Consider the total cost to NPAC of Computer systems Support, Administrative Staff, Equipment Maintenance, Operating Expenses (from paper to phones and SU network charges). In 96-97, this is estimated as $966K including fringe benefits but not overhead and not including charges earlier this year for staff who will not be replaced and assuming CASE picks up all of Woodcock’s salary. Of this money, $386K is borne by New York State through InfoMall and UDC grant (which included maintenance). Note adding 15% overhead leads to a total State charge of $444K which leaves (from $635K total state budget) $190K to directly apply to InfoMall’s mission (economic development). These numbers indicate that InfoMall is critical for supporting NPAC’s infrastructure and that it is quite hard to deliver perceived value for money unless State understands that InfoMall has the role of providing infrastructure. There is another critical role for InfoMall and Living Textbook funds. Thus NPAC currently is a partner in 3 of NSF supercomputer center recompetition proposals (Cornell, NCSA and San Diego) with a total budget for NPAC of about $900K per year – about 60% for education and 40% for HPCC research. This competition required 1:1 institutional matching. Currently InfoMall and Living SchoolBook supply most of the match. I do not expect to get this full requested budget but if Cornell is chosen (where our budget is $600K per year) we should get a lot of it as we are critical to their proposal. (In my humble opinion). If we add InfoMall/UDC/directly budgeted items in all NPAC’s state and federal grants, we find a total (with fringe but no overhead) of $746K. The difference ($966K-$746K plus earlier charges this year) is borne on discretionary account and re-budgeted grants. Note that in 96-97 NPAC’s grants call for a salary expenditure of $1.92M (no overhead but including fringe) on “researchers”. This will be underspent (or spread by no cost extensions to grants) by a total of $686K (35%) due to unwillingness/inability to hire due to NPAC’s uncertain future and need to re-budget to support infrastructure. Note the net result that NPAC is understaffed on essentially all projects. We have in consequence to work hard and that is a problem for all NPAC researchers. Can one reduce these overhead costs? One way is to drastically change or reduce NPAC’s mission. Fox can certainly run a sweet $250K per year research program which pays his summer salary and a student or two and needs no infrastructure. I doubt if this is very interesting to either Fox or the University. In some perfect world, much of the infrastructure would be supplied by the University as a benefit of “overhead” charged to grants. However NPAC’s infrastructure (people and machines) are not idle and not significantly lower in quality than counterparts in ECS. (Maybe higher in some cases). Thus I don’t see it as realistic to transfer infrastructure responsibility to ECS. As discussed in options section one can capture components of NPAC (course teaching) and support through some modest augmentation of ECS. I personally know of no way of running an enterprise like NPAC without infrastructure comparable to that we now have. I note that previous managers Deborah Jones, Diane Wigand and Gene Woodcock have not proposed any changes to NPAC which would reduce costs significantly. Indeed NPAC’s infrastructure is probably more efficient now than it ever has been as much has been trimmed. Once one decides on next steps for NPAC, I do think one could focus infrastructure and reduce costs a little. However I would expect the infrastructure cost still to be about $850K per year (reduced from $966K). This does not qualitatively change scope of problem. (Note that grants such as vBNS connection will help infrastructure directly but again these are relatively small items). I also expect that this infrastructure could support a larger research effort than we now do as NPAC has cut- back research costs rather more than infrastructure as reduced staff still needs a variety of central machines as well as expertise of particular infrastructure people. 3)Options for NPAC’s Future First I note that it will be hard to save NPAC – defined in general as a nationally significant enterprise led by Fox. Thus I would recommend first identifying components that are internally viable and making certain these survive. These include a)An enterprise at SU working with Warner and naturally led by Lipson – call this NeatCenter. This should use “small business” IPR rules. b)An enterprise at SU following on the highly successful Web courses CPS606,616 ECS400 and using good if not bleeding edge research Web techniques in teaching. This could be led by McCracken. This should use University department IPR rules. c)Individual researchers able to raise money from outside sources. Clear candidates here are Furmanski and Bernholdt. This concept is obviously problematic in the long run unless they get regular faculty positions at SU. Both individuals have the talent to be good candidates in ECS and Chemistry respectively. They may not of course have desired expertise or be competitive on a particular head to head comparison if a suitable spot opens up. If we can set up newNPAC, this could subsume/ link synergistically with the above sub- components. I propose that the key to newNPAC is to set up: d)An organization similar to CITE proposed earlier by Fox, Goldberg and Lipson. This would in its research component enforce the “small business” IPR rules. It is different from b) in that it would have substantial research both in education and related areas. Note that NPAC’s work in education technologies has stemmed from either HPCC funding (see our submissions to NSF supercomputer re-competition) or the Rome Laboratory CIV or from working with businesses (InfoMall). Thus most of what NPAC does now can be incorporated as a research arm of CITE. University funding of base department in CITE would reduce reliance on InfoMall but I believe InfoMall will be essential initially and that newNPAC=CITE would do excellent technology transfer as its research would be very relevant to industry!