NHSE ---- Computational chemistry repository is in pretty good shape. Just a few technical things that need to be done before David and Yang can deal with it themselves. One is to set it up using the Repository in a Box, which I will do as soon as the next version becomes available, supposedly in a week or two. I will be helping to beta test it while doing this, before it gets an official release, probably sometime in January. Some other technical issues such as making AskNPAC About Chemistry robust are going to be discussed at a meeting with Gang and systems early next month. The random number generator repository is proceeding fairly well. We've discovered some more software from the Web, some of my software is now packaged and documented, some is still being worked on. Ko is working on checking and testing all the code so that it can be given "reviewed" status (i.e. checked and evaluated in an NHSE Review paper) when it is added to the HPCnetlib software catalog. I am doing a few last edits on content and HTML for my NHSE Review article. Ko is also working on some new and improved parallel generators (and tests), especially for HPF. I am still editing Saleh's work on a physical optimization roadmap, and this may take some time. He has also found and cataloged some software. The goal is to add this to the large software catalog that the NEOS people at Argonne have come up with, and use this to greatly improve the optimization section in netlib and HPCnetlib. In January I plan to start working on HPFA, to integrate it with NHSE using the RIB, to add more codes, and to add evaluations of code and compilers that Hon did before he left, and to improve roadmap stuff, in particular to include the latest version of Hon, Sanjay, et al's paper. I will also be thinking about new repositories. I would like to do a Monte Carlo repository (which would include random numbers as a subset). John Salmon mentioned to me that he might be interested in doing an astrophysics repository, so I'll talk to him about that. A more general particle dynamics repository might be useful (N-body, PIC, etc). CFD would clearly be a good idea. But in all these cases, we would have to collaborate with other groups, perhaps augmenting work in DoD modernization program. I plan to talk to David and others (e.g. Danny Powell?) to get their ideas about this. Another thing I was thinking of doing was to look carefully at netlib and HPCnetlib to try to find any areas that are missing or lacking quality and/or up-to-date software, particularly in HPCnetlib. I think there are some things such as graph algorithms, sorting and searching, etc, that are not hard-core "numerical algorithms" and may not be well represented in netlib, but are of general use. I know that the random number generator section of netlib was quite poor, and for HPCnetlib was non-existent, and I suspect the same is true of other areas. So this would not be another repository, but it would add a lot of useful code to NHSE. I am also slowly fixing up the roadmap. I plan to look through HENSA to try to find useful stuff to add to it. CSNet ----- Gary says he has a deal pending (if the lawyers can come to an agreement) with MCI. By the time that rolls out (probably a few months from now), he wants Webboard, LivingPage, and a couple of other similar things to be "productized". Before then, there are some fixes and additions needed in order to get the functionality pretty much how we want it, and to make the codes more modular. I'm hoping this will only take another month or two, then another couple of months to do the documentation, prettification, improving the user interface, fine-tuning, etc. I'm not sure what the status of the distance learning proposal to New York State is. Gary's been busy with the MCI deal and installing new POPs and I haven't gotten around to talking to him about it. I will bring it up when I see him next Friday. I mentioned to Denny that perhaps he could talk to Gary about getting some money from Gary, and he said he'd talk to him. We certainly should be able to get some if and when the MCI deal goes through, but I would think we should be entitled to at least a small amount from the New York State Constellation project money. I am currently spending an average of about 25% of my time on this stuff. Other stuff ----------- Denny wants me to demo Tango to a few people who might be interested, so I'm going to set it up on my laptop and see how the Windows version works and how it does over a 28.8 connection (often much slower since the Internet from NYC to Syracuse can sometimes be much slower than modem speeds). Alvin is still writing up. I expect he will have a first draft by end of January. I am aiming to give him his qualifying exam at the end of this month, and he's going to give me a couple more chapters. The chapters he's given me so far have been very good. His written English has improved a lot. Saleh and I are still working on final edits to his paper. I think we have finally sorted out all the problems and should have it ready to submit for publication next month. He is also going to submit a paper summarizing both simulated annealing and mean-field annealing results to a conference on scheduling problems to be held in Toronto next year. I've been corresponding with Su about his cluster labeling paper, and fixed all the problems except for his "proof" of the complexity of the algorithm. He's given me a few revisions but none have been acceptable. He just sent me his latest version which I'm going to study over Xmas. If we still can't come up with something solid, I'm just going to remove the proof from the paper and just say that the empirical results show that the algorithm is O(log N) (just as John and I did in our previous papers, since we couldn't prove it) and submit the thing to be published, since it's dragged on way too long. If we find a good proof at a later date it could be incorporated into Su's second paper, which still has some more editing to be done and probably won't be ready for a few months yet.