- creating wiki-like resource for sw available for NSF research, and possibly even communicating some kind of rating system - how reliable this is, start that with this project
- agreement to put information about software on the wiki by some!
- Survey - somone post question, get a lot of feedback - on all of the issues - licensing, distribution, feedback
- a maillist -
- perhaps a single forum, and split if traffic is too heavy
- forum to announce tutorials? forum for releases?
- advertising for students, jobs, postdocs?
- annual report - impact on my discipline, impact on others - get language on software, useability, - not sure how easy to find this in annual reports, but - would be good for template on not just information for NSF, but information to report -
- one more suggestion - do you have any nuggets - picture, and blurb on what you are doing - could this be adapted to "do you have any software" -
- maybe as a community, do a web page, disseminate around the directorate
- public web presences, start with this group, and then not limit it -
- how do you do this
- do you want to start this in wikipedia, for the rating system?
- involve some software engineering experts, get comments from them -
- monthly seminar on sw - could be a way to have information, sw engineers, sociology, provide some more input -
- structure in a way with education component - maybe speakers with interesting ways to do this stuff eg sw eng, and some headliners -
- have quite short talks, that lead to a discussion -
- 2 or three major meetings (SC, XSEDE12, APS, ... pick 2 of them every year, and do a half day or 2 hour session there -
- for XSEDE summer conference, in July, monday is tutorials, happy to entertain tutorial proposals from this group, have BoF sessoins -
- virtual seminars, have very short talks, on some topic - regression testing, micro talks - this would work very wlel
- how do you present -
- can record the sessions as well -
- lots of experience with software engineering, code reviews on other people's code, feedback on your codes - do people welcome this feedback -
- the big value would be making code that other people could read
- value of code reuse is architectural, seeing how the module works, not just syntax -
- regarding XSEDE's conference, if we were able to take a 2 hour session in one of our tracks, include this in the call - what would this look like to this group - submit something to present to each other, if you could come up with some notions there,
- scientific software track?
- mechanisms for getting credit
- recognition for students - award for best scientific software of the year -
- should it be domain specific?
- have prizes at the XSEDE conference already, could do something like this for scientific software
- poll of XSEDE users, 6,
- poll of DOE incite - 2-3
- others: use local resources?
- publishing - any venues to publish, details on the code?
- there was an Amber publication, now in 2009 in JcompChem - every decade - special issue of jcompchem on MD codes - not details though
- actual citation for Amber, this goes into the citation, do let you
- flash: parallel computing -
- concurrency and computation practice and experience
- journal of scientific computing
- international journal of HPC apps - large scale scientific simulation series on cutting edge platforms, doing modifications to code in flight - the way you have to plan simulations...
- differences? parallel computing is about code architecture, lagrangian framework on top of eulerian framework - parallel - IO, computational science and discovery,
- Computing in science and engineering also have special issues -
- Bioinformatics - have software tracks, a whole bunch of proliferating journals with sw tracks, but reviews can be strange -
- MD - nobody really publishes the software, talk about theory, mention implementation, manual indicates what citation to use if you use accelerated portion of code -
- publish something new, new algorithm - mention the software as a consequence of the new item -
- this happens with flash too, publish new algorithm, mention the software -
- most cited paper for flash, 2001 AP J -
- and review papers of the software - amber, and MD in journal, people mistakenly use this as citation for amber -
- people cite by version -how do you decide who should be on the list -
- each PI decides who should one the list, need to contribute for a couple of years a substantial contribution - 30 people -
- component paper - authors of code left field, would they be authors of a subsequent paper?
- in GAMESS, have main citation for GAMESS - if download GAMESS, users agree to use this citation - and per method, put in citation for each method (documented in GAMESS manual) - paper associated with each new theory, this is how contributors are acknowledged, and all contributors to GAMESS appear in the top of the output of the code
- postdocs: 2 are here - how do you feel about your career path in this, are you encouraged to write software -
- Sarom did undergrad at George Mason University (GMU) and graduate work at Iowa State University (ISU), happy to write code - part of the graduate work in the quantum theory research group at ISU
- just defended, in first year of postdoc, looking at gpu porting of methods available in GAMESS - hope this means you won't need a large supercomputer in order to pursue a career in academic research
-
-
{"serverDuration": 81, "requestCorrelationId": "5c75407a2ff26bb1"}