Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!sharkey!atanasoff!mehl From: mehl@atanasoff.cs.iastate.edu (Mark M Mehl) Newsgroups: news.groups Subject: Re: CALL FOR DISCUSSION: comp.math and comp.stat Message-ID: <2066@atanasoff.cs.iastate.edu> Date: 3 Dec 89 02:57:45 GMT References: <8911281924.AA15699@ncar.UCAR.EDU> <551@tauceti.Morgan.COM> Reply-To: mehl@atanasoff.cs.iastate.edu Organization: Dept of EE/CprE -- IOWA STATE UNIV Lines: 34 amull@Morgan.COM (Andrew P. Mullhaupt) writes: >The proposal for comp.math seems to address issues already dealt with >by the existence of sci.math.num-analysis and sci.math.symbolic. What >do we need comp.math for that we don't get from these already? I agree 100%. Discussions about LINPAK or ODEPACK should be held in sci.math.num-analysis. If this group needs to be split so as to "attempt" to separate the theoretical issues from the programming issues, then a new group like sci.math.num-analysis.software (or s.m.n-a.theory) should be created. This hierarchical nomenclature also simplifies the initialization file for the nn newreader (~/.nn/init) because the nn user simply has to include: sequence sci.math.num-analysis to read all sci.math.n-a.*.* groups in that hierarchy. The same argument holds for the proposed comp.stat news group. I think a separate news group to discuss statistical software would be useful, but name it: sci.math.stat.software (or s.m.s.num-analysis.software). This would be the place to hold discussions about BMDP, SAS, and SPSSX. Most of the groups in the comp.* hierarchy are directed towards computer science and computer engineering theory and practice. Discussions about numerical methods (and its practice) most appropriately belong under a math hierarchy rather than a computer hierarchy. -- /\ Mark M Mehl, alias Superticker (Supertickler to some) <><> Internet: mehl@atanasoff.cs.IAstate.edu \/ UUCP: {{mailrus,umix}!sharkey,hplabs!hp-lsd,uunet}!atanasoff!mehl Disclaimer: You got to be kidding; who would want to claim anything I said? Brought to you by Super Global Mega Corp .com