Open Source Mathematical Software
OpenSource
MathematicalSoftware
Mathematicalsoftwarehasgreatlycontributedtomathemat-icalresearch,enablingexcitingadvancesinmathematicsandprovidingextensivedataforconjectures.Perhapsthreeofthemostwell-knownapplicationsofcomputationtomathemat-icalresearcharetheresolutionofthefour-colorconjecturebyAppelandHakenin1976(thoughitisnowreprovenwithlessneedforcomputerveri cationbyN.Robertson,D.P.Sanders,P.D.SeymourandR.Thomas),ThomasHales’sproofofKepler’sconjecture,andtheformulationoftheBirchandSwinnerton-Dyerconjecture,whichgrewoutofextensivenumericalcomputation.
Opensourcesoftware,suchasTEX,MozillaFirefox,andLinuxhashadaprofounde ectoncomputingduringthelastdecade,andwehopethatopensourcemathematicalsoftwarewillhaveasimilarpositiveimpactonmathematics.
Ithinkweneedasymbolicstandardtomakecomputermanipulationseasiertodocumentandverify.Andwithallduerespecttothefreemarket,perhapsweshouldnotbedependentoncommercialsoftwarehere.Anopensourceprojectcould,perhaps, ndbetteranswerstotheobviousproblemssuchasavailability,bugs,backwardcompatibility,platformin-dependence,standardlibraries,etc.OnecanlearnfromthesuccessofTlikeMacaulay2.EXandmorespe-cializedsoftwareIdohopethatfundingagenciesarelookingintothis.—AndreiOkounkov,2006Fieldsmedalist(see“InterviewswiththreeFieldsmedalists”NoticesoftheAMS,54(3)(2007),405–410).
Thetermopensourceisde nedatdoc.guandang.net(includingcommercialcompaniesorthedefensedepartment)shouldbeabletoinspectopensourcesoftware,modifyit,andshareitwithothers.
Onekeydi erencebetweenmathematicaltheoremsandsoftwareisthattheoremsrequirelittlemaintenance,where-asmathematicalsoftwarerequiressubstantialandpotentiallyexpensivemaintenance(bug xes,updateswhenalgorithmsorlanguageschange,etc.).Mathematicalresearchusuallygen-eratesnodirectrevenueforresearchers,andlikewiseopensourcemathematicalsoftwareisfreetoshareandextend,soitrarelygeneratesrevenue.Volunteere ort,donations,and nancialsupportfromtheNSFandotherorganizationsisthuscriticaltothesuccessofopensourcemathematicalsoftware.
ThereisaproofinthearticlebyCampbelletal.inTheAtlasofFiniteGroups—TenYearsOn(1998)thatdescribeshowmanyseparatesoftwarepackageswere“easilyused”todeducevariousmathematicalfacts—nocodeisgiven,andsomeoftheprogramsareproprietarysoftwarethatrunsonlyonhardwaremanyyearsoutofdate.Suchproofsmay
becomeincreasinglycommoninmathematicsifsomethingisn’tdonetoreversethistrend.
SupposeJaneisawell-knownmathematicianwhoan-nouncesshehasprovedatheorem.Weprobablywillbelieveher,butsheknowsthatshewillberequiredtoproduceaproofifrequested.However,supposenowJanesaysatheoremistruebasedpartlyontheresultsofsoftware.Theclosestwecanreasonablyhopetogettoarigorousproof(withoutnewideas)istheopeninspectionandabilitytouseallthecomputercodeonwhichtheresultdepends.Iftheprogramisproprietary,thisisnotpossible.Wehaveeveryrighttobedistrustful,notonlyduetoavaguedistrustofcomputersbutbecauseeventhebestprogrammersregularlymakemistakes.
IfonereadstheproofofJane’stheoreminhopesofextendingherideasorapplyingtheminanewcontext,itislimitingtonothaveaccesstotheinnerworkingsofthesoftwareonwhichJane’sresultbuilds.Forexample,considerthefollowingquotefromtheMathematicatutorial1:
ParticularlyinmoreadvancedapplicationsofMathematica,itmaysometimesseemworth-whiletotrytoanalyzeinternalalgorithmsinordertopredictwhichwayofdoingagivencomputationwillbethemoste cient.[…]Butmostoftentheanalyseswillnotbeworthwhile.FortheinternalsofMathemati-caarequitecomplicated,andevengivenabasicdescriptionofthealgorithmusedforaparticularpurpose,itisusuallyextremelydi culttoreachareliableconclusionabouthowthedetailedimplementationofthisal-gorithmwillactuallybehaveinparticularcircumstances.
Nojournalwouldmakeastatementliketheaboveabouttheproofsofthetheoremstheypublish.Increasingly,pro-prietarysoftwareandthealgorithmsusedareanessentialpartofmathematicalproofs.ToquoteJ.Neubüser,“withthissituationtwoofthemostbasicrulesofconductinmathemat-icsareviolated:Inmathematicsinformationispassedonfreeofchargeandeverythingislaidopenforchecking.”
Fulldisclosure:Thesecondauthorstartedanewmath-ematicssoftwaresystemin2005calledSAGE(seedoc.guandang.net),whichcombinesPython,GAP,Singular,PARI,Maxima,SciPy,etc.withseveralhundredthousandlinesofnewcode.SAGEreceivescontributionsfrommanymathe-maticiansworldwidethatsynthesizethelatestalgorithmsfromabroadrangeoftopicsintoacomprehensivetoolkitformathematicalresearch.
—DavidJoyner
U.S.NavalAcademy,Annapolis
wdj@usna.edu—WilliamStein
UniversityofWashington,Seattle
wstein@u.washington.edu
11页 2财富值 Open source software – ... 19页 免费 Open Source Cloud Softwa... 暂无评价 25页 免费 Open Source Mathematical... 暂无评价 1页 免费搜...
Open Source Mathematical... 暂无评价 1页 免费 FREE OPEN SOURCE SOFT... ...source softwareopen-source softwareopen-source softwareopen-source software隐藏>...
25页 免费 Open Source Mathematical... 暂无评价 1页 免费 FREE OPEN SOURCE SOFT... 暂无评价 14页 免费 Open source software – ... 19页 免费喜欢...
17页 免费 Open Source Mathematical... 暂无评价 1页 免费 FREE OPEN SOURCE SOFT... 暂无评价 14页 免费 Open source software – ... 19页 免费喜欢...
Reliability Evaluation Metrics for Open Source Software Selection_专业资料。Vo. o3et0011N.,Sp.21 Junlfala1n dec n ntmanai ora ks1etS Vo...
Open Source Software 暂无评价 12页 免费 open source cloud comput... 6页 ...11页 2财富值 Open Source Mathematical... 暂无评价 1页 免费搜...
CLIP 302.0 The Amos Project An Approach to Reusing Open Source Software ...For example, a suite of mathematical software may have inside packages which...
open source of software 29页 2财富值 Economics of Open Source... 27页 免费 Security of Open Source ... 17页 免费 Open Source Mathematical... 暂无评...
OpenSim Open-source Software to Create and Analyze Dynamic Simulations of Movement...fidelity of the underlying mathematical model of the neuromusculoskeletal ...
热门文档