欢迎来到天天文库
浏览记录
ID:7295937
大小:592.28 KB
页数:12页
时间:2018-02-10
《when databases attack a guide for when to stick to files》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库。
1、CHAPTER12WhenDatabasesAttack:AGuideforWhentoSticktoFilesTimMcNamaraMyMastersdissertationstillfeelslikesapersonaldefeat.Atleastfourmonthsofthenine-monthprojectweresabotagedbecauseIdidn’tunderstandtheimplicationsofthetechnologychoicesthatIwasmaking.Thischapterwillbeabitofapostmortemoftheprojectan
2、dawalk-throughofastrategyIshouldhaveused:storingplain-textdataon-disk,insteadofinadatabase.HistoryButfirst,alittlemoreaboutmystory.In2010,IwasundertakingaMastersinPublicPolicyatVictoriaUniversityofWellington.Iwasfocusingmyeffortsontheargumentssurroundingopendatawithinthesciencesector.Specifical
3、ly,Iwantedtoknow:dotheargumentsthatacademics,officials,politicians,andthepublicalign?Ihadthesensethatopendataandopengovernmentmeantquitedifferentthingstodifferentpeople,andIwantedtoquantifythat.Gettingaccesstoinformationaboutwhatofficialsthoughtaboutopendatawasfairlyeasy.TherewasamajorreviewofN
4、ewZealand’spubliclyfundedresearchbodiesbeingundertakenatthetime.PolicyadvicebetweendepartmentsisavailableundertheOf‐ficialInformation1982,whichisNewZealand’sfreedomofinformationstatute.Infor‐mationfrompoliticiansareeveneasiertofind,astheytalkallthetime.Allpressreleasesaresyndicatedviascoop.co.n
5、z.Butwhatabouttheviewsofthepublicatlarge?Thepoliticalblogosphereisveryactive.Whileblogshavediminishedinqualityanddepthwiththematurityofsocialmedia,passionatepeopleseemtodrawthemselveswillinglyintotheflameofpoliticaldebate.151BuildingMyToolsetIneededtolearntextanalysisandnaturallanguageprocessin
6、g(NLP)techniques,butmysupervisorandIwereconfidentthatitwouldbepossibletodrawsomeinferencesfromthislargecorpusofmaterial.ThegeneralapproachwastocrawltheNewZealandpoliticalblogosphere,identifydiscussionthreadsaboutopendata,identifypeoplewhocommentonthistopic,thenpulloutthethemesofthesecomments.Iw
7、asalsointendingonspendingsometimeclassifyingcommenters.Perhapswecouldseeiftherewerethingsspecialaboutcommentersinterestedinopendata.Thedissertationitselfwasasuccess,butnoneofmyworkfromtheblogospherecouldbeincluded.Partwaythroughth
此文档下载收益归作者所有