UnNews:NewTextCompressionAlgorithmInvented

ThisarticleispartofUnNews,yoursourceforup-to-the-microsecondmisinformation.

14August2006

Thisfragmentofa4th centuryGreekmanuscript provesthatthecompression algorithmwasknownto theancients,buthasn't beenrediscoveredtillnow.

ITHACA,NY--ResearchersatCornellUniversityinupstateNewYorkunveiledanewtextcompressionalgorithmtoday.Theirfindi ngs,publishedinthismonth'sissueoftheacclaimedJournalofCompressologyarethefirstmajoradvancementinthisfieldindeca des.Computerscientistsworldwidehailedthenewmethod,sayingitwillusherinaneweraofstoringandmanipulatingdata.Leadr esearchersJohnArdellexplainedthealgorithminlaymen'sterms,"Throughacomplexparsingprocess,thetexttobecompress edisanalyzed,andallblankspacesareremovedfromit.Thisresultsinalosslesscompressionofabout10to20percentonaverag e."Expertssaythemethodisbrilliant,especiallysinceitcanbeusedontextwritteninanylanguage-notjustEnglish.Previousatte mptsattextcompressionhavefocusedonletterandwordpatternscommontospecificlanguages,butthenewalgorithmistrulyun iversal.Workonthealgorithmbegantwoyearsago,whengraduatestudentTerryMatthewsstumbledontheideawhileinstantme ssagingcolleagueJoeFraser."Imadeatypowhiletalkingtohim,"heexplains,"andaccidentallyleftoutsomespaceswhiletypin g'sodidyouhookupwiththathotgirlatthepartylastnight?'Tomyamazement,Joewasabletounderstandmyquestiondespitemy mistakes."Thestudentsweren'tsureatthetimehowthecompressionhappened,butthroughdiligentworkwithProfessorArdell,t heexactalgorithmwaseventuallydiscovered.Historiansclaimsthatancientcivilizationsmayhaveactuallyusedasimilarversio nofthealgorithm."Writingmaterialswerescarce,"explainsPrincetonProfessorMcSnobly,"soearlyGreeksandRomanswrote manytextswithoutusingspacesorpunctuation.Upuntilnow,howtheydidsowasamystery-theknowledgemusthavebeenlostd uringmedievaltimes."Thecompressionmethodwillhaveawiderangeofpersonalandcommercialapplications.Institutionsworl dwidearerushingtoimplementthenewalgorithm,whichisexpectedtosavemillionsofdollarsworthofharddiskspace.UnNewsis oneofthefirstmediaorganizationstodeploythealgorithm,whichisnowappliedtoallpublishedstories.WikiNewsvowstoswitchoversoon,butsofarhasbeenplaguedbytechnicaldifficulties.

• "" "CornellUniversity". "", August14,2006