WPC j(C-#aHov}-.#aÍ*Ӯ4y#eiRN &͌7L[3~+ {4|){C~, I~2K0&!dԾo-\4}nVȈ;qZ` l߳nk$=!ȋ=laHIEc{ȶb)k՜ ]!^fL z/7p.奁T2=;u'`R4!a\1o_rBUU@Ȅ7]-=ߌnjZ2gUOBW1߃{bbo.EBrv#W^ ߰_X,p]umb26p̊B=E1YpPN@_ =X0J+}26ӣ|S^׷a(CIW~d+lRH,&5!u0#m=޼[sQ\+# %U: 0*# h5 0  k 1  N^ w4 m 0e  1 72B 1ut 72 0c 1~ 72B777UTy""Arial Regular($XXXXUSUS.,$X,X$      =dd=HP LaserJet 5M PCL,,,,0C:\COREL\WORDPE~2\TEMPLATE\CUSTOM~2\ANSIJ3~2.WPTC:\Corel\WordPerfect_Office_2002\Template\Custom WP Templates\ANSI J3 New Document.wpt  'USUS.,$X,X$   _>(# J3/02310r1  X _NCITS_Ԁ/J3ANSIFortranStandardsCommittee>(#CraigT.Dedo , TechnicalChangesNN:(#November13,2002 L (#Page  1  ڀof  2  (3$ !XXUSUS.,$X,X$      =Ndd=  ݁Level 1Level 2Level 3Level 4Level 5Level 1Level 2Level 3Level 4Level 5Level 1Level 2Level 3Level 4Level 5Level 1Level 2Level 3Level 4Level 5Level 1Level 2Level 3Level 4Level 5n- Ql-+R20023|xU )USUS.,$X,X$   _.H8C:\ANSI_J3\2002\02-310r1.wpdHC:\ANSI_J3\2002\02-310r1.wpd/  (3$ !XXUSUS.,$X,X$      =Ndd=   (5hCEKQW]cioAutoList11.1.1.1.1.1.1.1.1.3#37=CIQYag1.a.i.(1)(a)(i)1)a)i)(;3$2#  0  .3  0  (EhCEKQW]cioAutoList21.1.1.1.1.1.1.1.1.ANSI J3 New Document  ( ' $Craig T. Dedo$0Craig T. Dedop17130 W. Burleigh Place Brookfield, WI 53005-2759 00-3nn .    3 2/ Z<Century Schoolbook Regular !XXUSUS.,$X,X$      =Ndd=  _@" TechnicalChanges   To:0  0(#(#04(#(#J34(#4(# From:0  04(#(#CraigDedo` 4(#4(# Date:0  04(#(#November13,2002( 4(#4(# Subject:0 4 TechnicalChanges L4(#4(#  (5h83Date:28Oct2002   To:J3 H  From:RichardMaine l  Subject:Technicalchanges 4  Thefollowingareproposedpubliccommentsrelatingtotechnicalissues. h  1.Ibelieveittobeamistaketointroducetheconceptof T  Ѐ_iostat_Ԁvaluescorrespondingto"harmlessconditions"   ЀintheFLUSHstatement.Thisseemsinconsistentwithall   ЀtheotherIOstatements.Forexample,rewindingor P  Ѐbackspacingafilethatisalreadyattheinitialposition t Ѐmightbeconsidereda"harmlesscondition".Whydoesit < Ѐhavenoeffecttoflushafilethatdoesnotexist,but  Ѐcausesa"harmlesscondition"toflushafilethatisnot p Ѐappropriateforsomeotherreason?Thisrequirementisalso 8 Ѐarguablycontradictedbythefirstsentenceof9.11. \ ̀Furtherthenormativetextdescribingthisconditionusesthe  Ѐadjective"processor-dependent"somuchastomakeit X Ѐpointlessandill-defined.The"suchas"doesnotconstitute  | Ѐausefuldefinition,andthebitaboutinappropriateunitsis D Ѐinanote,soallwereallyhaveisthestatementthat   Ѐprocessor-definednegativevaluesindicateprocessor-defined x Ѐconditions. @ ̀Ithinkthiswholeconceptisapoorlyintegratedlast-minute  ,! Ѐadditionthatstandsoutasanomalous.Itisnotnecessary !" ЀfororuniquetothefunctionalityoftheFLUSHstatement. `"# J3recommendsthefollowingedits: #L% [207:11-12]0 ` Deletethewordsstartingwith orandreplacewith, oraprocessordependent $& negativevalueiftheflushoperationisnotsupportedfortheunitspecified.%'` (#` (# [207:12+]0 4 0` 4(#4(#DeleteNote9.57.'l )` (#` (# 2.IbelievetheCOMPATIBLEroundingmodetobeamistake(and (!+ Ѐtobestrangelynamed).Theonlydifferencebetween h)", ЀCOMPATIBLEandNEARESTreliesonanexactequalitytest 0*#- Ѐforafloatingpointvalue.Furthermore,thisexactequality *T$. Ѐtestwilloftenbeatestforexactequalitytoanumberthat +%/ Ѐcannotberepresentedexactly(whatbinaryfloating-point ,%0 Ѐnumberisexactlyhalfwaybetweenthedecimalnumbers P-&1 Ѐ0.1and0.2?).Thisisthekindoftestthatprogrammers .t'2 Ѐareincessantlywarnedagainstbytextbooksandevenbysome .<(3 _Ԁcompilers.  ̀Itisalsomanifestlyunclearwhatitisthatthismodeis 4 Ѐsupposedtobecompatiblewith.Itwouldseemjustas X ЀmeaningfultonamethemodeRALPH.   J3Response:J3believesthefeatureshouldstayasitisdefined. T   References   @ 02-007r3,Fortran2000CommitteeDraft    [EndofJ3/02-310r1]