On Thu, 31 Dec 2009 12:35:15 -0500, P S wrote: >On Thu, Dec 31, 2009 at 12:08 PM, McKown, John wrote: > >> Hum, I have exactly the opposite opinion. I dislike Windows' "case >> preserving" but "case ignoring" feature. I think that Mac OSX is like >> Windows in this as well (and it is UNIX based). If it is going to ignore >> Since perhaps 10.2; certainly 10.3, OS X provides the option of formatting a filesystem as case-sensitive or case-insensitive. It is rumored that many vendor apps fail on a case-sensitive FS. I've encountered only one, and it was fixed in a later release.
>> case, then it should automagically change all file names to either all upper >> or all lower case in the filesystem. And, if really desired, somebody could >> write a UNIX filesystem which does this same "case ignoring" (by up casing >> or lower casing the input) when doing a file lookup and "automatic >> upper/lower casing" when storing the file in the filesystem. Of course, >> people do like to "CamelCase" file names to make them more readable. And I >> agree that I can't see why anybody would want the files: foo , Foo , fOo , >> FOo, and so on in a directory. >> Identifiers, at least. In typography, the HTML (SGML?) entities Á and á have different meanings. It was one obvious way of distinguishing them; opinions will vary as to which is best. Then if the bitmaps or outline instructions are stored in a filesystem, Aacute.bmp and aacute.bmp are obvious choices. >OK, *not* trying to start a war, honestly curious: why do you dislike it? >(The 8.3 filenames are, of course, all uppercase, btw.) Why should it upper- >or lowercase the filenames? What would that buy? > Ease in sorting and searching. If I were implementing such a filesystem, I'd store the names in a single case followed by a bitmap indicating which case to display. >Mind you, an *option* to say "DISPLAY FILENAMES IN ALL upper/lower CASE" >would be a plausible option. CamelCase is useful for readability -- studies >have proven that (he says, without citation but with confidence). > Long ago, I experimented with STOW from assembler. It uncomplainingly creates member names in mixed case, and worse. At that time, I could operate on such members by selecting them from a member menu, but not by typing the names on the command line, whether quoted or not. My strong bias is that such rules should be enforced by the filesystem, not haphazardly by the APIs in various languages: typing 'FooBar' in Assembler, JCL, TSO, C, etc. should all produce the uniform results because the string should be passed ASIS to STOW, BLDL, DESERV, etc. which should make the decision on how to handle mixed case: o full case-sensitivity or o Lower case is invalid or o case insignificant to the FS, but preserved for display. At the lowest level, the z/OS filesystem is fully case-sensitive. Assembler, at least, respects this (How could it be otherwise? Well, the BLDL and STOW macros could convert the case before issuing the SVC.) Other APIs should respect the programmers' conventions even as assembler does. And I've experimented with Samba and filenames differing in only case of the characters. It seems sensitive to version of client. I recall both seeing one of the names converted to 8.3, but later seeing both names displayed as in the UNIX directory, but when I click on one it's unpredictable which one will open. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html