home *** CD-ROM | disk | FTP | other *** search
- >From allegra!phri!roy@seismo.UUCP Fri May 2 22:44:44 1986
- Date: Mon, 28 Apr 86 11:55:42 edt
- From: allegra!phri!roy@seismo.UUCP (Roy Smith)
-
- In section C.3 of the April 3rd draft, it is proposed that ioctl()
- be replaced with iocntl(). While I aggree with the basic arguments, I
- think iocntl() is a bad name for the new call.
-
- Somebody once said that your code should pass the telephone test;
- if you can't read it to somebody over the phone and have them understand
- it, it's too complex. It has also been said (again I can't remember by
- who) [Oh, Kernighan and Plauger, among many others, if you're just looking
- for an authority to cite -- mod.] that it is bad style to use variable
- names that differ only slightly; using both maxij and maxji for distinct
- but similar things is bound to cause confusion when somebody has to read
- it later. [Or port code that abuses flexnames, though not a problem here.
- -- mod.]
-
- Surely "ioctl" and "iocntl" fail both these tests of programming
- clarity. I'd like to see the name "iocntl" changed without changing the
- proposed functionality. Perhaps "iofunc" or even "iox". More descriptive
- would be "special_io", but this is bad if you want to appease 7 character
- name compilers.
-
- [Any comments from the people to whose hearts "fcntl" may have once been
- near and dear? -- mod]
-
- Roy Smith, {allegra,philabs}!phri!roy
- System Administrator, Public Health Research Institute
- 455 First Avenue, New York, NY 10016
-
- Volume-Number: Volume 6, Number 15
-
-