Where should the type information be 146
On Tue, 29 Mar 2005 18:03:27 -0700 in alt.folklore.computers, Steve
Precedence tables are merely a summary of the grammatical rules which specify the operator precedence: no such tables exist in the Standard.
IMHO boolean variables in programs should be considered anathema comparable to GOTOs. I have no problems with boolean operations or results, just the ability to store anything which *implicitly* encodes a previous state, and which is then used subsequently to conditionally perform actions. The problem seems to be that programmers get confused between boolean variables' use as a record of previous state, and its use to have actions performed. I've seen worse spaghetti programs result from boolean variables being used interchangeably between records of previous states and triggers for subsequent actions, than I've ever seen using GOTOs (except one using ForTran buttIGNed GOTOs excessively).
One could say the same for any expression that is not fully parenthesized in any language: the implication then is that *no* operator precedence should be specified in *any* language; but conventions about precedence are useful, as they avoid specifying the obvious; where operations are not obvious, comments are available for specifying intentions clearly.
Any language which you can't use in such a manner is probably less useful than C: I'm sure experts could come up with similarly obfuscated code in PL-I.
This is a feature with C, not a bug. In C, just as with many buttemblers, and, or, xor and not are used in arithmetic roles in addition to roles in...
-- Thanks. Take care, Brian Inglis Calgary, Alberta, Canada
fake address use address above to reply
Where should the type information be 147
Alt Folklore Computers from Newsgroups