home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!corton!babbage!imag!imag.fr!dumas
- From: dumas@imag.fr (Olivier Dumas)
- Newsgroups: comp.lang.pascal
- Subject: Re: TP units question
- Message-ID: <39704@imag.imag.fr>
- Date: 11 Sep 92 07:28:12 GMT
- References: <1992Sep4.141203.28370@news.columbia.edu> <dmurdoch.123.715622406@mast.queensu.ca> <1992Sep9.152408.842@iplmail.orl.mmc.com>
- Sender: usenet@imag.fr
- Reply-To: dumas@bounty.imag.fr
- Organization: BullImag
- Lines: 25
- Nntp-Posting-Host: samoa
-
- In article <1992Sep9.152408.842@iplmail.orl.mmc.com>, steveg@iplmail.orl.mmc.com (Steve Gabrilowitz) writes:
-
- > Another disadvantage of doing this is that if you make any changes to the
- > parameter list in the routine, you have to change it in two places (interface and
- > implementation) instead of just one. Also, when you do this then TPC /M will
- I don't agree with that. If you put the name of a procedure in the interface,
- you need to put the parameter list, but when you implement it, you don't have
- to repeat the parameter list.
- Of course, repeating it makes the life easier when you read the implementation.
-
- btw, it is wiser not to interface a locally-used procedure, for maintenance reasons (as said by Duncan Murdoch in an earlier article)
-
- > have to recompile all the units that use the one with the routine you just
- > modified, where if all the changes had been made in the implementation section it
- > would not be necessary.
- >
-
- --
- +------------------------------------------------------------+
- | Olivier DUMAS -- e-mail : dumas@bounty.imag.fr |
- | Unite Mixte BULL-Imag -- ZI de Mayencin -- Gieres - France |
- +------------------------------------------------------------+
- Mathematics may be defined as the subject in which we never know what we
- are talking about, nor whether what we are saying is true.
- -- Bertrand Russell (1872-1967)
-