home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.mac.programmer
- Path: sparky!uunet!think.com!sdd.hp.com!ux1.cso.uiuc.edu!news.cso.uiuc.edu!alexia!cole
- From: cole@alexia.lis.uiuc.edu (Sandra Stewart-Cole)
- Subject: Re: Why Control Segmentation?
- References: <1993Jan09.090047.8491@cosoco.com>
- Message-ID: <C0Mo2n.BDE@news.cso.uiuc.edu>
- Sender: usenet@news.cso.uiuc.edu (Net Noise owner)
- Organization: University of Illinois at Urbana
- Date: Sun, 10 Jan 1993 07:42:22 GMT
- Lines: 14
-
- The more control the better, but there are powers one should never need. There
- is a clear advantage to having well thought out and implemented segment
- planning. Aside from the memory concerns of needing to keep multiple segments
- around, there is always the simple performance penalties of far calls and the
- jump table. I think the level of control is really a matter of working style.
- With the by-file control of Think C, I get fine results, but I also have a lot
- of different files. Provided the programmer designs the code well, there ought
- to be no excuse for having an object splittable between segments (i.e. at the
- function level) If I could split easily by object instead of file, I would
- have an easier management problem in Think C. I doubt I would ever use an easy
- ability to split my objects by function, but if I did a lot more procedural
- work, it might be nice (i.e. maybe encourage good object design) to NOT allow
- segmentation by function inside objects, but allow it for procedural code
- (maybe by thinking of procedures as tiny objects...)
-