C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  Coding standard rule for using struct's..?          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message John Thilenius Guest Posted: Thu Aug 21, 2003 10:46 am    Post subject: Coding standard rule for using struct's..? If you were tasked with creating a guideline which defines what a struct can/cannot contain, or should/shouldn't do, what would it look like? One reason for the guideline might be to prevent struct's from becoming large nested classes, which can happen over time. Some example guidelines might look like the following: Extreme... 1) A struct should not contain any member functions, or objects containing member functions. Restrictive... 2) The only methods a struct should contain are those which support member initialization, assignment & copying of the struct. Open... 3) Use your own judgment. Typically, use a struct for ... Assuming you have lots of newbie developers, what would your guideline look like? Thanks in advance, John Thilenius Boeing Software Engineering [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Jerry Feldman Guest Posted: Fri Aug 22, 2003 9:38 am    Post subject: Re: Coding standard rule for using struct's..? On 21 Aug 2003 06:46:37 -0400 John Thilenius <john.m.thilenius (AT) boeing (DOT) com> wrote: Quote: If you were tasked with creating a guideline which defines what a struct can/cannot contain, or should/shouldn't do, what would it look like? One reason for the guideline might be to prevent struct's from becoming large nested classes, which can happen over time. Some example guidelines might look like the following: Extreme... 1) A struct should not contain any member functions, or objects containing member functions. Restrictive... 2) The only methods a struct should contain are those which support member initialization, assignment & copying of the struct. Open... 3) Use your own judgment. Typically, use a struct for ... I would probably go for 2 since they already have default constructors and destructors. And, I would add some verbage to suggest places where a struct might be used more appropriately than a class. Or, an even more extreme approach is to simply not allow structs. -- Jerry Feldman <gaf-nospam-at-blu.org> Boston Linux and Unix user group http://www.blu.org PGP key id:C5061EA9 PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9 [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Torsten Robitzki Guest Posted: Fri Aug 22, 2003 9:49 am    Post subject: Re: Coding standard rule for using struct's..? Hi, John Thilenius wrote: Quote: If you were tasked with creating a guideline which defines what a struct can/cannot contain, or should/shouldn't do, what would it look like? One reason for the guideline might be to prevent struct's from becoming large nested classes, which can happen over time. Some example guidelines might look like the following: Extreme... 1) A struct should not contain any member functions, or objects containing member functions. Restrictive... 2) The only methods a struct should contain are those which support member initialization, assignment & copying of the struct. Open... 3) Use your own judgment. Typically, use a struct for ... Assuming you have lots of newbie developers, what would your guideline look like? 4) Learn the difference between class and struct and judge on knowledge and rational reasons rather then on guidelines. I think it's a common misbelieve that one can turn a newbie into a competent software developer with a guideline. regards Torsten [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top llewelly Guest Posted: Fri Aug 22, 2003 4:18 pm    Post subject: Re: Coding standard rule for using struct's..? John Thilenius <john.m.thilenius (AT) boeing (DOT) com> writes: Quote: If you were tasked with creating a guideline which defines what a struct can/cannot contain, or should/shouldn't do, what would it look like? One reason for the guideline might be to prevent struct's from becoming large nested classes, which can happen over time. Some example guidelines might look like the following: Extreme... 1) A struct should not contain any member functions, or objects containing member functions. Restrictive... 2) The only methods a struct should contain are those which support member initialization, assignment & copying of the struct. Open... 3) Use your own judgment. Typically, use a struct for ... Assuming you have lots of newbie developers, what would your guideline look like? [snip] 4) All programmers will be taught what POD means, what operations are safe for POD, and not safe for non-POD. All programmers will be taught struct != POD. Code review will analyze each struct for POD-ness and that it is correctly documented it as POD or non-POD. There will be a programming education session on POD, Friday at 10am. In my expierence, there are two views of struct, which together dominate nearly all C++ programming: (a) struct is 'just like in C', and (b) struct is 'just like class but with different default access' . ( (b) is of course the correct view, but bear with me. ) People of the (a) view will memcpy structs about with impunity. They will assume that if struct A is a 'prefix' of struct B, A and B are layout compatible. And so on. In short, they expect struct == POD, even if they've never heard the term 'POD'. My problem with (1), and even more so with (2), is that they allow non-POD structs, and yet seem to be intended to cater to people who think (a) - but if a struct has base classes, access qualifiers, reference members, or non-POD members (and (1) allows non-POD members! ), it can't be safely treated in the ways those of the (a) view expect - but since they don't know about POD-ness, they'll trust the guideline, and write non-POD structs that match the guideline. The result will all too often be hard-to-reproduce bugs that defy your best efforts and most effective tools. If you feel you must have a 'protect the ignorant' guideline for this situation: 5) All structs must be POD, and all structs be designed, implemented, and reviewed by people who understand POD-ness. Note that (3) essentially requires (4), and even (5) requires (4) for some portion of the development team. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Attila Feher Guest Posted: Sat Aug 23, 2003 3:28 am    Post subject: Re: Coding standard rule for using struct's..? Francis Glassborow wrote: Quote: In article <3F438E1E.1449DE87 (AT) boeing (DOT) com>, John Thilenius [email]john.m.thilenius (AT) boeing (DOT) com[/email]> writes Assuming you have lots of newbie developers, what would your guideline look like? [SNIP] Now making a nested private type a struct seems reasonable. It would also seem reasonable to use a struct for any type whose state you elect to make public and note that as soon as such a type has been published you are locked to that decision for that type. I would add to this something which is not my tought (as usual ), but seems to be appropriate. Traits (and other "collection of data static at compile time) seems to be also a place where the use of a struct is appropriate. Traits (once published) are such that they only get extended and they basically have only "public interface", a loads of static members. Kind of a little namespace. :-) -- Attila aka WW [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top John Potter Guest Posted: Sat Aug 23, 2003 3:29 am    Post subject: Re: Coding standard rule for using struct's..? On 22 Aug 2003 05:49:08 -0400, Torsten Robitzki <Torsten (AT) Robitzki (DOT) de> wrote: Quote: 4) Learn the difference between class and struct and judge on knowledge and rational reasons rather then on guidelines. The difference is spelling and nothing else. class C : B { == struct C : private B { private : struct S : R { == class S : public R { public : What rational reasons derive from these observations? John [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Andrea Griffini Guest Posted: Sat Aug 23, 2003 2:42 pm    Post subject: Re: Coding standard rule for using struct's..? On 22 Aug 2003 23:29:32 -0400, John Potter <jpotter (AT) falcon (DOT) lhup.edu> wrote: Quote: What rational reasons derive from these observations? That rational reasoning doesn't help much in C++ wonderland. Study helps more. Andrea [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Torsten Robitzki Guest Posted: Sat Aug 23, 2003 5:48 pm    Post subject: Re: Coding standard rule for using struct's..? John Potter wrote: Quote: On 22 Aug 2003 05:49:08 -0400, Torsten Robitzki wrote: 4) Learn the difference between class and struct and judge on knowledge and rational reasons rather then on guidelines. The difference is spelling and nothing else. class C : B { == struct C : private B { private : struct S : R { == class S : public R { public : What rational reasons derive from these observations? Now, as the newbies know the difference they might decide to use a struct when there is the need for a class with a lot of public members. But my original point was that writing a coding standard about the usage of the key words struct and class to prevent newbies from making design mistakes is pointless. It would be better to educate them and thus give them the ability to make a reasonable decision on there own. regards Torsten [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Peter Schneider Guest Posted: Sat Aug 23, 2003 5:57 pm    Post subject: Re: Coding standard rule for using struct's..? Hi, John Thilenius, Boeing Software Engineering, wrote: Quote: If you were tasked with creating a guideline which defines what a struct can/cannot contain, or should/shouldn't do, what would it look like? It's a design question, rather than a coding question, not at all C++ specific. "Though shalt not publish implementation details", needless to remind a Boeing employee -- after all, some papers in my Software Engineering textbooks came from Boeing. The broader the audience of the publishing, the worse. For oneself, in a limited piece of code, it is a convenience for a bad typer. In a library interface, it's normally a major mistake. Quote: Assuming you have lots of newbie developers, what would your guideline look like? Not differently. Give them as much education and as much freedom as possible. They'll experience that their convenient little innocent data structures frequently need to be rewritten, once all side effects and special cases are discovered and all extensions are to be implemented. If you simply forbid structures, they'll use bundles of arrays or something worse instead. It's not about rules, it's about insight here. Regards, Peter [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Philippe Mori Guest Posted: Sat Aug 23, 2003 8:07 pm    Post subject: Re: Coding standard rule for using struct's..? Quote: Open... 3) Use your own judgment. Typically, use a struct for ... For enums. Define a struct that has an enum in it. Other classes inherit from your struct. For enum, I prefer namespace (you can an using in function that uses it a lot). I do not recommend classes that inherit from structure and give access to all structure data members. Better to have the structure as a private data member and provide accessor in the class. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top James Dennett Guest Posted: Sun Aug 24, 2003 11:23 pm    Post subject: Re: Coding standard rule for using struct's..? John Potter wrote: Quote: On 22 Aug 2003 05:49:08 -0400, Torsten Robitzki wrote: 4) Learn the difference between class and struct and judge on knowledge and rational reasons rather then on guidelines. The difference is spelling and nothing else. class C : B { == struct C : private B { private : struct S : R { == class S : public R { public : What rational reasons derive from these observations? The fact that the different spelling has different connotations to human programmers. "struct" makes many think in C-like terms, and "class" pushes them towards thinking more in terms of interfaces. These non-technical differences are important, of course: most code is at least as much about communication between humans as it is about communication between a human and a computer. If I'm going to write access qualifiers at all, I tend to write "class". If everything is to be public, I tend to write "struct". Unless local guidelines/rules override that, in which case I do whatever they dictate (unless it's crazy, and then I'd do what they dicate until/unless I can get them changed). -- James. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top James Dennett Guest Posted: Tue Aug 26, 2003 8:39 am    Post subject: Re: Coding standard rule for using struct's..? John Potter wrote: Quote: On 24 Aug 2003 19:23:31 -0400, James Dennett <jdennett (AT) acm (DOT) org> wrote: John Potter wrote: The difference is spelling and nothing else. class C : B { == struct C : private B { private : struct S : R { == class S : public R { public : What rational reasons derive from these observations? The fact that the different spelling has different connotations to human programmers. "struct" makes many think in C-like terms, and "class" pushes them towards thinking more in terms of interfaces. I believe that, but there are other humans. We're not all clones yet, then. Probably a good thing. Quote: I suspect that this has much to do with the age of the human. The novice is likely not coming from C and has never heard of a struct. Given a guideline of public before protected before private and most derivation should be public, it is obvious that struct has the right defaults and should always be used. Given that struct, unless manually overridden, defaults to exposing everything, it is obvious that class has the right default and should always be used ;) Quote: Is this a call for education on the prejudiced connotations of the ancient like me? Guidelines might be easier and more appropriate. I think it comes down to taste. Any purely logical reasons will pale into insignificance next to human factors, and the chance of getting complete agreement among a large group of people on an issue that is subjective is close to zero. Probably the thing that does matter is to teach the truth, that in C++ class and struct (as used to declare class types) are almost identical, and that any choice for when to use which is a convention rather than being based on a language restriction. -- James. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         C++Talk.NET Forum Index -> C++ Language (Moderated) All times are GMT Page 1 of 1   Jump to: Select a forum C/C++----------------C LanguageC Language (Moderated)C++ Language (Moderated)C++ language (comp.lang.c++)C++ language, library and standards Non-English C/C++ Newsgroups----------------C++ (German)C++ (French)  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.