C++Talk.NET C++ language newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  The Gist of Object Oriented Programming Goto page 1, 2, 3  Next          C++Talk.NET Forum Index -> C++ Language (Moderated) View previous topic :: View next topic   Author Message Roger Smythe Guest Posted: Mon Apr 19, 2004 5:48 am    Post subject: The Gist of Object Oriented Programming A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. This was an attempt to explain the gist of OOP to programmers accustomed to the structured programming paradigm. I tried to explain OOP in terms of ideals that can be striven for, even though these ideals may never be perfectly achieved in real systems. Does anyone have any ideas on how to improve this explanation? The goal is to provide the gist of the benefits of OOP to structured programmers in no more than a single short paragraph. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Ian McCulloch Guest Posted: Mon Apr 19, 2004 9:51 pm    Post subject: Re: The Gist of Object Oriented Programming Roger Smythe wrote: Quote: A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. This was an attempt to explain the gist of OOP to programmers accustomed to the structured programming paradigm. I tried to explain OOP in terms of ideals that can be striven for, even though these ideals may never be perfectly achieved in real systems. Does anyone have any ideas on how to improve this explanation? The goal is to provide the gist of the benefits of OOP to structured programmers in no more than a single short paragraph. Does that imply a 'structured programming' style cannot have a well-defined interface? Surely not, you could also be describing top-down design which doesn't necessarily imply object orientation. Not necessarily an improvement, but I have a different perspective on this: Structured programming: 'code' and 'data' are completely separate entities, the code is completely static in the sense that any change in program flow is explicitly coded by the programmer as some test on the data that results in two or more different paths through the code. OO programming: introduces the notion of an 'object', which is a way of packaging code and data together. Objects of a different type can be substituted transparantly, as long as they satisfy the inferface (Liskov substitution principle), in which case the code path may depend implicitly on the actual type (which is really 'data'). Functional programming: code and data are essentially equivalent. Pure OO (as taught in my undergraduate classes) is a mapping between objects and real-life (or imagined) entities (Employee, ATM machine, Animal etc etc). I've never found this approach useful in practice, perhaps because I don't know how to apply it to my usual problem domain (computational physics), maybe it is useful elsewhere. Cheers, Ian McCulloch [ 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: Tue Apr 20, 2004 10:30 am    Post subject: Re: The Gist of Object Oriented Programming "Roger Smythe" <Roger314159 (AT) hotmail (DOT) com> writes: Quote: A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. This definition seems equally true for structured programming. In fact, it seems like a general definition for 'design method'. Quote: This was an attempt to explain the gist of OOP to programmers accustomed to the structured programming paradigm. I tried to explain OOP in terms of ideals that can be striven for, even though these ideals may never be perfectly achieved in real systems. Does anyone have any ideas on how to improve this explanation? The goal is to provide the gist of the benefits of OOP to structured programmers in no more than a single short paragraph. [snip] The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Tue Apr 20, 2004 7:26 pm    Post subject: Re: The Gist of Object Oriented Programming In message <s8ygc.48267$K_.1098501 (AT) bgtnsc05-news (DOT) ops.worldnet.att.net>, Roger Smythe <Roger314159 (AT) hotmail (DOT) com> writes Quote: A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. This was an attempt to explain the gist of OOP to programmers accustomed to the structured programming paradigm. I tried to explain OOP in terms of ideals that can be striven for, even though these ideals may never be perfectly achieved in real systems. Does anyone have any ideas on how to improve this explanation? The goal is to provide the gist of the benefits of OOP to structured programmers in no more than a single short paragraph. Well to start with I would never have recognised OO from your first paragraph. Try Object orientation is a view of programming where the available behaviour is strictly constrained by the type of the data. For example, if you have two values that represent dates, subtracting one from the other will result in a number of days, addition, multiplication and division are not only meaningless but attempts to apply them to such values are diagnosed errors. Of course I am definitely not happy with my paragraph because trying to encapsulate such a major paradigm in a single paragraph requires skills that I do not have. -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit For project ideas and contributions: http://www.spellen.org/youcandoit/projects [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Peter Olcott Guest Posted: Tue Apr 20, 2004 7:42 pm    Post subject: Re: The Gist of Object Oriented Programming "Ian McCulloch" <nobody (AT) example (DOT) com> wrote Quote: Roger Smythe wrote: A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. This was an attempt to explain the gist of OOP to programmers accustomed to the structured programming paradigm. I tried to explain OOP in terms of ideals that can be striven for, even though these ideals may never be perfectly achieved in real systems. Does anyone have any ideas on how to improve this explanation? The goal is to provide the gist of the benefits of OOP to structured programmers in no more than a single short paragraph. Does that imply a 'structured programming' style cannot have a well-defined interface? Surely not, you could also be describing top-down design which doesn't necessarily imply object orientation. You have taken what I have said out of context, and thus provided a different meaning than the one that I provided. The required context is "such that these component parts are completely independent of each other except for their well-defined interfaces" In other words 100% complete encapsulation. Quote: Not necessarily an improvement, but I have a different perspective on this: Structured programming: 'code' and 'data' are completely separate entities, Yes this seems to be one of the primary distinctions. Quote: the code is completely static in the sense that any change in program flow is explicitly coded by the programmer as some test on the data that results in two or more different paths through the code. OO programming: introduces the notion of an 'object', which is a way of packaging code and data together. Objects of a different type can be substituted transparantly, as long as they satisfy the inferface (Liskov substitution principle), in which case the code path may depend implicitly on the actual type (which is really 'data'). Yes this is the crucial difference. Now here is the hard part. Try to accurately describe this difference, not, in how the difference works, but, in terms the the quantity of benefits of this difference. In other words why is this better? and how much better is this than structured programming? Quote: Functional programming: code and data are essentially equivalent. Pure OO (as taught in my undergraduate classes) is a mapping between objects and real-life (or imagined) entities (Employee, ATM machine, Animal etc etc). I've never found this approach useful in practice, perhaps because I don't know how to apply it to my usual problem domain (computational physics), maybe it is useful elsewhere. Cheers, Ian McCulloch [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Peter Olcott Guest Posted: Tue Apr 20, 2004 10:33 pm    Post subject: Re: The Gist of Object Oriented Programming Quote: A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. This definition seems equally true for structured programming. In fact, it seems like a general definition for 'design method'. Yet the structured paradigm does not have nearly the same degree of encapsulation... "such that these component parts are completely independent of each other except for their well-defined interfaces." Quote: The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. This is the kind of material that I was looking for. Because of the capability to create user defined types, OOP can more closely correspond to the real world. This makes systems much more self-documenting, but, more importantly, provides these systems with a fully elaborated design, this being the design of the real world system that is being modeled. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top John G Harris Guest Posted: Wed Apr 21, 2004 12:14 am    Post subject: Re: The Gist of Object Oriented Programming In message <8665bvvrir.fsf (AT) Zorthluthik (DOT) local.bar>, llewelly <llewelly.at (AT) xmission (DOT) dot.com> writes <snip> Quote: The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. The idea that there is a mapping from things inside the computer to things outside can be tricky to describe, and is sometimes misleading. For instance, where in the real world do you find Sonic The Hedgehog and Lara Croft? What exactly is a digital camera image, and where is it? Why is OO so useful when doing BigInt arithmetic? John -- John Harris [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Antoun Kanawati Guest Posted: Wed Apr 21, 2004 12:29 am    Post subject: Re: The Gist of Object Oriented Programming llewelly wrote: Quote: "Roger Smythe" <Roger314159 (AT) hotmail (DOT) com> writes: A means for the progressive decomposition a problem space into increasingly simpler component parts such that these component parts represent higher levels of conceptual abstraction, and are completely independent of each other except for their well-defined interfaces. [snip] The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. I've always had an issue with Objects and The Real World. Whether you're doing procedural or object-oriented programming, the "Real World" is out, and all you have is some abstraction, which [hopefully] captures the essence of the problem under consideration. The primary characteristic of object oriented programming is that it provides a notation for the classification of program entities (classes), the specification of their interfaces (method names, exposed variables), and the control of generalizations and specializations. Different OO-languages provide different supports for various aspects of specification. In Smalltalk, you have no control over the interface, anyone can, post-facto, add methods to a class. In C++, you have very fine control over specialization points (virtual, non-virtual, pure-virtual, and inheritance). Java's 'final' also allows some fine tuning. All of the above is doable, and has been done, with procedural frameworks (e.g.: the X[-windows] class system). Object oriented languages formalize these approaches, solidify them and make them more robust and usable. In other words, OOP simplifies some aspects of expressing abstract models in program designs and implementations, because it provides notation that is better suited for this sort of work. The conceptual similarities to "The Real World" cannot be avoided no matter how you're programming. If they're not there, you're doing the wrong thing. The expression of the concepts, which affects the visibility of the similarities, is where object oriented notations have an impact. On a related note: lately, "Aspect Orientation" is picking up steam because Object Orientation does not seem to capture the real world just right. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Walt Karas Guest Posted: Wed Apr 21, 2004 9:15 am    Post subject: Re: The Gist of Object Oriented Programming I think the first thing you'd have to make clear is that there is not a clear consensus as to what OO means. We can only give our own point of view. OO (I would say) originated in the idea of abstract data types. Traditionally, a "type" had been thought of as a way of representing a set of values in computer memory (e.g. an "int" is represented as 32 binary bits using 2's complement for negative values). But an abstract data type was defined by a set of operations that could be performed on a values of the type that would result in other valid values of the type. Algorigthms written to use abstract data types were easier to understand, less error prone in that invalid values were avoided, and more portable to different architectures. In orthodox OO-speak, the abstraction of data types is called "encapsulation". Once you look at types as sets of operations, a definition like "type B is type A plus operations X, Y and Z" seems straight-forward. This is inheritance. It saves alot of repetition in the definition of B. It also makes it easy to define programming languages where an algorithm written for values of type A also works for type B. The next step forward is a programming language where you can express things like "type B is type A, execpt A's operations U and V work differntly for B, and B has the additional operations X, Y and Z". And as long as the changed operations U and V are "conceptually" the same, algorithms written for type A still work for type B. This is called (for extreme want of a better term) polymorphism. It's now a small step to see the usefulness of data types that are just sets of operations, with no associated set of values. The only purpose of these data types is to be inherited from by other data types. Such a data type can be useful even if some or all of its operations have only "conceptual" definitions, and require the inheriting type to provide a compilable definition for the operation. This idea expresses itself as abstract classes in C++, and interfaces in Java. Another swell thing is if you can write an algorithm for some type T that uses operations X, Y and Z and relies only on a "conceptual" definition of X, Y and Z. Then, the algorithm works for any type having opearations X, Y, and Z with the same conceptual definition. The "type" (the set of operations X, Y and Z) become anonymous and implicit. This is what C++ templates allow you to do. Most people seem to feel that this goes beyond OO, and we need to use the term "generic programming" for this capability. It seems to me that SmallTalk also had this capability because of its weak typing, and it was considered to be covered by the term OO. But I've seen it argued that SmallTalk programmers where never interested in this concept (of writing algoritms that operate on ad-hoc sets of operations, rather than base classes), so it shouldn't be thought of as falling under OO. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Peter Olcott Guest Posted: Wed Apr 21, 2004 9:23 am    Post subject: Re: The Gist of Object Oriented Programming Quote: Well to start with I would never have recognised OO from your first paragraph. Try Object orientation is a view of programming where the available behaviour is strictly constrained by the type of the data. For example, if you have two values that represent dates, subtracting one from the other will result in a number of days, addition, multiplication and division are not only meaningless but attempts to apply them to such values are diagnosed errors. Of course I am definitely not happy with my paragraph because trying to encapsulate such a major paradigm in a single paragraph requires skills that I do not have. Nonetheless that was a pretty good attempt. When Bjarne Stroustrup was here and presented a brief description of his language, it was very similar to what you have said. UDT, might be the essence of OOP. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Francis Glassborow Guest Posted: Wed Apr 21, 2004 2:09 pm    Post subject: Re: The Gist of Object Oriented Programming In message <XG8hc.7899$um3.191094 (AT) bgtnsc04-news (DOT) ops.worldnet.att.net>, Peter Olcott <olcott (AT) worldnet (DOT) att.net> writes Quote: The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. This is the kind of material that I was looking for. Because of the capability to create user defined types, OOP can more closely correspond to the real world. This makes systems much more self-documenting, but, more importantly, provides these systems with a fully elaborated design, this being the design of the real world system that is being modeled. However I know a number of OO experts who will dance up and down with irritation at this view of OO because it leads to a simplistic approach. It certainly helps the novice get started but there is more to it and a fixation on that view will inhibit the student from going further. I can remember an enlightening conversation with the designer of some software for managing an oil refinery. The gist of it was that the 'objects' were not the physical equipment but the processes. The trouble is that it takes a great deal of time and open-minded experience to develop wisdom. Sound-bytes are often greatly inhibit the acquisition of wisdom. -- Francis Glassborow ACCU Author of 'You Can Do It!' see http://www.spellen.org/youcandoit For project ideas and contributions: http://www.spellen.org/youcandoit/projects [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Hyman Rosen Guest Posted: Thu Apr 22, 2004 12:17 am    Post subject: Re: The Gist of Object Oriented Programming Francis Glassborow wrote: Quote: However I know a number of OO experts who will dance up and down with irritation at this view of OO because it leads to a simplistic approach. In my opinion, OO is a programming technique, and the objects are programming constructs which help you write the program to do what it needs to do. The objects may model some aspect of the real world or they may simply be useful internal constructs or both, whatever gets the job done. Arguing about the "meaning" of the objects in a program is one of the most useless endeavors I can think of. [ 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: Thu Apr 22, 2004 3:02 pm    Post subject: Re: The Gist of Object Oriented Programming John G Harris <news0 (AT) nospam (DOT) demon.co.uk> writes: Quote: In message <8665bvvrir.fsf (AT) Zorthluthik (DOT) local.bar>, llewelly [email]llewelly.at (AT) xmission (DOT) dot.com[/email]> writes snip The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. The idea that there is a mapping from things inside the computer to things outside There is not 'a mapping'. There are endless possible mappings. It is up to the designer to devise one that aids understanding. If such a mapping is difficult to determine, OO may not be appropriate. Quote: can be tricky to describe, and is sometimes misleading. If OO design ideas are tricky or misleading in a particular problem domain USE SOMETHING ELSE. Quote: For instance, where in the real world do you find Sonic The Hedgehog and Lara Croft? What exactly is a digital camera image, and where is it? Why is OO so useful when doing BigInt arithmetic? I don't think OO is at all useful for BigInt arithmetic. BigInt arithmetic naturally operates on values, not on entities. In fact, I think OO is a hindrance when it comes to most kinds of numerical types. (Especially in a langauge like C++, which has rather primitive OO.) [ 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: Thu Apr 22, 2004 3:05 pm    Post subject: Re: The Gist of Object Oriented Programming Francis Glassborow <francis (AT) robinton (DOT) demon.co.uk> writes: Quote: In message <XG8hc.7899$um3.191094 (AT) bgtnsc04-news (DOT) ops.worldnet.att.net>, Peter Olcott <olcott (AT) worldnet (DOT) att.net> writes The benefit of OO is OO-designed objects have behavior conceptually similar to real-world entities, and OO-designed types have behavior conceptually similar to real-world categories. This is the kind of material that I was looking for. Because of the capability to create user defined types, OOP can more closely correspond to the real world. This makes systems much more self-documenting, but, more importantly, provides these systems with a fully elaborated design, this being the design of the real world system that is being modeled. However I know a number of OO experts who will dance up and down with irritation at this view of OO because it leads to a simplistic approach. It certainly helps the novice get started but there is more to it and a fixation on that view will inhibit the student from going further. Agreed, but I couldn't think up a short paragraph which didn't have that drawback. I'll agree that there is danger in simplification, but sometimes I think one difference between the beginner and a master, is that the master has identified and unleared many of the simplifications that were helpful or necessary to him as a beginner. I don't believe it's possible to avoid all simplifications. Quote: I can remember an enlightening conversation with the designer of some software for managing an oil refinery. The gist of it was that the 'objects' were not the physical equipment but the processes. I think at some point the student must learn that OO can be used to map entities which are not real-world things to program objects, but I don't know fit that into a short paragraph for a beginner. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] Back to top Dave Harris Guest Posted: Thu Apr 22, 2004 3:06 pm    Post subject: Re: The Gist of Object Oriented Programming [email]Roger314159 (AT) hotmail (DOT) com[/email] (Roger Smythe) wrote (abridged): Quote: This was an attempt to explain the gist of OOP to programmers accustomed to the structured programming paradigm. I think we need to see how you defined "Structured programming" I generally take that to mean the Single Entrance, Single Exit approach to coding individual routines. As such it was at a lower level than OO, which is a way of organising routines into modules. Quote: I tried to explain OOP in terms of ideals that can be striven for, even though these ideals may never be perfectly achieved in real systems. For me it a specific strategy for achieving those ideals, whether or not that strategy is successful. Other strategies may do better without having to be called OO. "Object oriented" is not a synonym for "good". Specifically: OO is a decompositional paradigm which takes access to common representations as its relatedness metric. The goal is to manage dependencies. We want to avoid situations where changes to one module causes changes to another module, because that can lead to knock-on effects propagating through the code without limit. The OO insight is that changes to a representation are likely to cause changes to the code which manipulates it. Therefore we organise modules around the representations, to form Abstract Data Types. I prefer the word "representation" to "data", even though in C++ we are usually talking about data, because I believe the insights of OO can carry across to situations in which there is no mutable state. So OO is programming with ADTs. Some people call that "Object based", saying that it is not OO without inheritance and dynamic polymorphism. I disagree. I think there is a kind of mental flip which happens when you go from structured programming to OO, and I think it happens with ADTs. The other stuff is nice, of course. A good OO language should support, in order of importance: (1) User-defined abstract data types. (2) Message-sending or some other form of dynamic polymorphism. (3) Inheritance or some other form of explicit reuse. (4) Automated memory management. (The last one is controversial in this newsgroup - drop it if it causes offense.) -- Dave Harris, Nottingham, UK [ 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 Goto page 1, 2, 3  Next Page 1 of 3   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.