AppletTalk.com Java discussions newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  Problem with lack of pass-by-result          AppletTalk.com Forum Index -> Java Help View previous topic :: View next topic   Author Message Richard Guest Posted: Wed Sep 15, 2004 9:50 am    Post subject: Problem with lack of pass-by-result Java newbie. As a Java newbie, I'm charged with learning the language by maintaining a pretty good-sized pile of code. The conceptual problem I'm having is with the C convention of using pass-by-result (and pass-by-value-result) function parameters to readily show which functions use and change variables outside of the function scope. In Java, pass-by-result vars are supposedly verbotem (although check BufferInputStream.read(byte[] b, int off, int len) for at least one exception to that rule that I've seen so far). So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? -- What matters most? Back to top Stefan Schulz Guest Posted: Wed Sep 15, 2004 10:51 am    Post subject: Re: Problem with lack of pass-by-result Richard wrote: Quote: Java newbie. As a Java newbie, I'm charged with learning the language by maintaining a pretty good-sized pile of code. Good luck ;) Quote: The conceptual problem I'm having is with the C convention of using pass-by-result (and pass-by-value-result) function parameters to readily show which functions use and change variables outside of the function scope. In Java, you usally do not care outside of the classes implementation. Each of them should be self-contained, not showing any of the internal state to the "outside world". Quote: In Java, pass-by-result vars are supposedly verbotem (although check BufferInputStream.read(byte[] b, int off, int len) for at least one exception to that rule that I've seen so far). Not really, you pass an Object (an array of bytes) which is passed by value, just like everything else in java. However, as with all object references, you can actually _change the state_ of that Object. That you do this in a much more direct fashion then with most other Objects does not change that. (Think of passing objects as passing a pointer to a struct, if you come from the C corner). Quote: So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? The convention is that only the Class internals should care. Noone outside the class should even know about the private members of the class which may be affected by a method call. As for documentation within the class, there are no hard-and-fast rules for this, AFAIK. Just remember that you should not depend on internal state. (For example, you should not care if a method getResource() just returns a reference to one of the class' fields, or computes an entirely new Resource. This may even change!) Back to top Matt Humphrey Guest Posted: Wed Sep 15, 2004 11:33 am    Post subject: Re: Problem with lack of pass-by-result "Richard" <rh310 (AT) hotmail (DOT) com> wrote Quote: Java newbie. As a Java newbie, I'm charged with learning the language by maintaining a pretty good-sized pile of code. The conceptual problem I'm having is with the C convention of using pass-by-result (and pass-by-value-result) function parameters to readily show which functions use and change variables outside of the function scope. In Java, pass-by-result vars are supposedly verbotem (although check BufferInputStream.read(byte[] b, int off, int len) for at least one exception to that rule that I've seen so far). Java very consistently (with no exceptions) follows the rule of pass-by-value where the value of an object is its reference. All objects including arrays of primitives (byte [], int []) are represented by a reference to their associated data. These references are passed by value as input-only parameters. The references may not be changed, but the contents to which they refer may be. This concept is fundamental to understanding Java. Any object that is passed may have its contents modified, but the reference to the object itself cannot be changed as a parameter. In the read () method above, the (reference to) the array of bytes is passed in and the method fills out the contents of the the array. Note that Java primitives (int, short, byte, long, char) are not objects. There are no references to primitives. http://mindprod.com/jgloss/callbyvalue.html http://mindprod.com/jgloss/callbyreference.html Quote: So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? Method documentation indicates whether the objects refered to by parameters are intended to be modified. Cheers, Matt Humphrey [email]matth (AT) ivizNOSPAM (DOT) com[/email] http://www.iviz.com/ Back to top Richard Guest Posted: Wed Sep 15, 2004 8:29 pm    Post subject: Re: Problem with lack of pass-by-result [email]terra (AT) spacetime (DOT) de[/email] wrote... Quote: Richard wrote: Java newbie. As a Java newbie, I'm charged with learning the language by maintaining a pretty good-sized pile of code. Good luck It's been an "interesting" three months, that's for sure. Quote: The conceptual problem I'm having is with the C convention of using pass-by-result (and pass-by-value-result) function parameters to readily show which functions use and change variables outside of the function scope. In Java, you usally do not care outside of the classes implementation. Each of them should be self-contained, not showing any of the internal state to the "outside world". Understood. Some of the classes I have to maintain are very large, and their class variables are modified by several different methods. It's not an impossible problem, I just have a lot of documentation on what's changing what to create and maintain. Feels primitive. :) Quote: In Java, pass-by-result vars are supposedly verbotem (although check BufferInputStream.read(byte[] b, int off, int len) for at least one exception to that rule that I've seen so far). Not really, you pass an Object (an array of bytes) which is passed by value, just like everything else in java. However, as with all object references, you can actually _change the state_ of that Object. That you do this in a much more direct fashion then with most other Objects does not change that. (Think of passing objects as passing a pointer to a struct, if you come from the C corner). Now this is very interesting. I did not know this. Time to go write some sandbox code. Quote: So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? The convention is that only the Class internals should care. Noone outside the class should even know about the private members of the class which may be affected by a method call. As for documentation within the class, there are no hard-and-fast rules for this, AFAIK. Just remember that you should not depend on internal state. (For example, you should not care if a method getResource() just returns a reference to one of the class' fields, or computes an entirely new Resource. This may even change!) -- What matters most? Back to top Thomas G. Marshall Guest Posted: Fri Sep 17, 2004 1:36 am    Post subject: Re: Problem with lack of pass-by-result Stefan Schulz coughed up: Quote: Richard wrote: Java newbie. As a Java newbie, I'm charged with learning the language by maintaining a pretty good-sized pile of code. Good luck ;) The conceptual problem I'm having is with the C convention of using pass-by-result (and pass-by-value-result) function parameters to readily show which functions use and change variables outside of the function scope. In Java, you usally do not care outside of the classes implementation. Each of them should be self-contained, not showing any of the internal state to the "outside world". In Java, pass-by-result vars are supposedly verbotem (although check BufferInputStream.read(byte[] b, int off, int len) for at least one exception to that rule that I've seen so far). Not really, you pass an Object (an array of bytes) which is passed by value, just like everything else in java. However, as with all object references, you can actually _change the state_ of that Object. That you do this in a much more direct fashion then with most other Objects does not change that. (Think of passing objects as passing a pointer to a struct, if you come from the C corner). So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? The convention is that only the Class internals should care. Noone outside the class should even know about the private members of the class which may be affected by a method call. As for documentation within the class, there are no hard-and-fast rules for this, AFAIK. Just remember that you should not depend on internal state. (For example, you should not care if a method getResource() just returns a reference to one of the class' fields, or computes an entirely new Resource. This may even change!) I'll only add that should you /really/ need to be sure that a called routine is not mucking with your object, you can clone it first, though doing so without really knowing what's going on is very bad form. Similarly, if you are writing a method such as: public void send(Object sentO) {...} and you really need sentO to stay intact long after your method returns because you are entering it into a larger structure or what have you, your method can clone that object first and use the copy. -- http://www.allexperts.com is a nifty way to get an answer to just about /anything/. Back to top iamfractal@hotmail.com Guest Posted: Fri Sep 17, 2004 2:07 pm    Post subject: Re: Problem with lack of pass-by-result Richard <rh310 (AT) hotmail (DOT) com> wrote Quote: Java newbie. So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? C newbie. What's the convention in C? ..ed www.EdmundKirwan.com - Home of The Fractal Class Composition Back to top Chris Smith Guest Posted: Fri Sep 17, 2004 3:38 pm    Post subject: Re: Problem with lack of pass-by-result Thomas G. Marshall wrote: Quote: I'll only add that should you /really/ need to be sure that a called routine is not mucking with your object, you can clone it first, though doing so without really knowing what's going on is very bad form. Yes, with emphasis on the "very bad form" part. Cloning is useful in situations where you need to make guarantees and can't trust the code you're executing, such as in managing security while running dynamically loaded mobile code from an outside source. It's not a reasonable solution to have a mess on your hands with your own application code. Doing the latter would be taking the first step to eventually rewriting the whole application. -- www.designacourse.com The Easiest Way to Train Anyone... Anywhere. Chris Smith - Lead Software Developer/Technical Trainer MindIQ Corporation Back to top Thomas G. Marshall Guest Posted: Fri Sep 17, 2004 4:22 pm    Post subject: Re: Problem with lack of pass-by-result Chris Smith coughed up: Quote: Thomas G. Marshall wrote: I'll only add that should you /really/ need to be sure that a called routine is not mucking with your object, you can clone it first, though doing so without really knowing what's going on is very bad form. Yes, with emphasis on the "very bad form" part. Cloning is useful in situations where you need to make guarantees and can't trust the code you're executing, such as in managing security while running dynamically loaded mobile code from an outside source. It's not a reasonable solution to have a mess on your hands with your own application code. Doing the latter would be taking the first step to eventually rewriting the whole application. You bet. However for some things it is critical, and for others it is perfect for defensive programming, both of which are my point. For example, take a look at this method in mumble/java/awt/GridBagLayout.java: /** * Sets the constraints for the specified component in this layout. * @param comp the component to be modified * @param constraints the constraints to be applied */ public void setConstraints(Component comp, GridBagConstraints constraints) { comptable.put(comp, (GridBagConstraints)constraints.clone()); } Because GridBagLayout needs to use the passed constraints *long after this method is done with*, it is cloned to prevent the caller from reusing his constraints object and mucking with the innards and screwing up the layout down the road. -- Forgetthesong,I'dratherhavethefrontallobotomy... Back to top Richard Guest Posted: Fri Sep 17, 2004 5:35 pm    Post subject: Re: Problem with lack of pass-by-result [email]iamfractal (AT) hotmail (DOT) com[/email] wrote... Quote: Richard <rh310 (AT) hotmail (DOT) com> wrote Java newbie. So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? C newbie. What's the convention in C? Stylistically, the strongest way I've seen (and used) is passing all referenced vars as function parameters, possibly with additions to the prototype that document intended use*. Second strongest way is to just pass all the changed vars as function parameters. Third strongest way is to declare externs inside the function body. * such as: #define IN #define OUT #define INOUT int function_prototype( int a IN, int * b OUT, int * c INOUT ) ; Granted there's an awful lot of old C around that doesn't do this, written without a lot of concern for maintainability (which is OK; maintainability isn't a globally high priority). -- Globalists want a world of China's labor rate and US retail prices. Back to top Richard Guest Posted: Mon Sep 20, 2004 6:22 pm    Post subject: Re: Problem with lack of pass-by-result [email]matth (AT) ivizNOSPAM (DOT) com[/email] wrote... Quote: "Richard" <rh310 (AT) hotmail (DOT) com> wrote in message news:MPG.1bb0e67c7c6255d3989dfb (AT) news (DOT) verizon.net... Java newbie. As a Java newbie, I'm charged with learning the language by maintaining a pretty good-sized pile of code. The conceptual problem I'm having is with the C convention of using pass-by-result (and pass-by-value-result) function parameters to readily show which functions use and change variables outside of the function scope. In Java, pass-by-result vars are supposedly verbotem (although check BufferInputStream.read(byte[] b, int off, int len) for at least one exception to that rule that I've seen so far). Java very consistently (with no exceptions) follows the rule of pass-by-value where the value of an object is its reference. All objects including arrays of primitives (byte [], int []) are represented by a reference to their associated data. These references are passed by value as input-only parameters. By which I think you mean that I can't create another Object inside of the method and assign its reference value to the reference value of the Object parameter. But I can copy the contents of the new Object into the paramter Object, using the parameter's reference. Seems like simply a way to syntactically insure that any change of Object parameter contents from inside a method is a deliberate act, rather than the indirection happening as an unplanned side-effect. Quote: The references may not be changed, but the contents to which they refer may be. This concept is fundamental to understanding Java. Any object that is passed may have its contents modified, but the reference to the object itself cannot be changed as a parameter. In the read () method above, the (reference to) the array of bytes is passed in and the method fills out the contents of the the array. Note that Java primitives (int, short, byte, long, char) are not objects. There are no references to primitives. http://mindprod.com/jgloss/callbyvalue.html http://mindprod.com/jgloss/callbyreference.html So, what's the Java convention for documenting which class-scope variables (those declared outside of methods) are used and-- specificially--CHANGED by which methods? Method documentation indicates whether the objects refered to by parameters are intended to be modified. Cheers, Matt Humphrey [email]matth (AT) ivizNOSPAM (DOT) com[/email] http://www.iviz.com/ -- Globalists want a world of China's labor rate and US retail prices. Back to top Thomas G. Marshall Guest Posted: Mon Sep 20, 2004 7:35 pm    Post subject: Re: Problem with lack of pass-by-result Richard coughed up: Quote: matth (AT) ivizNOSPAM (DOT) com wrote... ....[rip]... Quote: Java very consistently (with no exceptions) follows the rule of pass-by-value where the value of an object is its reference. All objects including arrays of primitives (byte [], int []) are represented by a reference to their associated data. These references are passed by value as input-only parameters. By which I think you mean that I can't create another Object inside of the method and assign its reference value to the reference value of the Object parameter. But I can copy the contents of the new Object into the paramter Object, using the parameter's reference. Sure, but regardless, it is still /pass by value/. And you are not allowed to copy all object contents: the reflection package restricts you based upon member scope. And as I stated in another ng, even if you /had/ the ability to copy full contents of one object to another's reference, bit for bit, it would /still/ be PBV because: Inside the method, the formal parameter when used as the LHS of an equation /cannot/ affect the actual parameter. Quote: Seems like simply a way to syntactically insure that any change of Object parameter contents from inside a method is a deliberate act, rather than the indirection happening as an unplanned side-effect. I'd argue that it's not a syntactic issue, but a paradigm one. And paradigms by themselves carry syntactic rules, but not necessarily the other way around. ....[rip]... -- http://www.allexperts.com is a nifty way to get an answer to just about /anything/. Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         AppletTalk.com Forum Index -> Java Help All times are GMT Page 1 of 1   Jump to: Select a forum Java Newsgroups----------------Java DiscussionsWeb-based java games comp.lang.java 3D Graphics API's for Java Java: Support and criticism Java System Announcements comp.lang.java.api JavaBeans Java and CORBA Java and Databases comp.lang.java.developer Java GUI Toolkits Java Help comp.lang.java.javascript JVM, native methods and hardware comp.lang.java.misc Java Language Programming Security and Java comp.lang.java.setup Java Software Tools comp.lang.java.tech ibm.software.java.linux  Non-English Java Newsgroups----------------Java Language (German)Language Java (Français)Java (Italian)  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.