Comments on: Singleton’s new Tinfoil Hat called “Don’t clone me!” http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/ Jing Xue's Weblog Tue, 06 Jan 2009 22:51:19 +0000 http://wordpress.org/?v=2.7 hourly 1 By: Howard D'Souza http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/comment-page-1/#comment-30319 Howard D'Souza Fri, 02 Mar 2007 21:52:11 +0000 http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/#comment-30319 Avoiding implementing Cloneable is easy. A bigger problem arises with Serializable. I could argue that it is sometimes necessary for a singleton to implement Serializable (though I could argue the flip side just as easily). Implementing cloning through in-memory (de)serialization is not only more robust, but ensures deep copying of fields. So... pop quiz. How does one prevent cloning of Serializable singletons. Answer: Don't try. If you think of design patterns as something that makes your life easier, but you need to cooperate for it to work - you won't be so concerned about playing Singleton-nazi. Just put a caveat-emptor in your Javadoc and move on to the next candidate :-) Avoiding implementing Cloneable is easy. A bigger problem arises with Serializable. I could argue that it is sometimes necessary for a singleton to implement Serializable (though I could argue the flip side just as easily). Implementing cloning through in-memory (de)serialization is not only more robust, but ensures deep copying of fields. So… pop quiz. How does one prevent cloning of Serializable singletons.
Answer: Don’t try.
If you think of design patterns as something that makes your life easier, but you need to cooperate for it to work - you won’t be so concerned about playing Singleton-nazi.
Just put a caveat-emptor in your Javadoc and move on to the next candidate :-)

]]> By: Jing Xue http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/comment-page-1/#comment-29966 Jing Xue Tue, 27 Feb 2007 18:16:48 +0000 http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/#comment-29966 LOL, yes, Ricky, you can have my job. :) I should have probably added that I ask that question a lot in interviews, not because I want to encourage singletons, but because it's something most developers know about, and can act as a good platform for some more interesting follow-up questions regarding design patterns and multi-threading issues, etc. LOL, yes, Ricky, you can have my job. :)

I should have probably added that I ask that question a lot in interviews, not because I want to encourage singletons, but because it’s something most developers know about, and can act as a good platform for some more interesting follow-up questions regarding design patterns and multi-threading issues, etc.

]]> By: Ricky Clarkson http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/comment-page-1/#comment-29939 Ricky Clarkson Tue, 27 Feb 2007 13:46:25 +0000 http://www.digizenstudio.com/blog/2007/02/26/singletons-new-tinfoil-hat-called-dont-clone-me/#comment-29939 I would implement a singleton as a bag of static things, unless it needed to implement some interface. That shows it for what it is commonly used for, a set of global variables. Then I would remove it from my code, before the smell gets too bad. Do I get the job? I would implement a singleton as a bag of static things, unless it needed to implement some interface. That shows it for what it is commonly used for, a set of global variables.

Then I would remove it from my code, before the smell gets too bad.

Do I get the job?

]]>