Re: java serialization alternative
Bob Lee <[email protected]> Wed, 9 Jul 2008 10:21:38 -0700
| Newsgroups | gmane.comp.windows.devel.java.advanced |
|---|---|
| Message-ID | <[email protected]> |
First off, I'm not a fanboy of anything, and I haven't drank any kool-ade. Have you ever known me to be dogmatic about anything? 2008/7/9 Ted Neward <[email protected]>: > Odd. I can't remember the last time Serialization caused either my feet to > fall off, or my code to throw a runtime exception at the worst possible > moment. > That's probably because you're smart enough to not use serialization for anything non-trivial. I wasn't so smart and still support major applications which heavily depend upon serialization. You can't accuse me of not knowing serialization. I know the spec inside and out. As the core lib lead for Android, I've had to implement serialization itself (thanks for the help, Harmony!). I was the first to implement serialization support for cglib proxies. Out of necessity as a user, I've implemented tools for debugging and unit testing serializable code ( http://tinyurl.com/63g5hj). How anyone lived without traces for serialization exceptions before Java 6 is beyond me. I'm certainly not the only one who thinks this way about serialization--take a poll. I'm just the only one stupid enough to argue with Jeff. If Google wants to reinvent Serialization in a more statically-bound, > strongly-typed format, there are obviously cases where they found it > useful, > and more power to them, particularly given that they've given it away to > the > rest of the world. Perhaps Google will play nicely in the Java community > and > open a JSR to include PB's into the JDK 7 release; perhaps not, and perhaps > that doesn't make a whit of difference. I care about the approach, not the implementation. Use Thrift, or OR mapping, or XML binding, or whatever for all I care. FWIW, tools like Terracotta appear to have similar problems to serialization. > But watching you two bicker over > something trite that essentially reflects the fact that each of you has > drunk different Kool-Ade is really quite an inefficient use of bandwidth > and > time. > This is analogous to a static vs dynamically typed language debate. It's simple: if you care about maintainability, use a statically typed language and a statically typed serialization framework. If you don't, use Python and flattening. And yes, I'm grumpy, because this is just the tip of the iceberg. Now I get > to watch dozens upon dozens upon hundreds of "PBs vs. Serialization vs. > XML" > debates that will--all of them--reflect nothing more than the inherent > biases and fanboi-isms of their blog authors, when in fact the comparison > is > apples vs. oranges vs. coconuts, and nobody will want to stop the shouting > long enough to realize that. You're right, "PBs vs. Serialization vs. XML" is an "apples vs. oranges" comparison. PBs can produce and consume XML for example. Bob =================================== This list is hosted by DevelopMentorĀ® http://www.develop.com View archives and manage your subscription(s) at http://discuss.develop.com