i thought this was so funny i had to pass it on:
explains that a fictional drama film in which U.S. President Bush is assassinated received “a prestigious award at the Toronto Film Festival.” Good call, judges!
In the film, computer effects were reportedly used to simulate Bush’s last seconds. Too bad, really - a waste of perfectly good CPU time. Probably fun to watch in slow motion over and over, though.
i recently read a comment to a blog post (don’t remember exactly what site) where Robert Ramey (author of Boost.Serialization) complained (rightfully) that libs11n.so was really big (8MB, if i recall correctly). While this isn’t really a correct blanket statement, it is true if you enable all Serializers and build with debugging options turned on.
So… today i added support to the build process for disabling any unwanted/unused Serializers. When enabling, for example, only the expat Serializer, and disabling debug, the library now compiles down to a lean 200k. When disabling *all* Serializers, the library is a miniscule 136k on my machine. However, disabling all Serializers leaves you with a libs11n which can’t actually load or save anything, so doing so is pretty useless. That said, with just a bit more work in the source tree the Serializers can be built as DLLs. In the 1.0 series they were all built as DLLs by default, but they were changed to built-ins during the 1.1.x refactoring to help make the tree more portable to non-Linux platforms.
This feature, if we can call it that, will be available in the next release (1.2.5), though there is no planned date for that release.
It was 3 years ago - September 2003 - when i finally understood C++ well enough to begin to implement libs11n. It was a task i had been contemplating since a year or more before, but the real work began 3 years ago this month. The part which had taken me so long to understand was how to use abstract object factories to load types at runtime. Once that was cracked, coding went into high gear. About 15 months after the project started, version 1.0 was release (on January 1, 2005). One year later version 1.2 was released. Today there is some experimental work going on for 1.3, but for all intents and purposes 1.2.4 is still the Golden Master.
Without the ongoing feedback from s11n’s users, this project wouldn’t have come nearly as far as it has. It would still be standing amongst the special-purpose or half-completed attempts at serialization frameworks. Instead, however, it stands nearly alone at the top of that pile of solutions. In my humble opinion, libs11n is one of only two complete, generic serialization frameworks for C++, the other being Robert Ramey’s Boost.Serialization library. There are a number of less generic implementations out there, but none of them come close to the capabilities of both libs11n and Boost.Serialization.
If you’re a user of libs11n, please take a moment to pat your computer on the back and tell the software happy birthday. (If you’re not a libs11n user, you’re probably not reading this post.)
Happy birthday, libs11n! And a huge thanks to the community which makes this project possible.