Adam Curry has been playing with BitTorrent:
I’ve been using BitTorrent for a couple of days now.
This is promising stuff.
The BT application is fired automatically from your browser when you click a link to a .torrent file. I like this approach, since users are used to having ‘helper applications’ launch from their browser.
Jon Tirsen hits the nail right on the head here!
Persistence is only complicated because we make it complicated. Somewhere back in time, it was drilled into our heads that we must use relational databases for persistence. For many applications this makes a lot of sense, but the complications it adds to every day development is just sometimes not worth the effort. Huge advances have been made in the Java world in O/R persistence tools such as JDO and Hibernate (.NET still has some catching up to do here) but using a relational database is just a major headache sometimes.
The LAN connection on my router is hosed. I’m sure it’s either a matter of swapping NICs or cords. *Sigh* No SSHing from work today…
Greg Klebus said something yesterday/last night/this morning that sounded right. (Things are tricky when you comment on something written on the other side of the world.):
But there is hope here. More and more manufacturers do provide Linux support. It sometimes may be kind of unofficial, weak, hard to find (sic), but still is there. And growing.
So, folks, don’t ask what Linux can do for you, ask what you can do for Linux. Buy only devices that are supported on Linux. If you have one that’s unsupported, bug the manufacturer asking for Linux support (at least one email). Otherwise, Microsoft is going to kick Linux off the desktop market completely.
Kerneltrap has the story:
GCC 3.2.1 (the last in the 3.2.x series, it seems) has been released. Lots of bugfixes in here. Notably, some MMX, SSE, and x86-64 bugs have been fixed.
Additionally, the method of constructing the list of directories to be searched for header files has been revised. If you specify a standard include dir with -I (for example: -I/usr/include), the compiler will ignore it; this is so you don’t change the search order for standard include dirs.
For more information, visit the GCC 3.2 changes page, or look at the Release Notes. This bug release follows GCC 3.2.0 which was released in mid-August [story].
Werner Vogels has released his source so that you can run your own C# SciMark tests:
At popular request I have put the C# version of the SciMark benchmark up for download. Please keep in mind that this is a simple benchmark and that mainly the comparison with the Java version makes sense. For example the random number generator in the C version has no “synchronization”. If you remove the synchronization from the the Java and C# versions you will see that the performance of the Monte Carlo section is close to the C version. The zip file contains the C#, J# and C versions. The pure java version you will have to download from NIST.
If you really want to do some decent benchmarking you will later to wait until we release the microbenchmarks (threads, matrices, etc),and the ports of JavaGrande and Linpack to C#.