I often hear people talking about which technology is better between C / C++ / C# .Net / java. Most of the time, it’s more a political/brotherhood/community thing (like football, even thought frenchies like me aren’t so proud of their team now) than a technical talk. I find it absurd. Computer science is about efficiency and making more money. You can take into account the pleasure you have to work on a technology (as you will be more likely to be more efficient with it), but it should still remain about efficiency.
I found out that there is real test than the little one i did here. In fact the main interest of the post is this spreadsheet : I recently did a simple and stupid comparison test between C++ (g++) and C# .Net (mono). My point here is that C# can actually be considered as a very fast language. It allows automatic hardware-specific optimization. The only real drawback you have in the .
I talked about this earlier. The mono tools for Visual Studio is a commercial product that easily enables you to test and port your Windows applications to mono. As I told before this is really a great idea. Because I think Visual Studio is the best IDE and Linux is the best server OS (in my opinion, it’s more efficient and more robust than Windows Server). So, I think it’s the perfect gateway between this two very different (but very close since Mono appeared) environments to achieve greatness.
I’ve switched from my two three years old dedicated servers to one brand new virtual server. Reasons are : These servers costed me too much and they were becoming old (risk of failure increases). It wasn’t worth it. I spent last night doing that because I didn’t want to interrupt anybody using these servers. My two servers were running some Debian and I’m now switching to a CentOS virtual server. I was a little bit worried at first that CentOS would have a crappy package management system, but its yum is in fact working the same way as Debian’s apt-get and OpenSuse’s zypper.
Yes, I have tested MonoVS with the version 0.2.2641 (on both client and server). I installed OpenSuse 11.1 and added the MonoVS software repository, and everything worked ! I would have prefer to get it from SVN in order to use it in my Debian hosts but the mono development team seems to have removed it from their SVN repository. So, the Mono Remote Debugger for Visual Studio works, but there still some bugs.
Just a little post for all these people who seem to think Mono is just an other short-term open-source software. I’ve used it for quite some time with a production “real time” network server, which is running for something like 6 months now, and it performs very well. I do everything on my Windows host and then copy and launch the final app on the Linux host. But there are still two problems :
I talked some time ago about a library I made to take advantage of the kernel network events. I now release it and explain how to use it. It can help people to do some little network software without knowing where to start from. I built it for network servers made to communicate with remotely connected embedded chips. I wanted to be able to always stay in touch with a huge number of chips without any real cost.
As I already told before, I love the Mono project. It enables to run the powerful Microsoft .Net Framework on UNIX/Linux/BSD systems. I recently wanted to test a very cool feature of ASP.Net on a mono server. So I did a little apt-get install lighttpd mono-fastcgi-server2 -y The feature I wanted to try was a web scripting method ( with the [WebMethod] attribute) exporting some JSON directly from your method return value.
I have built a simple tcp/udp network server library. I used it on lots of little programs and on one pretty important vehicle tracking server. The first version was using one thread for each connected client. It was working but it consumed a lot of memory. When the server reached something like 3000 simultanneous connections, the kernel was killing it for consuming to much memory. I built a second version which was using network events.
Recently, I was faced with a little problem. I built a .net program which calls a MapPoint WebService. It worked fine on Windows but failed on Mono/Linux with a “401 Unauthorized” error. As it really made no sense, I decided to listen to the network communication. It did it with wireshark on my computer and tcpdump on the Mono/Linux host. And by looking at the header of the HTTP request, I noticed they were some slight differences.