Hacker News Reader

Comment by iaskwhy | Hacker News Reader

I guess you need to change the Hardware to "Win" to see it shining.

Indeed, mono performance isn't great. That is the problem with C#, you need ultimately to run it on IIS for web platform deployment, sadly under anything else it is rather slow.

Also these tests are a bit unfair, as I think they give benefit in some areas to frameworks that cache by default, rather than by request.

Ultimately, what I want to know is total cost, that means hardware, licenses / support agreements, development time, developer ease of hiring.

That isn't to say these aren't interesting, but they are rather apples to oranges. As I posted in the other flamebate thread about MS stack > *, I have set a note in my calandar to help improve the example for MVC.Net (https://news.ycombinator.com/item?id=5976540). Hopefully I will be able to come then when not in an exclusive agreement, they suck I know, but it pays my mortgage!

I missed your later reply in that other thread. If the node.js implementation is caching results from the database, it should be modified to not do so. The original implementation did not, but it has seen several updates from node.js SMEs. Can you point to what makes you suspect that it is caching? The requirements do not allow for caching [1]. Future test types (perhaps in Round 7) will allow for caching, but none of the present test types allow for it.

[1] http://tewebdev.techempower.com/benchmarks/#section=code

Even then it tends to run behind Go and Node.js in most tests. Once Round 7 includes MSSQL we'll see a more accurate picture.

I don't find it a problem seeing it running behind Go and Node.js, I believe C# is another beast so it's a trade-off.

Java still beats it on Windows.

I don't see a framework-to-framework comparison. The .NET equivalent to a servlet is a "HTTP Handler"

Reply on news.ycombinator.com