I want an internal static implicit operator :)
February 22, 2006 at 7:02 pm | In C#, GIMP | 2 CommentsHere is the problem: I have a struct of type GimpRGB which is basically the C# version of the same structure in C. I need this because I have to marshall functions between C and C#. Now in C# I use a class RGB which encapsulates this GimpRGB struct. In the API of my library GIMP# I don’t want to see this GimpRGB struct, so I use an internal property in the RGB class to get this struct: Continue reading I want an internal static implicit operator :)…
More Heartmath analysis
February 22, 2006 at 1:21 pm | In skepticism | Leave a CommentIn my previous analysis on the Heartmath article about predicting the future I already showed that with a very simple model it’s easy to predict that future, because of their faulty test set-up. I have given this some more thought and instead of running the simulation I derived mathematical formulas for the chance of predicting the future. Continue reading More Heartmath analysis…
Unit testing GIMP
February 20, 2006 at 8:28 am | In GIMP | Leave a CommentEveryone seems to agree that testing software is important. Hardly anyone does so in practice. Especially open source software can benefit from good testing and there are a lot of great tools available. Last week I wrote a GIMP# new plug-in that is based on NUnit. This plug-in does unit testing on all GIMP# classes and thus indirectly also unit tests GIMP itself.
For GIMP# this is a very important step to improve the quality even further, since I have written a lot of code (mostly C# wrappers around the GIMP API) that isn’t called by any plug-in yet. The unit tests will ensure that the wrapper will work as advertised.
The new plug-in will be available in GIMP# 0.8 and can be enabled with the –with-unittest option in the configure script.
Randomness for dummies
February 16, 2006 at 3:53 pm | In Ramblings, skepticism | 1 CommentI just found a nice article (in Dutch). It describes an experiment from Ruper Sheldrake that “proves” that people know when they are stared at. In this article they debunk that experiment by showing that the test setup described by Sheldrake wasn’t actually that random at all. He has made the same mistake as the people from the Heartmath institute that claim that your heart can predict the future. Look in one of my previous blogs for more details.
Oh, and if you want to be utterly bored, you can read Sheldrake’s articles here and here.
Linux and moms
February 14, 2006 at 8:33 am | In Ramblings | 2 CommentsNow imagine that my mom uses a Linux system. And let’s also imagine that she just bought a new 20 inch LCD because her eyes are getting worse and someone gave her a nice video card because her old one didn’t have DVI output. Then the following conversation over the phone could take place: Continue reading Linux and moms…
GIMP# progress
February 10, 2006 at 10:06 am | In GIMP | Leave a CommentWork on GIMP# is progressing nicely. The code is becoming leaner and cleaner and there’s still some low hanging fruit out there, for example replacing one of the enumerators by an implementation using yield. I also switched to using GTK# 2.0, which was relatively painless. Only downside is that some widgets have become obsolete and I really want to replace those before I release GIMP# 0.8
Of course we will only release 0.8 if there are also enough user visible changes, but some of the work on existing and new plug-ins is coming on nicely, especially the (new) Swirlies plug-in and the Splitter plug-in. Hopefully Max can also iron out a few bugs in the Picture Package.
More on performance
February 6, 2006 at 8:04 am | In C#, Programming | Leave a CommentI did some more performance testing, trying to compare C, C# and Java. The test was done using the swirlies calculation as described in one of my previous posts. We already found out that Java was about 4 times as slow as C# which was caused by Java using a more exact calculation of the standard trigonometry functions.
Last night I did the same “benchmark” with the original C code. Result of this unofficial test is that on Linux the C implementation was exactly twice as fast as the C# version, using Mono.
Although I think that two times as fast is still a large difference, this is not a bad result. Firstly because Mono is probably going to become faster in the (near) future so this difference will become smaller. And secondly, it looks that C# is the right choice for writing GIMP plug-ins since this is a lot faster than interpreted plug-ins written in Scheme, Perl, Python or any of the other available languages for GIMP.
Blog at WordPress.com. | Theme: Pool by Borja Fernandez.
Entries and comments feeds.
