Sorry for replying a bit late.... I don't read this section often.
What you're seeing or being made to download is the actual "code" of the image that you're supposed to be viewing. The reason you're seeing this is Yoshi 15's method basically takes the code of the image on the server, and throws it at your browser, assuming it'll sort everything else out on its own.
Why do some browsers handle it properly, and some don't? Or different servers produce different results? Most of the time when files are sent to your web browser, they're sent with a header that tells your browser how it should handle the image. Using Yoshi 15's method, the wrong information is sent in this header. Again, some browsers and webservers will deal with it better than others.
Try using the second method suggested by Mercutio, or the method I suggest, over in
Signature Chit-Chat.
How are these methods different? Instead of sending a raw image, they send just header code. This header code basically points your browser to another file somewhere on a webserver, which it downloads just as it would any other normal image, letting the server and browser handle the images the way they're used to.
Sorry if I've provided a bit too much unnecessary information. Worth being educational, though.