Why write your own rendering engine anymore?

I just installed and ran the latest version of Lavasoft’s Ad-aware Free (Anniversary Edition) and was disappointed by the general fuzziness of the text.

image

So, I grabbed one of my favorite tools, Spy++ to do a little bit of technology spelunking. The interesting thing was that there was only a single window in use:

image

The fuzziness of the text initially made me think GDIPlus, but a quick scan using Depends seemed to indicate that wasn’t the case:image

The text rendering just isn’t right:

image

For example, check out the kerning of the “e” and the “m” above.  I tried the same text in Photoshop (wondering if it was a bitmap from a common image generation tool):

image

But, it looks fine in every aliasing option.

It wasn’t GDI, or GDI+:

image 

And not WPF:

image

There are probably a handful of toolkits that it could be, but I don’t have the patience to try a match anymore.

I’m just not sure why they’d take the time to not use a standard Windows provided text / graphics engine? I’m not opposed to an architecture that minimizes HWND usage to increase efficiency or allow for some unique user experiences (like skinning), but why go this far? What’s the gain? When the quality of text isn’t near perfect, why wouldn’t you abandon the technique? Anyone know?

One Comment

Comments are closed.