Not reinventing the wheel

I found someone else's Berkeley DB wrapper online: bdbsharp. The programmer took a fairly different approach from what I was thinking of doing. Instead of compiling Berkeley DB in Managed C++ and working in a single project, he first compiled a thin C wrapper dll around the Berkeley DB project to hide some of the more strange aspects of the API and then wrote a wrapper in C# using P/Invoke on the C dll. It actually seems a lot more elegant that what I had in mind. The interface is a little strange, but I'm sure I'll get used to it.

So now that I know how I want to handle the backend, I've started work on the UI elements. Right now I'm using a RichEdit control for entry of markup text and have embedded Internet Explorer to do the display of the transformed text. Probably the RichEdit will be kicked out in favor of something more Ink friendly, probably an InkEdit control.

I want to be able to embed sketches directly into the final produced page without too much hassle, but I don't want to mix sketches and text directly in the user inputted markup like OneNote. I'm not entirely sure how I want to handle that, but I'm considering some sort of embedding tag that can be used to link pictures or other multimedia in the page. Then I'll attach a basic image editor that can mess with the embedded file.

That's for later though. I think the next step is to start thinking about the markup I want to use and start working on a parser for that. It would probably be easier to just use HTML, but that seems to go against the Wiki philosophy. Of course, the first question would be: do I really care about the Wiki philisophy? But at the very least, the full a href thing is definately too cumbersome for what I want. Another thing is that some of the html formatting, especially the & thing, I have trouble entering via the digitizing pen, so a different syntax is definately in order.