Lambda the Ultimate

Greetings,

I've posted here in the past and just had a few questions about the basics of parsing and perhaps wanted a bit of insight on parsing theory.

I'm currently writing a parser compiler, to better understand the formal side of parsing, and into that I hand-wrote the grammar description language's parser. One of the major things I've learned through this is how very little I really know now and knew when I started.

In the general case, when parsing the full file you begin from the start rule, and go from there, that much is pretty basic. The parts I get a bit hazy on are how to handle partial parses when faced with interactive environments, such as the user typing the text into Visual Studio and integrating your parser infrastructure into the Visual Studio Extensibility. The current method I use involves parsing the entire thing every time, with a asynchronous delay kicked off to pause each time a key is pressed, so it's not parsing everything literally every key press. The major question I have is how to merge the parse trees and isolate the impact of a change to the concrete/abstract tree that results from the user's text.

Is there research based off of interactive parsing I could look at to get an idea?

In the future I want to understand this concept well enough to automate its integration into the project I'm writing now, even if it's in the second or third iteration. I have a feeling it's complicated, perhaps more than I'm thinking now.

Insight is very welcome.

Also note to people posting links to published documents: I don't have any subscriptions to any journals/publication websites, so it might be difficult to follow if such sources are referenced since they usually request a monetary fee for retrieving their data.