Well, as you can see from the topics title, I am thinking about writing my own Client. It's not terrible difficult, just a lot of work, and now I have some questions for the people better in such things:

1) Would a hardcoded curesystem be better than something like...zMud or so? I could call on it directly after receiving the streams, so I thought it might be at least 0.2, 0.3 seconds faster?

2) Another thing, would it be actually worth it, comparing the time to code it with the time spendable in other things?

EDIT: Not to mention that I could actually code in a real language (C#) and not that bastardized thing of zMud-language.

It would be worth it totaly! You could sell it for credits or something. Advertise it being so good, even the Gods'll use it! And even then, sell it!!!

You can definitely write a system to be faster than anything written in a client's scripting languages. It'll be quite a bit more work to write an entire client, however, and I highly doubt it's worth it. Just think of the challenges you need to overcome before you even get to the really cool stuff, such as parsing ANSI color codes, dealing with disconnects, multi-threading to avoid contention between foreground (GUI) and background (triggers) execution, etc.

If you really want to do this, though, I recommend trying a COM system as a plugin to something like zMUD or MUSHclient first. You can get a lot of speed that way on processing the queues and priorities without the huge overhead of handling all communications and processing yourself.

You know how Whyte's mapper is a pre-parser that prints out to your client? Well, the mapper is just a plugin for his pre-parser - he has combat plugins as well. And they're smokin' fast. Just don't be tempted to start running in-client scripting at the same time.