Saturday, January 22, 2011

CMUD versus zMUD

The process of trying to get the mapper in CMUD working took far too long, and I don't just mean the couple of months when I had to put the project aside and work on other things. I spent a few hours on it yesterday and didn't quite get it working. More importantly, I just got too frustrated with really basic instability issues.

Just to be clear, at this point, I don't have anything near a system set up in CMUD. All I have is a few triggers (manually made there) and the Rainbow Queue system, which is a few triggers and a few aliases. That's it. And yet I still have some troublingly basic instability issues. Notably:
  • I gave up on importing zMUD settings. Even importing from a very bare-bones zMUD file created for the purpose with nothing in it causes CMUD to lock up, 100% repeatably. What could possibly be unusual about my system to cause this?
  • Clicking on any of the other tabs on a trigger causes it to lock up. You can't even test if text matches it. This is not 100% but is highly repeatable: it seems any trigger in the form {some text|some other text|a third possibility} stands a good chance of doing it, though if I delete and recreate it sometimes the new one works.
  • Right now, whenever I try to run CMUD from the Start Menu shortcut, it simply beeps and closes down without any other action. I can see it appear in Task Manager and then vanish. This is true even if I delete and recreate the shortcut. Double-clicking on the actual EXE works, though. Bizarre.
I'm not happy about abandoning CMUD. The promise of working on Windows versions later than XP is a big one to lose. The improvements, small though they are, in the programming language are nice. The chance to use ATCP or GMCP, and the hope of future versions using these for the mapper (and using that XML imported map) are huge. But if, in version 3, after years of people saying "too unstable" and Zugg saying "no, really, I've done a lot of work on stability this time", I can't even get the program to run from a shortcut, or examine a trigger in a bare-bones settings file, the fact that I might be able to work around these and make a usable system is not sufficient. I'd be having to fight with CMUD, instead of work with it, ten times more than I do with zMUD. And that's saying a lot.

Technically I'm not saying "never" because, who knows, CMUD version X.Y might actually be serious about fixing the stability issues. However, at this point, I think I can hope for MUSHclient or Mudlet to take seriously the shortcomings in their mappers faster. So I suppose that means, disappointingly, that I have to stick with zMUD and try to make my Rainbow system in it after all.

So now I'm cranky about all the time I wasted on CMUD. From the fact that months have gone by and I've been able to put in about one week's work on this, you can conclude that time is a significant factor. (Even when I take a whole week off from work, inevitably I get given a lot of chores to do and half of it gets used up before I can start.) I dig into working on Rainbow reluctantly, worried that that effort will all be wasted too, because Mudlet will have gotten usable before I can finish. But I've got to do something. Waiting isn't working either.

No comments: