> > I looked at the source/sample files (didn't actually install it.). Given > that, these are its shortcomings: > > 1) scheme like syntax (may or may noot be a shortcoming depending on your > like of it) Nothing more to say here. > 2) improperly written scripts could crash the entire process. I am currently writing on that part of code. You may expect correct error handling in a couple of days. It took so long to implement error handling because it is the less documented feature of Guile so it took more time to find a correct solution. > 3) each npc action requires a seperate script file. So one complex npc may > have half a dozen different > script files. Yes. I didn't see this as a problem, but I think you are right. I'll then implement a script...endscript tag to correct this (should not be a big problem because it would be very similar to the msg...endmsg structure). > 4) Require guile to be installed on system (ie, more software dependencies) > That's the price to pay if you don't want to rewrite a scripting language from scratch. > good things: > 1) Appears to have support for lots of different actions, so fairly > powerful 2) its real life code, vs something on the drawing board. > > my personal opinion is that as long as points 2 & 3 exist in the > shortcomings, its not useful. #2 is a major one, but #3 just in terms of > maintainabilty - I would much rather have something line a msgscript .... > endscript in the object, and just have a big buffer in the object that > contains the script. When something is done, it goes look at this buffer. > This way, the script and object using the script stay connected. > > Note this is just from looking over the code and docs. If I'm wrong on > any of this, I apologize to the author. > Do not apologize, you are 100% right. I put the patch on the list to get feedback like this to know what I should do next to improve it, so I thank your very much for giving your opinion. Chachkoff Y.