Simplest approach for a MMF2 Dev project - Opinions welcome!

[Artematopeya]

Hi there again,

Back in the MMF2 arena after a while, I'm need to replicate a compiler/interpreter architecture for a -now obsolete- DOS proprietary system based on a simple ad-hoc (command-type) language. I would v++++ appreciate your comments and suggestions on how to best approach it since I can't seem how to implement the logic part. To sum up, the system is like a (school-level) new computing language compiler used for language learning. The usual workflow was: the user writes a database definition + programme logic in an ASCII txt file using a simple pseudo-code semantic language, compiles to a data package that includes both the static data (no new data is ever created at runtime) + the app logic and then the compiled package is interpreted at runtime. I assume that this approach can be replicated by creating a -kind of- interpreter in MMF2 that provides the main execution flow from which: a) the static database can be loaded (could be created using a separate MMF2 app in the fashion of a e.g. game level editor type - data are simple variables, text arrays and certain types of objects that could be arranged using INIs or data objects); b) BUT... I can't see how to implement the pseudocode part. I must retain the functionality of the user being able to pseudo-programme his own kind-of apps, but i can't see how to make MMF2:
* either interpret a txt file with program pseudo-code logic at runtime, so that compilation is not necessary. The pseudo code could call what, external lua or python (basic too?) snippets via a scripting object at runtime?
* or: create a logic editor but then, how to transform the pseudo code into a MMF2 friendly language... compiling? %-o

I understand that all of this could be achieved high level languages, but then I wouldn't be using MMF2... I can't loose the pseudo-code facility, which is paramount.

Looking forward to hearing from you guys

[MuddyMole]

It's very hard to say without seeing a detailed specification for the language ("pseudocode" isn't the right word) - does it have a name? It could be anywhere from "borderline impossible and definitely not worth the hassle" to "not too hard at all, if you're good with string functions and fastloops" - largely depending on how dynamic the language is.

For example, with MMF2, people are always asking why they can't construct expressions from strings, like this:
"Alterable Value A(""" + Global String A + """)" where the global string contains the name of an object.

That kind of functionality would be incredibly useful, but it's also exactly this kind of dynamic code that will cause you massive headaches.

[ouly]

Hi Artematopeya,
In such situation I would probably create a new MMF's extension for that. The interpreter would be written in C/C++ and then act as interpreter and interface between MMF events and the other language. It's more or less the same as Basic, Lua and Python extensions.
ouly