Comparison of Mycology results

My own interpreter, written in version one of the D programming language. CCBI passes all the Mycology tests with flying colours, as I’ve obviously been testing against it.

This doesn’t mean that it’s correct according to the Funge–98 spec: if you feel my interpretation of something is wrong, go ahead and e-mail me.

cfunge is a POSIX-only Befunge interpreter written in C99 by Arvid Norlander, AKA AnMaster. It’s been tested against Mycology since before the first release and he’s been helpful in reporting bugs, so its standards support is very good.

Parts of its functionality are almost copy-pasted from CCBI, and I think AnMaster trusts CCBI a bit more than he should. In some cases, this means that CCBI bugs remain in it for some time. But that does help it in getting a full pass from Mycology.

Quite a few fingerprints are implemented in cfunge, though it shies away from those which require changes to interpreter-internal data structures.

Language::Befunge, by Jerome Quelin, was the only other interpreter in active development at the time of Mycology’s release. It’s written in Perl and claims to support arbitrary-dimension Funges: this is likely what makes it really slow—on my machine, it takes over 15 seconds to run Mycology.

It has received some attention as regards Mycology conformance, and as of 4.12 it gets all the way to the end with only a few minor BADs.

PyFunge is an interpreter by Kang “Tokigun” Seonghoon. It is written, as you may expect, in Python.

This is another one that’s seen improvements since the release of Mycology. It implements a lot of fingerprints now, and receives only a single BAD in all of Mycology: in the INDV fingerprint. This one can really be blamed on the poor INDV documentation, though.

Rc/Funge–98, by Mike Riley, is another POSIX-only interpreter written in C (though not C99, like cfunge). This one is set apart from all the others in that it:

• First implemented Befunge–98.
• Therefore first implemented Concurrent Funge.
• First supported a form of Mini-Funge (libraries written in Funge).
• Supports all the Cat’s-Eye fingerprints and defines a wallop of its own.

It’s got fairly good Befunge–98 support to boot, especially the latest versions: in the summer of 2008 Mike returned to the Funge world, beginning to rapidly bring his interpreter up to Mycology standards and creating many more fingerprints.

In Mycology terms, both 1.27 and 2.02.00 are unfortunately limited by their mistreatment of form feeds: as far as I can tell, they are considered to terminate the file in Befunge mode, when they should be ignored. (Perhaps they increment the z-coordinate even in Befunge?)

It must be noted, however, that it gets a full pass in TRDS! No minor feat, that. In the process of fixing his TRDS code for MycoTRDS he found a few bugs in MycoTRDS as well as CCBI, and we checked to see that our TRDS implementations (the only two known to exist) behave identically.

An interpreter in C++, Stinkhorn, originally named befunge98, by Lee Houghton now manages to get all the way to the end of Mycology with just a few minor BADs.

This was the first interpreter to be updated after Mycology was released. It’s nice to see that Mycology had some impact. Since then, of course, we’ve got all sorts of updates: newcomers like cfunge and returning giants like Rc/Funge–98.