Re: doc strings for both interfaces and concrete implementations

2012-08-04 Thread Warren Lynn
I won't object, and the alter-meta! approach allows you to do that. Of > course, it cannot provide documentation for the implementations, but the > interface docs may be extended with impl docs incrementally at the > location where the implementation is defined. That's at least better > than

Re: doc strings for both interfaces and concrete implementations

2012-07-30 Thread Tassilo Horn
Warren Lynn writes: >> In general, all different versions of a function should somehow do >> the same thing, so with separate docstrings you'd need to repeat >> yourself. A good guideline is to write the "big picture" first, >> followed by the meaning of the different parameters. >> > I agree th

Re: doc strings for both interfaces and concrete implementations

2012-07-29 Thread Warren Lynn
> In general, all different versions of a function should somehow do the > same thing, so with separate docstrings you'd need to repeat yourself. > A good guideline is to write the "big picture" first, followed by the > meaning of the different parameters. > > I agree the design should keep a

Re: doc strings for both interfaces and concrete implementations

2012-07-29 Thread Tassilo Horn
Warren Lynn writes: > Is there a way to create doc strings on both interfaces (procotols, > multi functions) and concrete implementations (protocol > implementation, multi-methods)? No, I don't think so. > Assuming we don't have it yet, what I am hoping for is, when I do (doc > symbol), or use

doc strings for both interfaces and concrete implementations

2012-07-29 Thread Warren Lynn
Is there a way to create doc strings on both interfaces (procotols, multi functions) and concrete implementations (protocol implementation, multi-methods)? For now it seems to me I can only have doc strings for the interfaces. If that is the case, I feel it is something worth fixing, because in