Re: [Haskell-cafe] How to determine if a FilePath is a directory name or regular file?

2009/6/22 Duncan Coutts :

On Mon, 2009-06-22 at 08:53 +0200, Deniz Dogan wrote:

...

2009/6/22 Colin Paul Adams :

...

...

...

...

>> "Judah" == Judah Jacobson writes:

   Judah> On Sun, Jun 21, 2009 at 11:12 PM, Colin Paul    Judah> Adams wrote:    >> I've been hoogling like bad to try to determine if a function    >> like this exists.    >>    >> getDirectoryContents returns sub-directories as well as file    >> names. I want only the latter, so I'm looking for a suitable    >> filter.

   Judah> Use System.Directory.doesDirectoryExist/doesFileExist.

Thanks.

it seems it's time i went to the optician again.

I'm not surprised that anyone would make the mistake. I think that the two functions should be named "isDirectory" and "isFile", but it seems that "isDirectory" was already taken by another function in System.Directory, which is quite unfortunate. "does" goes against the intuition one gets from pretty much everything else in Haskell, where "is" seems to be the convention. In fact, Hoogle only knows about three functions which start with "does".

One explanation is that isBlah asks "is this thing a blah", but we're not asking that because there is an indirection via the filepath. We're asking "does this filepath refer to a directory" not "is this filename a directory". The latter could be a function:

isDirectory :: FileInfo -> Bool

along with a hypothetical

getFileInfo :: FilePath -> IO FileInfo

Duncan

I think see what you mean, but I find the argument more of an excuse to the poor naming than a solid argument for it. Following the convention and intuition that most users have should be more important than making the (sometimes unnecessary) distinction between a directory and the path to it. -- Deniz Dogan