[Haskell-cafe] Data.Binary Endianness

Bryan O'Sullivan bos at serpentine.com
Tue Sep 11 12:16:45 EDT 2007

Jules Bean wrote:

> For these reasons, although it is very cool, I don't think it can be 
> recommended as a basis for long-term file format definitions.

Indeed, the authors have never claimed that this is what it's for. 
Unfortunately, because the authors haven't *disclaimed* this as a 
purpose, people have fairly reasonably assumed that this *is* the intent 
of the package.

In conversations with Don and Duncan, they've always been quite clear 
that Data.Binary is intended to shovel bits rapidly and with a 
reasonable interface.  All of the things of which you speak, and more 
useful ones such as RTTI and representation of cyclic data, ought to 
live in a higher-level library.  Said library merely hasn't been written 

> (All of the above speaks of the 'high-level' Data.Binary not the 
> 'low-level'.)

Data.Binary *is* the low-level Data.Binary :-)


More information about the Haskell-Cafe mailing list