[Haskell-cafe] Re: String vs ByteString

[Donn Cave]

Quoth "Bryan O'Sullivan" <bos at serpentine.com>,

> In the case of the text library, it is often (but not always) competitive
> with bytestring, and I improve it when I can, especially when given test
> cases. My goal is for it to be the obvious choice on several fronts:
>
>    - Cleanliness of API, where it's already better, but could still improve
>    - Performance, which is not quite where I want it (target: parity with,
>    or better than, bytestring)
>    - Quality, where text has slightly more test coverage than bytestring

That sounds great, and I'm looking forward to using Text in my
application - at least, where I think it would help with respect
to correctness.  I can't imagine I would unpack all my data right
off the socket, or disk, and use Text throughout my application,
because I'm skeptical that unpacking megabytes of data from 8 to
16 bits can be done without noticeable impact on resources.  I
wouldn't imagine I would be filing a bug report on that, because
it's a given - if I have a big data load, obviously I should be
using ByteString.

Am I confused about this?  It's why I can't see Text ever being
simply the obvious choice.  [Char] will continue to be the obvious
choice if you want a functional data type that supports pattern
matching etc.  ByteString will continue to be the obvious choice
for big data loads.  We'll have a three way choice between programming
elegance, correctness and efficiency.  If Haskell were more than
just a research language, this might be its most prominent open
sore, don't you think?

	Donn Cave, donn at avvanta.com