[Haskell-cafe] Enterprise Haskell -- help

[Jason Dagit]

On Thu, May 26, 2011 at 1:43 PM, Gaius Hammond <gaius at gaius.org.uk> wrote:
>
> On 26 May 2011, at 21:34, Jason Dagit wrote:
>
>> On Thu, May 26, 2011 at 1:09 PM, Gaius Hammond <gaius at gaius.org.uk> wrote:
>>>
>>>
>>> Over in OCaml-land, I have taken it upon myself to address this:
>>> http://gaiustech.github.com/ociml/
>>
>> Takusen already supports Oracle (and other rdbms) in a resource
>> precise and good performance way, if that's what you're talking about
>> implementing:
>> http://projects.haskell.org/takusen/
>
>
>
> Ah, interesting - I was only aware of HDBC and HSQL. Which in turn begs the
> question, why isn't this better known among (us) Enterprise dev types?

I'm not certain, but I think the biggest problems getting in the way
of Takusen's greater adoption are:
  * It requires a deeper understanding of Haskell than HDBC due to use
of functional dependencies, rank-2 types, and left-fold enumerators.
  * Less documentation than its competitors
  * The name is good (it's the japanese word for oracle), but doesn't
evoke "that's clearly a db library" in most english speakers.

As the current maintainer (and I've been a really lazy maintainer,
still hoping to find a proper long term maintainer), I'm interested in
any feedback about Takusen.  Things I would like to eventually
implement:
  * HDBC compatibility layer to make it easier to transition from HDBC
to Takusen
  * Replace the left-fold enumerator code with a proper iteratee
library (Takusen is by Oleg but predates his work that generalized the
approach).

Jason