Personal tools

Wanted libraries

From HaskellWiki

(Difference between revisions)
Jump to: navigation, search
m (Link .NET to salsa?)
(rm wanted .NET - since salsa now exists, add harpy to Could be improved section)
Line 17: Line 17:
 
=== Numerical algorithms ===
 
=== Numerical algorithms ===
 
Stuff to compute bessel functions, struve H1 etc.
 
Stuff to compute bessel functions, struve H1 etc.
 
=== .NET/Java Interoperability ===
 
We could use a .NET FFI library that lets you easily call into the .NET framework. Perhaps the (abandoned?) [http://galois.com/~sof/hugs98.net/ Hugs .NET FFI library] could be resurrected/updated and ported to GHC? Maybe a tool could even be created that automates the interoperability of Haskell and arbitrary .NET libraries?
 
 
: I'm working on a Haskell to .NET bridge for my undergraduate thesis. The goal is to provide access to arbitrary .NET libraries from Haskell, with a convenient syntax. I expect to have something releasable in a few weeks. [[User:AndrewA|AndrewA]] 13:33, 6 October 2007 (UTC)
 
: That is [[Salsa]] ?
 
   
 
=== AMQP Client ===
 
=== AMQP Client ===
Line 40: Line 34:
 
Problems with existing solutions, that could be improved.
 
Problems with existing solutions, that could be improved.
   
  +
* Harpy lacks support for x86-64 - supporting both x86 and x86-64 would be a boon.
 
* The core HTTP libraries need a lot of work. There is the beginnings of a useful library there, but many http codes are not handled, and there should be simple 'get' and 'head' functionality.
 
* The core HTTP libraries need a lot of work. There is the beginnings of a useful library there, but many http codes are not handled, and there should be simple 'get' and 'head' functionality.
 
* Binary serialisation support should be ported to bytestring, and added to the extralibs bundle
 
* Binary serialisation support should be ported to bytestring, and added to the extralibs bundle

Revision as of 00:54, 18 November 2008

This page provides a list of missing and found libraries for solving common real world programmer tasks.

If you think there should be a library for some common task, add it to this page, and someone may well step up and write it. If you know of a library to solve an open problem domain, link to it.

The first place to look is the full libraries list.


Contents

1 Missing libraries (by problem domain)

1.1 IMAP

A functional IMAP library would be nice, particularly one that can be used in conjunction with encryption and authentication. References have been found to a SoC project although it appears to be derelict.

1.2 Numerical algorithms

Stuff to compute bessel functions, struve H1 etc.

1.3 AMQP Client

It will be great to have a client for the open source protocol AMQ (http://en.wikipedia.org/wiki/Advanced_Message_Queuing_Protocol). This is a new transport protocol, solving problems currently addressed by Java Messaging Service, MQ, Tibco RV, etc. Financial companies such infrastructure to distribute massive amounts of market data (although the protocol can be used for effeciently distributing anything)

1.4 STOMP Client

The Streaming Text Orientated Messaging Protocol (http://stomp.codehaus.org/) is a messaging protocol supported by a good number of languages and message brokers. While not high-performance like AMQP it is apparently near trivial to implement and interact with.

1.5 Darcs Bindings

Many darcs users have requested an interface to darcs as a C library. Creating this library would involve deciding on the details of the API and generating the bindings using the Haskell FFI. Designing the API is mostly done for you by following the structure of the darcs source code and how the existing commands interface with the core of darcs.

1.6 OpenID Relying Party (Consumer) library

A library allowing (web) applications to authenticate users with an OpenID would be great. WikiPedia has a good description of what is expected from the relying party.

2 Could be improved

Problems with existing solutions, that could be improved.

  • Harpy lacks support for x86-64 - supporting both x86 and x86-64 would be a boon.
  • The core HTTP libraries need a lot of work. There is the beginnings of a useful library there, but many http codes are not handled, and there should be simple 'get' and 'head' functionality.
  • Binary serialisation support should be ported to bytestring, and added to the extralibs bundle
  • Support for wider Word values in bytestring.
  • Adding bytestring support to Parsec.
  • Simple process forking/popen, of type
    String -> IO String
    , is needed (the implementation should be in base)

If you have code that improves on something in the base libraries, you might consider submitting it to the libraries process, so it can appear in the standard libraries.

3 Existing libraries (by problem domain)

See the full libraries list.

3.1 GUIs

See GUI libraries

3.2 High performance string IO

See String libraries

3.3 Binary IO

See Binary IO

3.4 Compression

See Compressing data

3.5 Encryption

See Crypto libraries

3.6 Web frameworks

See Web frameworks

3.7 XML processing

See XML handling

3.8 Database access

See Databases

4 Libraries to package

Several cool libraries are out there, not yet in hackage / cabal format. Some on the hitlist , that would be great to package up would be:

5 C libraries

C libraries are often very popular, and cheap to bind to. A good task would be to just write bindings to C libraries, sorted by popularity.

Possible library rankings are provided by: