News in 0.4.1
-------------

* Building

  Builds cleanly on more platforms, including SunOS and old Redhat setups.


News in 0.4.0
-------------

* Cryptography

  New generic crypto subsystem, supporting ElGamal (public-key) and
  Twofish (symmetric) ciphers.


* FIFO buffers

  Support for transparent symmetric en/decryption of streams.


* Sockets

  Support for transparent key exchange and symmetric en/decryption of
  streams.


* Communications

  Support for transparent key exchange and symmetric en/decryption of
  streams.


* Network Virtual Terminal Protocol (NVTP)

  New subsystem, based on sockets. Simplifies the use of line-based network
  protocols, like FTP, HTTP, SMTP, POP3, etc.

  Supports transparent key exchange and symmetric en/decryption of
  streams.


* Simple Mail Transfer Protocol (SMTP)

  New subsystem, based on NVTP. Simplifies communication with SMTP servers
  and clients. Only works as a server in this release.

  Supports transparent key exchange and symmetric en/decryption of
  streams.


* Memory management

  New block allocator with vast speed benefits.


* Process management

  New function lets you spawn a child process with a socket connected to its
  standard input/output.


* Token trees

  Faster and more economical, especially for small data sizes.

  Uses block allocator for memory management.


* Balanced binary trees

  Allow you to keep and search among a huge number of key/data associations
  very quickly and economically. Scales well.


* Proxies

  Support for the new NVTP and SMTP subsystems.


* Other stuff

  Lots of new tests and benchmarks.

  Three examples.

  Quick-start build framework for applications based on Flux.

  Redesigned HTML pages, including some new documentation.



News in 0.3.0
-------------

* Token trees

  Recoded from scratch. Fundamental type changed from TTREE to TT. More
  object oriented design allows for backwards-compatibility even when
  huge changes are made internally.

  Ability to transparently use filesystem as data storage for large
  nodes.

  Faster node insertion. ([1] as opposed to [N], where N is the number of
  siblings on the branch).

  Hashing algorithm is RIPE-MD160.

  Lots of convenience functions.


* Markup trees

  Specializing in semantic markup representations, like XML.

  Based on token trees.


* FIFO buffers

  Major bugfixes.


* Sockets

  Re-implemented using FIFO buffers.


* Communications

  Now allow breadth-first serialization.

  Queues.


* Proxies

  Allow you to do non-blocking processing of multiple simultaneous tasks.


* Token nets

  Like token trees, but with network structure.

  Very basic.


* Documentation

  1/3 of the references done, in HTML.


* Other stuff

  Runtime test suite.

  Beautification tools.

  Simple HTML doc skeleton generator.
