This is the list of the planned futures developments. Presented here by
decreasing order of likelyness (It will depend on popular demand :-) 

- A 2D multi-channel extension of the MaxEnt module is in progress.
  This will permits multiple J-deconvolution in 2D.

- the internal peak table will be moved to an external dbm file.
  This will permits much more more freedom in defining new entries.
  Will also merge with the assignment code which is to come.
  A hash table will be added to rapidly find a peak from its
  spectral coordinates

- a 4D mode will be able once, (probably as soon as I manage to run one 
  on my spectrometer ;-)

- The size of the main memory should be configurable from the comman line
  thus having only one binary file to handle around.

- you will be able to return a value from a macro
  (with a RETURN (val) command, that will set a $returned value in the caller)

- The look mode in the cache_mad library will be developed some day.

- We plan to create interfaces between Gifa and "usefull" academic program
  (Crowd, Rasmol, XPLOR, etc...)

- The fitter will become 'generalized' and you will be able to fit
  any user-provided function onto your data.

- The generated postscript might even one day become 'conformant'

- The graphic might one day change to a display-list based one, to
  permit multiple display on-screen (kindof subplot() in matlab).
  Will permit multiple zooms (for instance to show a spin-system)

- EM and GM might change their syntax to fit with SIN