Things to do for the next release of ANSIG

• Check and clean up the examples.
• Document the curve fitting AL intrinsics.
• Document the AL data types.
• Document the spectrum folding and aliasing handling properly.
• Write AL procedures ass_15n_table.al, ass_15n_13c_table.al.
• Non-sequence names should be storable, and entries made for them in the assignments table. The trick is to record how many crosspeaks contribute to each assignment; when the crosspeaks are deassigned, then the entry can be removed when the counter reaches zero. This will make it possible to keep track of NOEs involving as yet unassigned sequence residues. What to call an entry in a sequence field that is not a member of the sequence? Set, Group, Name, Identifier, Unknown...
• Change CROSSPEAKS MOVE to move crosspeaks that have assignments and connections; in order to get rid of bug that shows up occasionally in AL procedure cp_move. This will require checks after the move to see that no inconsistencies are introduced.
• Fix the sec_struc AL procedure and header file so that larger proteins are handled properly.
• Matrix_statistics: output: min, max, stdev; input: sp, lodp, hidp. Matrix_histogram: output integer list; input: # bins, min, max, lodp, hidp.
• AL procedures to check assignments, find off-position crosspeaks, etc.
• Write AL procedure and macro to change contour levels easily.
• Write AL procedures that suggest crosspeaks that have not been assigned based on coordinates. Either for ambiguous crosspeaks, or 'ab initio'. How to do this for non-stereospecifically assigned nuclei ?
• Create a dictionary for DIANA-type designations.
• AL procedures to convert XPLOR assignments from/to DIANA.
• Check that DIANA NOE output actually works.
• When reading lists from symbol file, ignore elements which cannot be interpreted, and continue reading the file ?
• Allow reading more than one dictionary file. This would simplify adding non-standard residues.
• > GRAPHICS MARKER number: set number of markers used; circular buffer.
• Allow definition of which spectra to be displayed via explicit commands instead of order in spd. Requires a lookup table in iagrph.f. For simplicity, do this before opening graphics, not allowing changes thereafter. > SET GRAPHICS SPECTRUM spectrum slot.
• Allow 'GRAPHICS ATTACH' also for non-plane dimensions. I.e. allow explicit mapping of scroll-bars. How to deal with that annoying aspect-ratio problem ?
• Implement better integration methods: Integrate_fit (Denk's at al. method of comparing with resolved peaks) Create data base of peak lineshapes in F dimensions; by default a Gaussian of 'linewidth' width, but should be modifiable with the help of AL procedures. One data base file per spectrum. How should this behave close to diagonal ? Maybe use masks to avoid problem areas? How to behave when too ill-defined because crosspeaks are too close?
• Develop and/or include published peak-pickers for 3D/4D spectra. How to avoid rubbish, and still being sensitive?