TODO

unclassified:

- introduce cancellation everywhere (some progress made, what is wanted in the
  prefix case ?)
last update: 3 June 2007
-> cheak after release

SHORT TERM:
- restructure functions and file names. Remove empty placeholders.
- allow to break a grobnerloop, so that G and todo are produced with which only
  the last iteration/phase has to be redone
- allow to automatically break a computation after (time or number of runs or
  .. ?)
- try to speed up reducepol2 (so computations can be continued with long
  partial grobner bases): remove double loop and occur
- functions to check whether some G is a Grobner basis, or G and todo are a
  partial Grobner basis.
  -> only partially done
- avoid confusion with gbnp_doc.xml.in and gbnp_doc.xml by creating the
  gnbp_doc.xml in a different directory

-explain difference between LoadPackage("GBNP")
 and LoadPackage("GBNP","0",false).

+decide to establish GBR.pg as a field. Decision NOT PUBLIC

-document argument Num??Gens better

-choose for BaseQAPTS etc,. to have  four arguments; CHECK uitwerking en
 documentatie

-similar for DimQAPTS

-define Strong Grobner Basis

- In introduction to examples, specify version of FreeBSD and GAP version.

- Functions DetermineGrowth and DimsQATrunc print more than their output.
  Cheap option: Discuss this in explanatory sections (3.6.2 and 3.8.5).
  Posher option: turn output into a record containing the info printed.

-remove the word "Obstructions" from public part of Krook's code.

-"Timing info will be printed as well" bij level 2 van InfoGBNP. Hoe verhoudt
   zich dat tot InfoGBNPTime?

- Combine DetermineGrowth and DetermineGrowthObs

- Adjust output of DetermineGrowth and DetermineGrowthObs etc
 so that in similar cases similar output is given ("finite" vs "0" offends
 this rule from  DetermineGrowth and DetermineGrowthObs, respectively)

- ngens.g verwerken tot basis functies die min aantal variabelen
  en minimale moduulrang opgeven. Beantwoord:
  Read off the number of variables a list of polynomials needs to be
  defined. This would enable a default option (without the number t of
  generators of the free algebra) in most cases. Exception: when you have
  a case of polynomials in say t variables generating an ideal in the free
  algebra with more than t generators.
  NumAlgGensNP, etc.

+ rename
+        occurtree PTS ? -> no rename in this version
+        DONE: SGrobnerPTS -> SGrobnerModule
+        DONE: QAPTS -> QM
+        DONE: PTSNP -> NPM
+        DONE: LTermsNP -> LMonsNP
+        DONE: HilbertSeries -> HilbertSeriesQA
+        DONE: DetermineGrowth -> DetermineGrowthQA

+ check hilbertseries vs determinegrowth for [1,2,1], [2,2,1]
  -> polynomial of degree 2, that is correct (total vs ingremental)

- StrongNormalFormNPTrunc: voorbeeld: gebruik
  gewone StrongNormalFormNP ***Let op bovengrens graad***

LONG TERM:

- USES/USED IN are not up to date
- check input with more than one polys with same lt
- reduce PTS routines
- merge SGrobner variants (some progress made)
- SGrobnerLoop: ObsT(all) sorts todo, move this to the end of the loop in
  SGrobnerloop, combine this with a fase IIId that reduces todo without
  costly removing/inserting but by using a new StrongNormalForm variant
- merge ReducePol (some progress made)
- tree merge + trace inclusion in "all"
- 3-HEAP
- reduction of the length of traces
- GBNP.StrongNormalForm2TS all version?
- naming: StrongNormal versus SGrobner
- Add original relations to the Gröbner basis trace record ?
- Calculate the StrongNormalFormTraceDiff with respect to the Gröbner basis
  instead of with respect to the original relations
- It might be interesting to count monomials in the trace in example 24
- put the date in a single file (the GNUMakefile ?)
- add a link in the manual to the other available manual formats