FIXME-TODO
changeset 866 f537d570fff8
parent 794 f0a78fda343f
child 912 aa960d16570f
equal deleted inserted replaced
865:5c6d76c3ba5c 866:f537d570fff8
     1 Highest Priority
     1 Highest Priority
     2 ================
     2 ================
     3 
     3 
     4 - better lookup mechanism for quotient definition
     4 - better lookup mechanism for quotient definition
     5   (see fixme in quotient_term.ML)
     5   (see fixme in quotient_term.ML)
     6 
       
     7 - check needs to be fixed in mk_resp_arg (quotient_term.ML),
       
     8   see test for (_, Const _)
       
     9 
       
    10 - clever code in quotiet_tacs.ML needs to be turned into
       
    11   boring code
       
    12 
       
    13 - comment about regularize tactic needs to be adapted
       
    14 
       
    15 - Check all the places where we do "handle _"
       
    16   (They make code changes very difficult: I sat 1/2
       
    17   a day over a simplification of calculate_inst before
       
    18   I understood things; the reason was that my exceptions
       
    19   where caught by the catch all. There is no problem
       
    20   with throwing and handling exceptions...it just must
       
    21   be clear who throws what and what is thrown.)
       
    22 
       
    23 
     6 
    24 Higher Priority
     7 Higher Priority
    25 ===============
     8 ===============
    26 
     9 
    27 - If the user defines twice the same quotient constant,
    10 - If the user defines twice the same quotient constant,
    32   in quotient_def raises a bind exception.
    15   in quotient_def raises a bind exception.
    33 
    16 
    34 - If the constant definition gives the wrong definition
    17 - If the constant definition gives the wrong definition
    35   term, one gets a cryptic message about absrep_fun
    18   term, one gets a cryptic message about absrep_fun
    36 
    19 
    37 - have FSet.thy to have a simple infrastructure for 
       
    38   finite sets (syntax should be \<lbrace> \<rbrace>,
       
    39   look at Set.thy how syntax is been introduced)
       
    40 
       
    41 - Handle theorems that include Ball/Bex. For this, would 
    20 - Handle theorems that include Ball/Bex. For this, would 
    42   it help if we introduced separate Bex and Ball constants 
    21   it help if we introduced separate Bex and Ball constants 
    43   for quotienting?
    22   for quotienting?
    44 
    23 
    45 - The user should be able to give quotient_respects and 
    24 - The user should be able to give quotient_respects and 
    46   preserves theorems in a more natural form.
    25   preserves theorems in a more natural form.
    47 
    26 
    48 
    27 
    49 Lower Priority
    28 Lower Priority
    50 ==============
    29 ==============
    51 
       
    52 - Maybe a quotient_definition should already require
       
    53   a proof of the respectfulness (in this way one
       
    54   already excludes non-sensical definitions)
       
    55 
    30 
    56 - accept partial equivalence relations
    31 - accept partial equivalence relations
    57 
    32 
    58 - think about what happens if things go wrong (like
    33 - think about what happens if things go wrong (like
    59   theorem cannot be lifted) / proper diagnostic 
    34   theorem cannot be lifted) / proper diagnostic