Changeset 1153


Ignore:
Timestamp:
Mar 29, 2005, 12:12:10 PM (20 years ago)
Author:
duncan
Message:

more issues ..

File:
1 edited

Legend:

Unmodified
Added
Removed
  • inundation/ga/storm_surge/pmesh/documentation/issues.txt

    r1126 r1153  
    22
    33OPEN - issues that have not beeen resolved
     4Issue: If some triangles have tags and others don't then the data
     5structure (a list of lists) is dificult to use in Numeric.
     6 eg.
     7SyntaxError: invalid syntax
     8>>> a = array(['a','b']); print a
     9[a b]
     10>>> a = array(['a','b','']); print a
     11Traceback (most recent call last):
     12  File "<stdin>", line 1, in ?
     13TypeError: an integer is required
     14>>> a = array([['a'],['b'],['']]); print a
     15Traceback (most recent call last):
     16  File "<stdin>", line 1, in ?
     17TypeError: an integer is required
     18>>> a = array([['a'],['b'],['']]).astype(ListType); print a
     19Traceback (most recent call last):
     20  File "<stdin>", line 1, in ?
     21TypeError: an integer is required
     22>>> a = array([['a'],['b'],['c']]).astype(ListType); print a
     23Traceback (most recent call last):
     24  File "<stdin>", line 1, in ?
     25TypeError: an integer is required
     26>>> a = array(['a','b','']); print a
     27Traceback (most recent call last):
     28  File "<stdin>", line 1, in ?
     29TypeError: an integer is required
     30>>> a = array(['a','b','c']); print a
     31[a b c]
     32>>> a = array([['a'],['b'],['c']]); print a
     33Traceback (most recent call last):
     34  File "<stdin>", line 1, in ?
     35TypeError: an integer is required
     36>>> a = array([['a'],['a'],['a']]); print a
     37Traceback (most recent call last):
     38  File "<stdin>", line 1, in ?
     39TypeError: an integer is required
     40>>> a = array([['a'],['a'],['a']]); print a
     41Traceback (most recent call last):
     42  File "<stdin>", line 1, in ?
     43TypeError: an integer is required
     44>>> a = array([['a'],['a'],['a']], Character); print a
     45[[  [a]]
     46 [  [a]]
     47 [  [a]]]
     48>>> a = array([['a'],['a'],['']], Character); print a
     49Traceback (most recent call last):
     50  File "<stdin>", line 1, in ?
     51TypeError: an integer is required
     52>>> a = array([['a'],['a'],['','a']], Character); print a
     53Traceback (most recent call last):
     54  File "<stdin>", line 1, in ?
     55TypeError: an integer is required
     56>>> a = array([['a','j'],['a','j'],['j','a']], Character); print a
     57Traceback (most recent call last):
     58  File "<stdin>", line 1, in ?
     59TypeError: an integer is required
     60>>> a = array(['a','a',''], Character); print a
     61[[a]
     62 [a]
     63 [ ]]
     64Solution -  use a list structure, instead of list of lists
     65 
     66
     67Issue: Red mesh selected for refinement seems to be 'behind' the green
     68mesh after selecting all then clearing the selection.
     69Importance: low
     70Suggested Action: Fixed BUT now it draws the greens behind the reds.
     71It is much less obvious now though.
     72
     73
     74Issue:  Ole thought that the regional mesh max areas were incorrect
     75when a .tsh file was loaded.  Maybe it was that the overall max area
     76is not saved.  It's hard coded in MeshGenDialog.
     77Importance: Medium
     78Suggested Action: Fix when .tsh goes NETCDF
     79
     80Issue: Catch error if .tsh file is a bad format.
     81suggestion: quick and dirty try - except
     82Importance: Medium
     83
     84Issue (new feature): Have visualisation of verts be turned on and
     85off.  This will speed things up.
     86Importance: low
     87
     88Issue (new feature): Hve .tsh as a netscf fie, instead of an ascii
     89file.
     90Importance: Medium
     91
     92Issue: When a file is saved as and then saved later, sometimes the
     93file name is not known.   Is it that 'clear' is removing the name
     94(which it should).
     95
     96Issue:A .tsh file loaded into pmesh after smoothing will not have the
     97outline info, since this is removed when smoothing.  If this is a
     98problem, refactor loadASCII so it can write the outline as well.
     99Suggested Action: Document use
     100
     101Issue:vertices at the bottom of the canvas can be hard to visualise,
     102since you don't know if all vertices are on the screen. 
     103Suggested Action: have a 5% border
     104Importance: low to not do
     105
     106Issue: The mesh produced by triangle isn't of high quality near a low
     107angle formed by two user segments.
     108Reason:triangle
     109Tried:  Tried the triangle flags L,i,V (for info).  Using box.poly on
     110nautilus.  Couldn't get a better quality mesh.
     111
     112Issue: Get the naming right for mesh/triangulation/user/generated.
     113How about "outline + triangulation = mesh"
     114Importance: Medium -low
     115
     116Issue: S/W - too many methods with the same name on differnent classes
     117eg editWindow.  Makes things confusing.
     118Importance: low
     119
     120Issue: Normalise seems broken - what is it doing to the attributes?. 
     121Suggested Action: remove!
     122
     123Bug: Calling 'Generate Mesh' on empty document or If there are no
     124enclosed regions will cause trianglulation to crash pmesh.
     125Importance: low - not do, but document
     126Suggested Action: Catch them (with try) and do nothing
     127notes:   Fixing the 'Generate Mesh' crash is a P3 requirement.
     128
     129The zoom in/out is jittery
     130
     131The initial mesh window coordinates are in the fourth quadrant.
     132
     133Deleting a selected object is done with the scroll bar, not the
     134delete key.
     135
     136Issue: triangle calculates incorrect attribute values, some of the time.
     137Importance: Medium
     138Suggested Action: Preprocess data to smoth it out.
     139
     140Issue: pmesh can not be automatically installed in windows and linux,
     141triangle doesn't compile easily.
     142Installs best in Linux. 
     143Importance: Low
     144Suggested Action: delve into setup more so it works with a variety of
     145compilers.
     146
     147Issue: two vertices can be ontop of one another
     148Importance: Low
     149Suggested Action: See how it goes when used.  The cost of checking
     150probably doesn't outwiegh the benifit.
     151action taken: When loading an xya file duplicate vertices are removed.
     152
     153
     154Issue: When calculating mesh, have the pointer turn into an hour glass
     155Importance: Low
     156
     157****************
     158CLOSED - issues that have  beeen resolved
     159************
    4160
    5161Issue (Ole): I couldn't get pmesh to write mesh in the .msh format
     
    9165    2:  Allow pmesh to store as msh
    10166    3:  Have a tsh2msh converter
    11    
    12 
    13 Issue: Red mesh selected for refinement seems to be 'behind' the green
    14 mesh after selecting all then clearing the selection.
    15 Importance: low
    16 Suggested Action: Fixed BUT now it draws the greens behind the reds.
    17 It is much less obvious now though.
    18 
    19 
    20 Issue:  Ole thought that the regional mesh max areas were incorrect
    21 when a .tsh file was loaded.  Maybe it was that the overall max area
    22 is not saved.  It's hard coded in MeshGenDialog.
    23 Importance: Medium
    24 Suggested Action: Fix when .tsh goes NETCDF
    25 
    26 Issue: Catch error if .tsh file is a bad format.
    27 suggestion: quick and dirty try - except
    28 Importance: Medium
    29 
    30 Issue (new feature): Have visualisation of verts be turned on and
    31 off.  This will speed things up.
    32 Importance: low
    33 
    34 Issue (new feature): Hve .tsh as a netscf fie, instead of an ascii
    35 file.
    36 Importance: Medium
    37 
    38 Issue: When a file is saved as and then saved later, sometimes the
    39 file name is not known.   Is it that 'clear' is removing the name
    40 (which it should).
    41 
    42 Issue:A .tsh file loaded into pmesh after smoothing will not have the
    43 outline info, since this is removed when smoothing.  If this is a
    44 problem, refactor loadASCII so it can write the outline as well.
    45 Suggested Action: Document use
    46 
    47 Issue:vertices at the bottom of the canvas can be hard to visualise,
    48 since you don't know if all vertices are on the screen. 
    49 Suggested Action: have a 5% border
    50 Importance: low to not do
    51 
    52 Issue: The mesh produced by triangle isn't of high quality near a low
    53 angle formed by two user segments.
    54 Reason:triangle
    55 Tried:  Tried the triangle flags L,i,V (for info).  Using box.poly on
    56 nautilus.  Couldn't get a better quality mesh.
    57 
    58 Issue: Get the naming right for mesh/triangulation/user/generated.
    59 How about "outline + triangulation = mesh"
    60 Importance: Medium -low
    61 
    62 Issue: S/W - too many methods with the same name on differnent classes
    63 eg editWindow.  Makes things confusing.
    64 Importance: low
    65 
    66 Issue: Normalise seems broken - what is it doing to the attributes?. 
    67 Suggested Action: remove!
    68 
    69 Bug: Calling 'Generate Mesh' on empty document or If there are no
    70 enclosed regions will cause trianglulation to crash pmesh.
    71 Importance: low - not do, but document
    72 Suggested Action: Catch them (with try) and do nothing
    73 notes:   Fixing the 'Generate Mesh' crash is a P3 requirement.
    74 
    75 The zoom in/out is jittery
    76 
    77 The initial mesh window coordinates are in the fourth quadrant.
    78 
    79 Deleting a selected object is done with the scroll bar, not the
    80 delete key.
    81 
    82 Issue: triangle calculates incorrect attribute values, some of the time.
    83 Importance: Medium
    84 Suggested Action: Preprocess data to smoth it out.
    85 
    86 Issue: pmesh can not be automatically installed in windows and linux,
    87 triangle doesn't compile easily.
    88 Installs best in Linux. 
    89 Importance: Low
    90 Suggested Action: delve into setup more so it works with a variety of
    91 compilers.
    92 
    93 Issue: two vertices can be ontop of one another
    94 Importance: Low
    95 Suggested Action: See how it goes when used.  The cost of checking
    96 probably doesn't outwiegh the benifit.
    97 action taken: When loading an xya file duplicate vertices are removed.
    98 
    99 
    100 Issue: When calculating mesh, have the pointer turn into an hour glass
    101 Importance: Low
    102 
    103 ****************
    104 CLOSED - issues that have  beeen resolved
    105 ************
    106 
     167Closed: Couldn't be reproduced. suggested action 2 has already been
     168implemented.  More investigation required before calling an issue an issue..
    107169
    108170Issue: After selecting all for refinement then clearing the selection
Note: See TracChangeset for help on using the changeset viewer.