MIDAS / EXSAS Bug Report:


>From:            izzo Tue Jun 11 18:12:00 1996
>Originator:      Carlo Izzo

>Synopsis:        In logaritmic plots the bins having value 0 and 1 are 
                  plotted at the same level - Command: PLOT/HISTOGRAM
>Severity:        serious
>Priority:        high
>Category:        midas-graphic
>Class:           change-request
>Release:         95NOVpl2.0

>Environment:     System: OSF1 o09 V3.2 148 alpha
>Machine:         alpha
Description: In logaritmic plots the bins having value 0 and 1 are plotted at the same level (i.e., zero). It is important to differentiate between these values. This report was already sent on March 1995 (see PR/518)

STATUS: Fixed


>From             izzo Tue May 21 14:59:46 1996
>Originator:      Steve Weimer

>Synopsis:        Some necessary function is missing... 
                  - Command: EXTRACT/SLIT
>Severity:        serious
>Priority:        high
>Category:        midas-image
>Class:           support
>Release:         95NOVpl2.0

>Environment:     System: OSF1 o09 V3.2 148 alpha
>Machine:         alpha
Description: I have had no trouble extracting the profiles and plotting them (average/image + plot/row) however, since the results of the project will likely be shared it is necessary to display the location of the slits on the image and identify them with their corresponding plots. True, I could print out each one separately along with the plots, but we are looking at possible 100 slits around the [area of work]. The command extract/slit erases the overlay channel before allowing me to draw the next slit. Is there a way to preserve the overlay channel with each slit displayed? We have thought that perhaps a line could be drawn to represent the location, but it would be nice to show the outline of the slit. Is there a way to get the coordinates of the corners of the slit? Perhaps then a box could be drawn afterword, perhaps with draw/line. The extract/slit command displays the coordinates of the cursor, but when the slit is drawn it is positioned midway between the cursors. I suppose the cursors could be placed so that they are at the ends of the slits, but this might be difficult when the slit is angled. Is there a better (and easier) way?

The goal would be to have a picture of the xray image with all of the slits drawn on them and with some identification to each of their corresponding plots. Does this go beyond the capabilities of MIDAS/EXSAS ?

STATUS: Only partially solved


>From             izzo Mon May  6 17:25:42 1996
>Originator:      Carlo Izzo

>Synopsis:        Character array descriptors are not copied correctly 
                  - Command: COPY/DD
>Severity:        serious
>Priority:        high
>Category:        midas-system
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     System: OSF1 o09 V3.2 148 alpha
>Machine:         alpha
Description: Take two tables, table_a.tbl and table_b.tbl

Create a character array of 3 elements, 64 characters each, in table_a:

write/descr table_a.tbl remarks/c*64/1/3 " "

The descriptor is written correctly:

show/descr table_a.tbl remarks

frame: table_a.tbl (data = R4)

name: REMARKS (...)

type: character*64 no. of elements: 3

Now copy this descriptor to table_b:

copy/dd table_a.tbl remarks table_b.tbl

The output descriptor is an array of 192 elements (i.e. 64*3) of 64 characters each.

show/descr table_b.tbl remarks

frame: table_b.tbl (data = R4)

name: REMARKS (...)

type: character*64 no. of elements: 192

STATUS: Fixed


>From             izzo Fri Apr 26 17:06:26 1996
>Originator:      Carlo Izzo

>Synopsis:        Inconsistency - Command: STFINF
>Severity:        serious
>Priority:        high
>Category:        midas-system
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     System: OSF1 o09 V3.2 148 alpha
>Machine:         alpha
Description: In a FORTRAN program it is checked first, whether the file does exist and then whether it is already open

      INTEGER        IBUF(5),ISTAT,TID1,IOPEN1
      CHARACTER*128  OUTSTR,TNAME1
C
      CALL STFINF(TNAME1,99,IBUF,ISTAT)
      IF(ISTAT.NE.0) THEN
        WRITE(OUTSTR,211) TNAME1
  211   FORMAT(1X,'not existing file ',A)
        CALL STTPUT(OUTSTR,ISTAT)
      ENDIF
      CALL STFINF(TNAME1,0,IBUF,ISTAT)
      IF(ISTAT.EQ.0) THEN
        TID1=IBUF(1)
        IOPEN1=1
      ENDIF
With a non-existing file the first call works fine, but the second call of STFINF has then a return status of 0, which is obviously wrong. If we skip the first call to STFINF the second call returns the correct answer. Could it be that the internal initialization during the second call is not properly performed ?

STATUS: Fixed


>From:            izzo Mon Apr  1 13:47:51 1996
>Originator:      Carlo Izzo

>Synopsis:        Printer assignment lost when printing fails 
                  - Command: PRINT/TABLE, ASSIGN/PRINT
>Severity:        serious
>Priority:        medium
>Category:        midas-system
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     System: OSF1 o09 V3.2 148 alpha
>Machine:         alpha
Description: If the PRINT/TABLE command fails for some reason (e.g., the table to print doesn't exist) MIDAS correctly answers (e.g., that there is no such file or directory). But this destroyes the print-assignation. Any following print/tab-command with an existing table results in the message:

lpr: midtemp10.print: unknown printer

One has to reassign the printer with assign/print to proceed.

STATUS: Fixed


>From:            izzo Thu Mar 28 09:07:47 1996
>Originator:      Carlo Izzo

>Synopsis:        Bad termination in case of error - Command: PLOT/HISTO
>Severity:        serious
>Priority:        low
>Category:        midas-graphic
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     System: OSF1 o09 V3.2 148 alpha
>Machine:         alpha

>Description:
Take a MIDAS-table for example table.tbl an select it so that the selected
table is empty (for example: sel/tab table seq.lt.1).
Then type plot/histo table :column , where :column is any existing column
in table.tbl. At first you obtain the message "No selected entries or
empty column", which is quite correct. But then the plot/histo-program
crashes with the message:

  OSY_SPAWN: Child killed by signal 11
  Could not execute /usr/util/midas/95NOV/prim/exec/plothtbl.exe

If one wants to use this command in an automatic procedure the procedure
is terminated, which is a bit inconvenient. Of course one could check
before whether the table is empty or not, but plot/histo has already
recognized this, why can't it terminate proper then?

STATUS: Fixed

------------------------------------------------------------------------------

>From izzo Mon Mar 25 18:28:40 1996
>Originator:      Carlo Izzo

>Synopsis:        Wrong descriptor TSELTABL in output table 
                  - Command: SELECT/TABLE, COPY/TABLE
>Severity:        serious
>Priority:        high
>Category:        midas-table
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     
System: OSF1 o09 V3.2 148 alpha
Machine: alpha
>Description:
I have a table called "in" and I apply a selection:

sel/tab in seq.gt.1.and.seq.lt.13
copy/tab in out

The output table, consisting of 11 columns, still has the descriptor 
TSELTABL filled with 

	seq.gt.1.and.seq.lt.13

and this information is displayed also with the command SHOW/TABLE.
This is misleading even if, apparently, the selection is not effective
in the output table (it is enough to do a READ/TABLE to see this).

STATUS: Fixed

-------------------------------------------------------------------------------

>From izzo Mon Mar 25 11:22:26 1996
>Originator:      Carlo Izzo

>Synopsis:        The string "()" is written as "(0)" 
                  - Command: KEYWORD = "STRING"
>Severity:        serious
>Priority:        low
>Category:        midas-system
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     
System: OSF1 o09 V3.2 148 alpha
Machine: alpha
>Description:

Try the following:

outputc = " "
outputc = "()"
read/key outputc

The ontent of outputc is now (0).

The command

write/key outputc/c/1/2 "()"

works correctly instead.

STATUS: Fixed

---------------------------------------------------------------------------

>From izzo Fri Mar 22 16:39:41 1996
>Originator:      Carlo Izzo

>Synopsis:        if columns RA and DEC already present, wrong output 
                  - Command: PRECESS/COORDINATES
>Severity:        serious
>Priority:        medium
>Category:        EXSAS
>Class:           sw-bug
>Release:         95NOVpl2.0

>Environment:     
System: OSF1 o09 V3.2 148 alpha
Machine: alpha
>Description:
I have created the following test table:

create/table test 2 10 null
create/column test :ra0
create/column test :dec0
compute/table test :ra0 = seq
compute/table test :dec0 = seq
preces/coor test 1950 2000

At this point the conversion is fine, and the new columns :ra and :dec 
are created.

Now, if I repeat

preces/coor test 1950 2000

:ra and :dec are NOT overwritten (as it is stated in the help), but :ra is
destroyed and two :dec exist .

STATUS: Ask Stefan...

------------------------------------------------------------------------------

>From xsun Mon Feb 19 11:46:44 1996
>Originator:      Xuejun Sun

>Synopsis:        NaN rotation angle in output - Command: ALIGN/IMAGE
>Severity:        serious
>Priority:        medium
>Category:        midas-image
>Class:           sw-bug
>Release:         95NOVpl1.0

>Environment:     
System: OSF1 o07 V3.2 148 alpha
Machine: alpha
>Description:
I have two tables with coordinates.  The first is table abc.tbl

 Sequence X_SKY     Y_SKY     IDENT
 -------- --------- --------- -----------
        1   -888.50  -2150.90           1
        2    640.30  -2150.90           2
        3    -98.90  -2304.50           3
        4  -1024.50    327.00           4
        5   1155.00      1.00           5
        6   1837.50   -542.00           6
        7   1526.00   2133.50           7
        8   -743.00    220.00           8
        9    137.80  -2356.00           9
       10    532.00   -462.00          10
       11    752.00   -625.00          11
       12  -1467.00  -1546.00          12
       13    572.00  -1546.00          13
       14  -1254.00    145.00          14
       15    -57.00   1514.00          15
       16     12.00   -134.00          16
       17    982.50    146.00          17
       18   1166.00   -634.00          18
 -------- --------- --------- -----------

The second table is def.tbl

 Sequence X_SKY           Y_SKY           IDENT
 -------- --------------- --------------- -----------
        1       -603.1580        285.4297           1
        2        197.7679       1586.9197           2
        3        -58.4966        877.4666           3
        4      -2784.8696       1468.2953           4
        5      -1365.2421       3153.1504           5
        6       -545.0750       3449.7898           6
        7      -2987.0117       4586.7769           7
        8      -2546.2441       1651.9287           8
        9        108.9568       1051.4546           9
       10      -1297.3107       2379.9978          10
       11      -1043.3289       2481.8950          11
       12      -1421.5696        109.7528          12
       13       -353.1594       1845.9406          13
       14      -2750.1882       1177.3655          14
       15      -3288.6675       2914.2986          15
       16      -1849.2178       2108.9766          16
       17      -1579.0734       3082.2595          17
       18       -818.6022       2829.7200          18
 -------- --------------- --------------- -----------
 
Finally I use the command

align/ima abc,:X_SKY,:Y_SKY def,:X_SKY,:Y_SKY E

and the output rotation angle is NaN

STATUS: Fixed

------------------------------------------------------------------------------

>From ajv Wed Jan 24 11:20:23 1996
>Originator:      Andreas Vogler

>Synopsis:         - Command: REBIN/ROTATE
>Severity:        serious
>Priority:        medium
>Category:        midas-image
>Class:           support
>Release:         95NOVpl1.0

>Environment:     
System: OSF1 o04 V3.2 148 alpha
Machine: alpha
>Description:
Does not work if the rotation angle is 0 or very small.

E.g.  rebin/rot  a40 a40_2  0.00001  crashes with the following error 
massage: OSY_SPAWN: Child killed by signal 11
Could not execute /usr/util/midas/95NOV/prim/exec/genzz1.exe

This problem only appears if the image is larger than 705x705 pixels

(the images werw extracted from the ESO digitized sky survey for fields 
greater than  20arcminx20arcmin; the above mentioned image a40 had 
1413x1413 pixels:

NAXIS:                    2
NPIX:                  1413      1413
START:                202.08307448147        -27.433089346243
STEP:               -0.47204372945220E-03    0.47205216579737E-03
IDENT:           dss10844
CUNIT:                           RA---TAN        DEC--TAN
LHCUTS:             642.0000       19445.00       642.0000       19445.00
 )

the file a40.bdf is available on the OSF cluster in /home/ajv

STATUS: Fixed

----------------------------------------------------------------------------

>From mjf Thu Jan 11 20:21:41 1996
>Originator:      Michael Freyberg

>Synopsis:        fails if there "are NO PHOTONS in this observation" 
                  - Command: INTAPE/RDF DISK RP150058 ANC,BAS
>Severity:        serious
>Priority:        high
>Category:        EXSAS
>Class:           sw-bug
>Release:         95NOVpl1.0

>Environment:     
System: OSF1 o02 V3.2 148 alpha
Machine: alpha
>Description:
intape/rdf DISK rp150058 anc,bas     -> leads to crash
intape/rdf DISK rp150058 anc           -> works fine 

reason: warning message:
***                                                  ***
*** ERROR: There are NO PHOTONS in this observation! ***
***                                                  ***

     Upgrading descriptor header with EXSAS descriptors...

(ERR) SCDWRx: events.tbl + TIM_SEL - INPINV
Invalid input arguments to MIDAS routine.


However, the ancillary file should be converted anyway, and it should be checked
whether the number of photons is non-zero and only THEN try to update the
TIM_SEL descriptor. Apparently, all other descriptors seem to be ok. As the number
of photons is checked anyway (see warning) it shouldn't be too difficult to add
some lines of code to avoid the program crash and to get the rest INTAPEd.
The reason WHY such a change is neccessary: it is really difficult to know in 
advance (here: without crash) HOW MANY photons an observation has, but trying
to INTAPE the BAS file leads to the crash without user's chance to decide how to
continue (if there aren't any photons I wouldn't do any spectra with this file but I
could still use the orbit and attitude - for some strange purposes).

STATUS: Fixed

--------------------------------------------------------------------------------

>From izzo Mon Sep 18 17:16:49 1995
>Originator:      Carlo Izzo

>Synopsis:        Wrong error bars when (linear) x range is <0 and Y is a log scale - Command: OVERPLOT/ERROR
>Severity:        critical
>Priority:        high
>Category:        midas-graphic
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
This is a quite interesting bug. A user wanted to plot a table with a linear X scale
and a logaritmic Y scale. Then he overplotted the error bars, and he found
out that they all went down to -infinity  (no, the errors were not compatible
with zero). And he gave me other examples of similar plots that worked. At the
end I realized that all the wrong plots were involving negative values for the
X scale. Eventually I was able to replicate the problem with a quite simple
data table...
>How-To-Repeat:
Consider the table:

  Table : test 

 Sequence X            Y            ERR         
 -------- ------------ ------------ ------------
        1  1.00000e+00  1.00000e+01  3.00000e+00
        2  2.00000e+00  1.00000e+01  3.00000e+00
        3  3.00000e+00  1.00000e+01  3.00000e+00
 -------- ------------ ------------ ------------

Now do:

plot/axes 0,4 0,2,-1 ? "X" "Y"
overplot/table test :X :Y        
overplot/error test :X :Y :ERR

Everything worked properly. Now let's shift the data to the negative X:

compute/table test :x = :x - 4
plot/axes -4,0 0,2,-1 ? "X" "Y"
overplot/table test :X :Y
overplot/error test :X :Y :ERR

Now the error bars are wrong.

STATUS: Fixed

-------------------------------------------------------------------------

>From izzo Wed Sep 13 14:16:27 1995
>Originator:      Carlo Izzo

>Synopsis:        ASCII files cannot be addressed using the form LOGICAL:filename - Command: COPY/LSDD (AND OTHER COMMANDS)
>Severity:        serious
>Priority:        high
>Category:        midas-system
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
The command COPY/LSDD requires the following parameters:

1) ASCII file name
2) input frame
3) output frame

Now, the files 2) and 3) can be given using the form LOGICAL:filename, where
LOGICAL is an environment variable in Unix or a logical in VMS. Not so for
the file 1). The reason of that is that 2) and 3) are accessed via the MIDAS
standard interfaces, that resolve the value of the LOGICAL variable before
to use it, while the file 1) is accessed without any interface, and the filename
remains unresolved (it is interpreted literally as "LOGICAL:filename").

This is a general problem in MIDAS, in the sense that it tends to leave the
user puzzled.

We propose to allow the user to access ASCII files using "logicals" in the
same way they are used to access MIDAS binary files.

STATUS: Fixed

-------------------------------------------------------------------------

>From mjf Fri Sep  1 15:28:54 1995
>Originator:      Michael Freyberg

>Synopsis:        PLOT/COLUMN and AVERAGE/COLUMN give different results - Command: AVERAGE/COLUMN, PLOT/COLUMN
>Severity:        critical
>Priority:        high
>Category:        midas
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System:  OSF 
Machine: alpha (o02)
>Description:
PLOT/COLUMN gives different results as AVERAGE/COLUMN.
The undocumented feature is a shift of 1 column which can be seen 
in the PSPC detector response matrix, which contains zero values 
for channels 1-7, channels 8 and above have non-vanishing values.
As the DRM has 256 columns accessing the 0th column or the 258th
column results in a warning message but not accessing the 257th column.
Moreover, comparing PLOT/COLUMN and AVERAGE/COLUMN for the
8th column gives surprising differences: while AVERAGE/COLUMN and
a subsequent PLOT (= PLOT/ROW) command displays the familiar values 
of the DRM the command PLOT/COLUMN is not willing to confirm this. 
Instead it insists on adding an extra column to satisfy the user's wishes.
>How-To-Repeat:
READ/DESCRIPTOR  EXSAS_CAL:drmpspc_ao1
!
plot/column EXSAS_CAL:drmpspc_ao1 0
plot/column EXSAS_CAL:drmpspc_ao1 1
plot/column EXSAS_CAL:drmpspc_ao1 8
plot/column EXSAS_CAL:drmpspc_ao1 9
plot/column EXSAS_CAL:drmpspc_ao1 256
plot/column EXSAS_CAL:drmpspc_ao1 257
plot/column EXSAS_CAL:drmpspc_ao1 258
!
plot/column EXSAS_CAL:drmpspc_ao1 8
average/column tmp = EXSAS_CAL:drmpspc_ao1 @8,@8 
plot tmp
average/column tmp = EXSAS_CAL:drmpspc_ao1 8,8
plot tmp
plot/column EXSAS_CAL:drmpspc_ao1 9
average/column tmp = EXSAS_CAL:drmpspc_ao1 @9,@9 
plot tmp

STATUS: Fixed

---------------------------------------------------------------------------

>From mjf Tue Aug  8 18:05:51 1995
>Originator:      Michael Freyberg

>Synopsis:        screen output cannot be turned off - Command: DRAW/SOURCE
>Severity:        non-critical
>Priority:        low
>Category:        midas-system
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Usually setting the keyword LOG(4) = 2 turns off any screen output.
LOG(4) = 0 enables screen output again. This does not work for the
command DRAW/SOURCE., which still gives the screen message
               Open source list file lslst1.tbl
I am using procedures wihich frequently draw selections of source
detection tables, unfortunately lots of unneccessary screen output is 
produced. 

>How-To-Repeat:
in principle like the following:
LOAD/IMAGE image1
HELP/KEY LOG
LOG(4) = 0
DRAW/SOURCE lslst1
LOG(4) = 2
DRAW/SOURCE lslst1

STATUS: Ask Stefan...

>From izzo Tue Aug  8 17:46:48 1995
>Originator:      Carlo Izzo

>Synopsis:        Cross larger and larger with increasing X position - Command: DRAW/CROSS 
>Severity:        serious
>Priority:        high
>Category:        midas
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
I have reported this problem earlier today, but the user was impatient so I had to solve it myself.

File: prim/display/src/draw.c

Line: 348

The code to fix  is:

               coord[0] = cpos[0];
               coord[1] = cpos[1];
               coord[2] = lenght;
               cpos[0] = coord[0] - coord[2];
               cpos[1] = coord[1] - coord[2];
               cpos[2] = coord[0] + coord[2];
               cpos[3] = coord[1] + coord[2];

I have made a "minimal" modification to

               coord[0] = cpos[0];
               coord[1] = cpos[1];
               coord[2] = lenght;
               cpos[2] = lenght;
               cpos[3] = 0;

The reason of this change is that next (line 390) the array coord is rewritten
using again the values of cpos[0-3]

The correct code fix would be probably to write (in the fix above) just

               cpos[2] = lenght;
               cpos[3] = 0;

(note that the only reason I initialize the value of cpos[3] is that it is used in the
call 

	coord[3] = CGN_NINT( cpos[3] );

line 393)

		Ciao,   Carlo

STATUS: Fixed

-------------------------------------------------------------------------------

>From izzo Tue Aug  8 14:25:11 1995
>Originator:      Carlo Izzo

>Synopsis:        The size of the cross (P3) is mistaken with the X  image coordinates - Command: DRAW/CROSS
>Severity:        critical
>Priority:        high
>Category:        midas-display
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Cross sizes increase with increasing X image (not world) coordinate.
>How-To-Repeat:
DRAW/CROSS

STATUS: Fixed

----------------------------------------------------------------------------

>From mjf Mon Aug  7 14:53:38 1995
>Originator:      Michael Freyberg

>Synopsis:        does not work for moderately large lists - Command: MERGE/SOURCE_TABLES
>Severity:        critical
>Priority:        high
>Category:        EXSAS
>Class:           sw-bug
>Release:         94NOVpl2.0

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
I have a "local" list (lslst1.tbl) with 245 sources and a "map" list (mslst1.tbl)
with 497 sources, the sum of 245+497 = 742 is below the given limit of 1000.
However, after the merging process lots of sources are missing in the merged
table (mplst.tbl, after selection :COFL.gt.0).  More precisely, the "bottom" sources
(i.e. highest values of :Y_SKY) are skipped. The result table has 1000 rows
though 742 should be sufficient, this points towards some internal problems
in the underlying programs. 

As up to 6 input lists are allowed I tried  to split the two input tables into 3 lists 
each (82,82,81 and 166,166,165 entries, resp.) but got slightly worse results 
(only 301 merged sources found instead of 303, about 360 expected).

Additionally, I couldn't find HOW the properties of the merged sources are 
determined, i.e. the (merged) source position, count rate etc.: are the 2 (or more)
input values weighted by likelihood or count rate or ... ? If I've missed the relevant
part in the documentation, could you please tell me where to get the information.

Finally, in the parameter file (task EXSM) I found the (extended) parameter 
FWHM_CAL_FILE with the default value (ROSAT,HRI) EXSAS_CAL:fwrohr.tbl.
Apparently this table consists of a single column (:INDEX) with 10 rows containing
the (real) numbers from 1 to 10. SHOW/TABLE gives 1 column, READ/DESCRIP
however shows that in the past more columns have existed. The same holds for
the ROSAT PSPC default table EXSAS_CAL:fwrops.tbl. According to description
in Appendix G (page G-2) it should contain the "FWHM of point spread function
of ROSAT XRT HRI" (or PSPC, respectively).

STATUS: Fixed

-------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Wed Aug  2 16:55:50 1995
>Originator:      Carlo Izzo

>Synopsis:        PMODE=0 doesn't work for contour plots - Command: PLOT/CONTOUR, SET/GRAP
>Severity:        serious
>Priority:        high
>Category:        midas-graphic
>Class:           sw-bug
>Release:         94NOVpl2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
After applying patch level 2, the pmode=0 doesn't remove the frame around
a contour plot, but removes just the remarks on top of the frame (frame: image1, 
ident: "blabla", etc).

Knowing that you are planning to add new pmodes we have tried negative pmodes
(up to -4) and positive ones (up to 5) but we could not remove the frame around the
data (the only way was to plot it with colour 0 and then overplot the contours in
black).
>How-To-Repeat:
set/grap pmode=0
plot/contour image [-100,100:100,-100]  etc...

STATUS: Fixed

------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Thu Jul 27 09:30:56 1995
>Originator:      Carlo Izzo

>Synopsis:        "Child killed by a signal" message - Command: CREATE/COLUMN
>Severity:        critical
>Priority:        high
>Category:        midas-table
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
We have the following table:

 Table : qsource                         [Transposed format]
 No.Columns :      74   No.Rows :       1 
 All.Columns:      71   All.Rows:       8         Sel.Rows:       1
 Sorted  by :Sequence  Reference:Sequence     
 Col.#   1:IND_SOL          Unit:                 Format:I7     I*4
 Col.#   2:IND_MPX          Unit:                 Format:I7     I*4
 Col.#   3:RA_2000          Unit:h:m:s            Format:A11    C*11
 Col.#   4:DEC_2000         Unit:d:m:s            Format:A10    C*10
 Col.#   5:X_IMA            Unit:ima_pix          Format:F5.1   R*4
 Col.#   6:E_X_IMA          Unit:ima_pix          Format:F7.1   R*4
 Col.#   7:Y_IMA            Unit:ima_pix          Format:F5.1   R*4
 Col.#   8:E_Y_IMA          Unit:ima_pix          Format:F7.1   R*4
 Col.#   9:CTS              Unit:1/sec            Format:E9.3   R*4
 Col.#  10:BG_CTS           Unit:1/(ima_pix-sec)  Format:E9.3   R*4
 Col.#  11:EXPOSURE         Unit:sec              Format:F10.2  R*4
 Col.#  12:E_CTS            Unit:1/sec            Format:F9.7   R*4
 Col.#  13:CELL             Unit:ima_pix          Format:I4     I*4
 Col.#  14:OFF_AX           Unit:arcmin           Format:F6.2   R*4
 Col.#  15:EXT              Unit:arcsec           Format:F6.2   R*4
 Col.#  16:EXT_ML           Unit:                 Format:F6.2   R*4
 Col.#  17:DIST_SCR         Unit:ima_pix          Format:F8.1   R*4
 Col.#  18:DIST_RIB         Unit:ima_pix          Format:F8.1   R*4
 Col.#  19:EXI_ML           Unit:                 Format:F6.1   R*4
 Col.#  20:EXI_MD           Unit:                 Format:I10    R*4
 Col.#  21:HR_1             Unit:                 Format:F9.2   R*4
 Col.#  22:HR_2             Unit:                 Format:F9.2   R*4
 Col.#  23:E_HR_1           Unit:                 Format:F9.2   R*4
 Col.#  24:E_HR_2           Unit:                 Format:F9.2   R*4
 Col.#  25:PRIO             Unit:                 Format:A6     C*6
 Col.#  26:CSQ_PL           Unit:1/DOF            Format:F6.2   R*4
 Col.#  27:CSQ_RS           Unit:1/DOF            Format:F6.2   R*4
 Col.#  28:VARIAB           Unit:                 Format:A8     C*8
 Col.#  29:F_CTS            Unit:                 Format:A5     C*5
 Col.#  30:F_A_E            Unit:                 Format:A5     C*5
 Col.#  31:F_BB             Unit:                 Format:A5     C*5
 Col.#  32:F_SPEC           Unit:                 Format:A6     C*6
 Col.#  33:F_HR             Unit:                 Format:A4     C*4
 Col.#  34:CHAN_L           Unit:                 Format:I6     I*4
 Col.#  35:CHAN_H           Unit:                 Format:I6     I*4
 Col.#  36:F_MLL            Unit:                 Format:A6     C*6
 Col.#  37:OBS_MOD          Unit:                 Format:I7     I*4
 Col.#  38:F_MPX            Unit:                 Format:A5     C*5
 Col.#  39:F_PL             Unit:                 Format:A4     C*4
 Col.#  40:F_RS             Unit:                 Format:A4     C*4
 Col.#  41:F_VAR            Unit:                 Format:A5     C*5
 Col.#  42:EXT_METHOD       Unit:                 Format:A8     C*8
 Col.#  43:CTS_METHOD       Unit:                 Format:I10    I*4
 Col.#  44:F_HR_1           Unit:                 Format:A6     C*6
 Col.#  45:F_HR_2           Unit:                 Format:A6     C*6
 Col.#  46:RADEXTR          Unit:arcsec           Format:F6.1   I*4
 Col.#  47:X_SKY            Unit:sky_pix          Format:F16.2  R*4
 Col.#  48:Y_SKY            Unit:sky_pix          Format:F16.2  R*4
 Col.#  49:DIST_RIB_NEW     Unit:ima_pix          Format:f8.1   R*4
 Col.#  50:a                Unit:Unitless         Format:A1     C*1
 Col.#  51:m                Unit:Unitless         Format:A1     C*1
 Col.#  52:d                Unit:Unitless         Format:A1     C*1
 Col.#  53:e                Unit:Unitless         Format:A1     C*1
 Col.#  54:n                Unit:Unitless         Format:A1     C*1
 Col.#  55:v                Unit:Unitless         Format:A1     C*1
 Col.#  56:l                Unit:Unitless         Format:A1     C*1
 Col.#  57:s                Unit:Unitless         Format:A1     C*1
 Col.#  58:i                Unit:Unitless         Format:A1     C*1
 Col.#  59:p                Unit:Unitless         Format:A1     C*1
 Col.#  60:f                Unit:Unitless         Format:A1     C*1
 Col.#  61:c                Unit:Unitless         Format:A1     C*1
 Col.#  62:w                Unit:Unitless         Format:A1     C*1
 Col.#  63:commentline      Unit:Unitless         Format:A15    C*15
 Col.#  64:LABEL            Unit:PSF radius       Format:E12.6  R*4
 Col.#  65:broad1           Unit:Unitless         Format:A1     C*1
 Col.#  66:broad2           Unit:Unitless         Format:A1     C*1
 Col.#  67:broad3           Unit:Unitless         Format:A1     C*1
 Col.#  68:hard1            Unit:Unitless         Format:A1     C*1
 Col.#  69:hard2            Unit:Unitless         Format:A1     C*1
 Col.#  70:hard3            Unit:Unitless         Format:A1     C*1
 Col.#  71:soft1            Unit:Unitless         Format:A1     C*1
 Col.#  72:I7               Unit:Unitless         Format:A1     C*4
 Col.#  73:                 Unit:Unitless         Format:A1     C*8
 Col.#  74:LABEL10          Unit:                 Format:E12.6  R*4
 Selection: ALL

>How-To-Repeat:
CREATE/COLUMN qsource :dist_scr_new R*4

The error is obtained by different users in different machines in different 
days, so I would say it is reproducible.

STATUS: Fixed

----------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Sun Jul 16 15:44:25 1995
>Originator:      Carlo Izzo

>Synopsis:        Wrong "see also" command - Command: LOG/OFF
>Severity:        non-critical
>Priority:        high
>Category:        midas-doc
>Class:           doc-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
In the "See also" section of the help for LOG/OFF (file /midas/94NOV/prim/help/log_off.hlq)
is listed the command LOG/TOFF. This command doesn't exist. The command
LOG/TOF exists instead.

STATUS: Fixed

----------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Sun Jul 16 14:00:18 1995
>Originator:      Carlo Izzo

>Synopsis:        Help files not according to template.hlq under system/unix - Command: SEVERAL COMMANDS INVOLVED
>Severity:        non-critical
>Priority:        high
>Category:        midas-doc
>Class:           doc-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
I am currently implementing a GUI for any command having a help file
according to the Midas convention (see file /midas/94NOV/system/unix/template.hlq)

The idea is to generate an appropriate GUI for any Midas (or Midas context)
command just retrieving the on-line help information from the appropriate
TeX files.

In order for the GUI to work properly, the help files should follow the defined convention
accurately enough. Below I report some quick-and-easy-to-fix "mistakes" in the
Midas help files:

- File /midas/94NOV/prim/help/edit_tabl.hlq (help for EDIT/TABLE):
  the parameter helps are not separated by "\ap\pa"

- File /midas/94NOV/applic/help/createfunc.hlq (help for CREATE/FUNCTION):
  the sentence ending with "work area;\" should end "work area;\\"
  (two backslashes, not just one - This is your character for "newline").

- File /midas/94NOV/prim/help/statistabl.hlq (help for STATISTIC/TABLE):
  Missing explanation for parameter "table" (even if trivial, it should
  be there!)

- File /midas/94NOV/prim/help/computcolu.hlq (help for COMPUTE/COLUMN):
  parameter "res_frame.columns" in syntax section
  should be "res.frame.column" as in parameter section

- File /midas/94NOV/prim/help/rebin_it.hlq (help for REBIN/IT):
  parameter "parm" in syntax section
  should be "param" as in parameter section

- File /midas/94NOV/prim/help/rebin_ii.hlq (help for REBIN/II):
  parameter "parm" in syntax section
  should be "param" as in parameter section

- File /midas/94NOV/prim/help/rebin_ti.hlq (help for REBIN/TI):
  parameter "parm" in syntax section
  should be "param" as in parameter section

- File /midas/94NOV/prim/help/convertabl.hlq (help for CONVERT/TABLE):
  parameter "FREQUENCY" in syntax section
  should be "method" as in parameter section

- File /midas/94NOV/prim/help/get_gcur.hlq (help for GET/GCURS):
  parameter "outp_spec" in syntax section
  should be "output_spec" as in parameter section

- File /midas/94NOV/prim/help/overplhist.hlq (help for OVERPLOT/HISTO):
  parameter "image" in syntax section
  should be "frame" as in parameter section

- File /midas/94NOV/prim/help/overplgrid.hlq (help for OVERPLOT/GRID):
  parameter "xy_opt" in syntax section
  should be "xy" as in parameter section

- File /midas/94NOV/prim/help/overplline.hlq (help for OVERPLOT/LINE):
  parameter "l_type" in syntax section
  should be "ltype" as in parameter section

- File /midas/94NOV/prim/help/show_icat.hlq (help for SHOW/ICAT):
  parameter "cat_name" in syntax section
  should be "catalog" as in parameter section

- File /midas/94NOV/applic/help/selectfunc.hlq (help for SELECT/FUNCTION):
  parameter "number" in parameter section
  should be "number,..." as in syntax section

- File /midas/94NOV/applic/help/replacfunc.hlq (help for REPLACE/FUNCTION):
  parameter "funci" in parameter section
  should be "fun1,fun2,..." as in syntax section

- File /midas/94NOV/applic/help/createfunc.hlq (help for CREATE/FUNCTION):
  parameter "funci" in parameter section
  should be "fun1,fun2,..." as in syntax section

- File /midas/94NOV/applic/help/createstar.hlq (help for CREATE/STAR):
  parameter "in_frm" in syntax section
  should be "in_frame" as in parameter section
  parameter "in_tab" in syntax section
  should be "in_table" as in parameter section
  parameter "out_frm" in syntax section
  should be "out_frame" as in parameter section

- File /midas/94NOV/applic/help/integrline.hlq (help for INTEGRATE/LINE):
  parameter "x.sta" in syntax section
  should be "x_sta" as in parameter section
  parameter "x_pos, range" in syntax section and parameter section
  should NOT contain blanks (in fact, this is misleading for the
  user, who would include them in his command line erroneously)

- File /midas/94NOV/applic/help/integrstar.hlq (help for INTEGRATE/STAR):
  parameter "out_tab" in syntax section
  should be "out_table" as in parameter section

- File /midas/94NOV/prim/help/regresline.hlq (help for REGRES/LINE):
  parameter "xi" in parameter section
  should be "x1,x2,..." as in syntax section

- File /midas/94NOV/prim/help/regrespoly.hlq (help for REGRES/POLY):
  parameter "xi" in parameter section
  should be "x1,x2" as in syntax section
  The help for the parameters "degree1" and "degree2" is missing.

- File /midas/94NOV/prim/help/modifygcur.hlq (help for MODIFY/GCURS):
  parameter "y_coord" in parameter section
  should be "y-coord" as in syntax section

- File /midas/94NOV/prim/help/copy_zoom.hlq (help for COPY/ZOOM):
  parameter "ITTnam" in parameter section
  should be "ITTdef" as in syntax section

- File /midas/94NOV/prim/help/print_acat.hlq (help for PRINT/ACAT):
  File /midas/94NOV/prim/help/print_icat.hlq (help for PRINT/ICAT):
  File /midas/94NOV/prim/help/print_fcat.hlq (help for PRINT/FCAT):
  File /midas/94NOV/prim/help/print_tcat.hlq (help for PRINT/TCAT):
  parameter "cat_nam" in syntax section
  should be "cat_name" as in parameter section

- File /midas/94NOV/prim/help/print_tabl.hlq (help for PRINT/TABLE):
  parameter "elem" in syntax section
  should be "elem1" as in parameter section

STATUS: Partially fixed

---------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Fri Jul 14 17:59:28 1995
>Originator:      Carlo Izzo

>Synopsis:        Undocumented parameters - Command: OVERPLOT/AXES
>Severity:        non-critical
>Priority:        high
>Category:        midas-graphic
>Class:           doc-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
The parameters x_offset and y_offset, indicated in the syntax line,
are not documented. They probably should be removed.

STATUS: Fixed

----------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Fri Jul 14 10:50:07 1995
>Originator:      Carlo Izzo

>Synopsis:        In the "see also" section of the help should be written CENTER/GAUSS - Command: CENTER/MOMENT
>Severity:        non-critical
>Priority:        high
>Category:        midas-doc
>Class:           doc-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
In the "See also" section of the on-line help (file centermome.hlq)  should be
written "CENTER/GAUSS", and not "CENTER/MOMENT" again.

STATUS: Fixed

--------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Wed Jul 12 17:02:53 1995
>Originator:      Carlo Izzo

>Synopsis:        Inconsistent parameter name - Command: REBIN/TT
>Severity:        non-critical
>Priority:        high
>Category:        midas-doc
>Class:           doc-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
In the syntax section the sixth parameter is called "parm", while in the
lines below (helps for each parameter) the same parameter is called "param".

This makes my automatic GUI generator fail! This is a quick fix, I just care that
it is done by the next MIDAS release - for the moment I can change that myself.
I agree that you may not care how and why my GUI generator fails, but
the on-line helps are supposed to follow some set of basic conventions anyway,
and with no exceptions...

STATUS: Fixed

----------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Wed Jul 12 16:53:24 1995
>Originator:      Carlo Izzo

>Synopsis:        Bad default - Command: FILTER/GAUSS
>Severity:        serious
>Priority:        medium
>Category:        midas-image
>Class:           change-request
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Since radx and rady determine the radius of the box in which the pixel 
are used to calculate the Gaussian weighting, shouldn't be that box at least,
say, 3 sigma wide, so that the Gaussian function is nearly zero at the 
boundaries of that box? So, I would have thought reasonable choices of the
input parameters would normally be: 

filter/gauss inframe outframe 9,9 9,3,9,3
filter/gauss inframe outframe 6,6 6,2,6,2

or even

filter/gauss inframe outframe 3,3 3,1,3,1

but not

filter/gauss inframe outframe 3,3 3,3,3,3

as it is now.

As I understand it, according to sampling theory, if the weights at the 
borders of the box with which you are convolving your picture are not 
about equal to zero, your Fourier transform will show some sinc-like
spikes. 

By making tests, I could also confirm that there are some substructures 
(artifacts) in my image if I use a box radius of one sigma, which are
gone if I use a box radius of 3 sigma. (And the computing time is even
for a 15*15 pixel box negligible). 
Anyway, I think a good default would be:

filter/gauss inframe outframe 9,9  9,3,9,3

STATUS: Fixed

------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Thu Jun 22 15:30:53 1995
>Originator:      Carlo Izzo

>Synopsis:        Still some errors with declination sign when close to zero - Command: COMPUTE/PRECESSION

>Severity:        critical
>Priority:        high
>Category:        midas
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Still some errors with declination sign when close to zero
>How-To-Repeat:
Mi_66 003> COMPUTE/PRECESSION 01,13,00 -1,8,0 1950 2000
Coordinates precessed to 2.0000000000E+03 are: 
R.A.:  01 hrs 15 min 3.331E+01 sec   Decl.: 00 deg 52 arcmin 1.015E+01 arcsec 

The result for declination should be -00 52 10.15   (negative, not positive)

STATUS: Fixed

------------------------------------------------------------------------------

>From mjf@rosat.mpe-garching.mpg.de Thu Jun 22 14:35:54 1995
>Originator:      Michael Freyberg

>Synopsis:        Failure in retrieving double precision columns (in tables) - Command: OUTTAPE/FITS   INTAPE/FITS
>Severity:        critical
>Priority:        high
>Category:        midas-dataio
>Class:           sw-bug
>Release:         94NOV

>Environment:     
System: Open VMS 
Machine: alpha
>Description:
We have the following table:

Midas 016> show/tab out

 Table : out                             [Transposed format]
 No.Columns :       7   No.Rows :      10
 All.Columns:       9   All.Rows:      16         Sel.Rows:      10
 Sorted  by :Sequence  Reference:Sequence
 Col.#   1:TIME             Unit:sc_sec           Format:F14.4  R*8
 Col.#   2:AMPL             Unit:channel          Format:I4     I*4
 Col.#   3:XPIX             Unit:sky_pix          Format:I7     I*4
 Col.#   4:YPIX             Unit:sky_pix          Format:I7     I*4
 Col.#   5:XDET             Unit:det_pix          Format:I5     I*4
 Col.#   6:YDET             Unit:det_pix          Format:I5     I*4
 Col.#   7:RAW_AMPL         Unit:channel          Format:I8     I*4
 Selection: ALL

We convert it into FITS with 

outtape/fits out.tbl out.mt

and back to table format with

intape/fits 1 out out.mt

Now in the newly created out0001.tbl table the :TIME column is filled with
zeroes.

STATUS: Fixed

-----------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Thu Jun  8 14:21:27 1995
>Originator:      Carlo Izzo

>Synopsis:        Program crashes accessing not existing keyword COVER - Command: AOV/TSA
>Severity:        critical
>Priority:        high
>Category:        midas-dist
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Fix: 

40-th line of file /midas/94NOV/contrib/proc/tsaaov.prg should be modified from

	IF COVER.LT. 1 THEN

to

	IF COVERPLO .LT. 1 THEN
>How-To-Repeat:
Just run the program

STATUS: Fixed

-------------------------------------------------------------------------------

>From mmp@rosat.mpe-garching.mpg.de Wed May 24 13:59:02 1995
>Originator:      Marguerite Pierre

>Synopsis:        It fails when descriptor START has too many significant figures - Command: PLOT/AXES
>Severity:        critical
>Priority:        high
>Category:        midas-graphic
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o07 V3.0 347 alpha
Machine: alpha
>Description:
It seems that when the descriptor START contains too many significant figures,
PLOT/AXES fails when graph is assigned to display where image is loaded
>How-To-Repeat:
crea/ima test 2,155,139  328.41266503905,17.684192005371,0.21372667106334E-03,0.20453814067878E-03 gauss 
load test
assign/grap display
plot/axes

Error message:
pixel outside valid range ...
(APP) application error - 2
*** FATAL: Cextrco: invalid coordinate input ...

Everything works when the first command of the above sequence is replaced by

crea/ima test 2,155,139 328.412,17.684,0.21372667106334E-03,0.20453814067878E-03 gauss 50,50,50,50

STATUS: Fixed

-------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Wed May 24 11:42:59 1995
>Originator:      Carlo Izzo

>Synopsis:        When more than 8 parameters given, you get "crazy" behavior - Command: READ/TABLE
>Severity:        non-critical
>Priority:        low
>Category:        midas-table
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
A user did a mistake (didn't read the help) and used the commands

READ/TABLE test :COL01 :COL02 :COL03 :COL04 :COL05 :COL06 :COL07
READ/TABLE test :COL01 :COL02 :COL03 :COL04 :COL05 :COL06 :COL07 @1..4
 
He said that the first works, and the second "you get only the first 4 columns 
listed but for all 10 rows!!! In this case it sometimes can happen (depends on 
the column names) that you obtain even NO output"

Now, I have told him to use

READ/TABLE test :COL01,COL02,COL03,COL04,COL05,COL06,COL07 @1..4

Nevertheless, a check should be done in order to issue the appropriate
error message when the user does such mistakes...

STATUS: Fixed

-------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Thu May 18 18:43:32 1995
>Originator:      Carlo Izzo

>Synopsis:        Hard to describe it in short! - Command: COMPUTE/TABLE , WRITE/TABLE , CREATE/TABLE
>Severity:        non-critical
>Priority:        medium
>Category:        midas-table
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Once a table is created and columns are defined, they can be initialized
using a COMPUTE/TABLE, and this is fine.

But if the first compute table is preceded by a WRITE/TABLE to a given
row, the next COMPUTE/TABLE will work just up to that row. This happens
obviously because the rows are defined just up to the position of the
highest write statement.

This behavior appears inconsisten (if not wrong) to virtually all the users.
One expects that to create a 100 rows table means to create a table
with 100 rows defined - and not just allocated.
>How-To-Repeat:
First, the following works fine:

crea/tab test 1 100 null
crea/col test :x 
compute/table test :x = 10

In the above example, all column elements are initialized to 10.

Instead the next fails:

crea/tab test 1 100 null
crea/col test :x 
write/tab test :x @10 1
compute/table test :x = 10

In this case the defined rows are 10, and only the first 10 elements of the
:x column are initalized to 10.

All users find this behavior silly, and it is worthless to try to explain them
the difference between allocated and defined rows.

STATUS: Fixed

-------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Wed May 17 13:39:17 1995
>Originator:      Carlo Izzo

>Synopsis:        Problem with table expansion mechanism when character columns present - Command: CREA/COLUMN, COPY/TT
>Severity:        serious
>Priority:        high
>Category:        midas-table
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
This is a correction to the description of this problem sent in a previous report
(It has the internal identification `midas-table/668').

Apparently, in order to reproduce the problem it is not essential that the character
columns are consecutive.
>How-To-Repeat:
Try to create a new column in a table with the following structure:

 Table : carlo2                          [Transposed format]
 No.Columns :      17   No.Rows :      43 
 All.Columns:      15   All.Rows:      48         Sel.Rows:      43
 Sorted  by :Sequence  Reference:Sequence     
 Col.#   1:name             Unit:                 Format:A6     C*6
 Col.#   2:v                Unit:                 Format:E12.6  R*4
 Col.#   3:pmm              Unit:                 Format:A1     C*1
 Col.#   4:bmv              Unit:                 Format:E12.6  R*4
 Col.#   5:phm              Unit:                 Format:A2     C*2
 Col.#   6:vmi              Unit:                 Format:E12.6  R*4
 Col.#   7:vrm              Unit:                 Format:A1     C*1
 Col.#   8:ra_deg           Unit:degree           Format:F10.5  R*4
 Col.#   9:dec_deg          Unit:degree           Format:F10.5  R*4
 Col.#  10:mem              Unit:                 Format:A2     C*2
 Col.#  11:x_sky            Unit:sky_pix          Format:F15.4  R*4
 Col.#  12:y_sky            Unit:sky_pix          Format:F15.4  R*4
 Col.#  13:exi_ml           Unit:                 Format:F7.1   R*4
 Col.#  14:ra_x             Unit:degree           Format:F9.5   R*4
 Col.#  15:dec_x            Unit:degree           Format:F9.5   R*4
 Col.#  16:dra              Unit:                 Format:E12.6  R*4
 Col.#  17:ddec             Unit:                 Format:E12.6  R*4
 Selection: ALL

	Please notice that this is a "saturated" table.

STATUS: Fixed

------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Tue May 16 16:10:19 1995
>Originator:      Carlo Izzo

>Synopsis:        Table expansion mechanism fails when character columns present - Command: COPY/TT, CREATE/COLUMN
>Severity:        serious
>Priority:        high
>Category:        midas-table
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
When the number of allocated columns in a table is used up, any subsequent
column creation (e.g. via COPY/TT or CREATE/COLUMN) implies the table
expansion (i.e. the table is copied into a new one having a larger number of
allocated columns).

This "expansion" mechanism typically fails when more than one character columns
in sequence are defined in the table (the system either crash, or the output
table is corrupted).

>How-To-Repeat:
Try with a table like this one:

 Table : /home/izzo/carlo.tbl            [Transposed format]
 No.Columns :      15   No.Rows :      43 
 All.Columns:      13   All.Rows:      48         Sel.Rows:      43
 Sorted  by :Sequence  Reference:Sequence     
 Col.#   1:name             Unit:                 Format:A6     C*6
 Col.#   2:pmm              Unit:                 Format:A1     C*1
 Col.#   3:phm              Unit:                 Format:A2     C*2
 Col.#   4:vrm              Unit:                 Format:A1     C*1
 Col.#   5:mem              Unit:                 Format:A2     C*2
 Col.#   6:v                Unit:                 Format:E12.6  R*4
 Col.#   7:bmv              Unit:                 Format:E12.6  R*4
 Col.#   8:vmi              Unit:                 Format:E12.6  R*4
 Col.#   9:x_sky            Unit:sky_pix          Format:F15.4  R*4
 Col.#  10:y_sky            Unit:sky_pix          Format:F15.4  R*4
 Col.#  11:ra_deg           Unit:degree           Format:F10.5  R*4
 Col.#  12:dec_deg          Unit:degree           Format:F10.5  R*4
 Col.#  13:exi_ml           Unit:                 Format:F7.1   R*4
 Col.#  14:ra_x             Unit:degree           Format:F9.5   R*4
 Col.#  15:dec_x            Unit:degree           Format:F9.5   R*4
 Selection: ALL

As you can see, this is an example of "saturated" table (13 allocated columns,
and 10 defined R*4 plus 6+1+2+1+2=12 characters = 12 bytes = 3 words
for a total of 13 defined words).

Then try CREATE/COLUMN, and the program either will crash, or the output
table will turn out to be corrupted.

STATUS: Fixed

------------------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Thu May 11 15:44:14 1995
>Originator:      Carlo Izzo

>Synopsis:        label/graph plot the string at the bottom of the page - Command: LABEL/GRAP
>Severity:        serious
>Priority:        high
>Category:        midas-graphic
>Class:           sw-bug
>Release:         94NOVpl1.2

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
It seems that when the YSCALE is set to some value, the command LABEL/GRAPH
doesn't place the label in the right (world coordinate) position when the Y range
of the plot is extending to too low values (these values depend somehow on
the value of YSCALE,: for instance, if in the example below -18.7 is replaced
with -17, the string BATSE is placed in the right position; in the same way, if
yscale is changed from -90 to -100,  again BATSE would be placed in the right
position).
>How-To-Repeat:
assign/grap p2c1.p nospool
set/grap
set/grap xscale=-145 yscale=-90 
plot/axes 7.5,29.5,2,-1 -18.7,-2.4,2,-1
label/grap BATSE 1.e+20,5.e-5 ? 0.7
copy/grap p2c1.p

(p2c1 is the name of out postscript BW printer)

STATUS: Fixed

-----------------------------------------------------------------------

>From izzo@rosat.mpe-garching.mpg.de Fri Apr 28 18:59:35 1995
>Originator:      Carlo Izzo

>Synopsis:         - Command: COMPUTE/PRECESSION
>Severity:        critical
>Priority:        high
>Category:        midas
>Class:           sw-bug
>Release:         94NOVpl1

>Environment:     
System: OSF1 o02 V3.0 347 alpha
Machine: alpha
>Description:
Still some errors with declination sign when close to zero!
>How-To-Repeat:
Mi_66 003> COMPUTE/PRECESSION 01,13,00 -1,8,0 1950 2000
Coordinates precessed to 2.0000000000E+03 are: 
R.A.:  01 hrs 15 min 3.331E+01 sec   Decl.: 00 deg 52 arcmin 1.015E+01 arcsec 

The result for declination should be -00 52 10.15   (negative, not positive)

STATUS: Fixed

----------------------------------------------------------------------------