Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.sao.ru/precise/Midas_doc/courier/courier-92jun.ps Äàòà èçìåíåíèÿ: Wed Mar 29 20:15:37 1995 Äàòà èíäåêñèðîâàíèÿ: Sun Sep 12 11:36:50 2010 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: earth's atmosphere

European Southern Observatory
Newsletter of the MIDAS Users' Community
MIDAS Courier
Volume 2 Number 1 June 1992
Editorial
As has been announced in the previous issue of
the Courier and in the recent MIDAS Users Meeting,
starting this year MIDAS will be officially released
only once per year; the next release is to be expected in
November (release 92NOV). For the Courier this has
some consequences. Since it was published shortly
after the semi­annual MIDAS release, part of its
contents was dedicated to new features in that release.
With the annual release to appear in November,
it is obvious that this kind of information will be
more suitable for the December issue of the Courier.
Clearly, the Courier will continue to inform you
about current developments in and plans for MIDAS.
However, in particular in the June issue, we will try
to obtain more user contributions. At the last MIDAS
Users Meeting several people have already indicated
their willingness to contribute.
During the MIDAS Users Meeting we again have
tried to get your judgement about MIDAS (installation,
application software, documentation, user­support) via
a questionnaire. A similar action was taken two years
ago. A quick look at the response and comparison
with the previous questionnaire shows that we are
on the right track: after verifying the basic MIDAS
commands, and rewriting several parts of the User
Manual and installation documentation, we found that
the users' appreciation has clearly improved over the
whole range of our activities. We also noticed that
a substantial number of questionnaires have not been
returned. We hope that you can find time to complete
and return them. It would help us to find out where
users see room for improvements.
In and outside ESO four major # long term #
MIDAS projects have started recently. These projects
should result in off­line MIDAS data analysis software
for four VLT instruments that were approved ESO's
Governing bodies in the last year as part of the
VLT Instrumentation Plan. These four instruments
are: ISAAC, Infrared Spectrometer And Array
Cameras; UVES, UV­Visual Echelle Spectograph;
FORS, FOcal Reducer/Low Dispersion Spectrograph;
CONICA, COud’ e Near­Infrared CAmera. The first
two instruments will be built by ESO, and MIDAS
data reduction software will be written in­house. The
two other instruments will be built by two consortia.
The MIDAS software for these instruments will be
created by these consortia. In these cases, members of
the MIDAS Group operate as consultants to assure that
the software is written according to MIDAS standards
and is compatible with other parts of the system. The
most recent issue of The Messenger (No. 67 # March
1992) contains detailed information about FORS and
CONICA.
a a a a
In this ESO­MIDAS Courier:
a
Editorial
a
1992 MIDAS Users Meeting
a
General, System and Applications News
a
AGL Upgrade to Version 3.6
a
Graphical Interfaces: First Experiences
a
MIDAS Used for Data Acquisition of IRAC 2
a
ROSAT Data Analysis with EXSAS
a
Expeditions into the MIDAS Jungle: Images and
Descriptors
a
MIDAS Questions and Answers
a
The MIDAS Patch File Summary
a
Problems found in the 91NOV release
a
a a

2 ESO­MIDAS Courier, Vol. 2, No. 1, June
1992
a a
1992 MIDAS Users Meeting
Dietrich Baade,
European Southern Observatory
aa
For technical reasons, the annual MIDAS Users
Meeting could in 1992 be held only in May. This
change from the rainy Aprils of previous years was
most audibly appreciated by many participants. But
also otherwise the meeting was held in a rather
constructive atmosphere.
The meeting followed the same pattern as in previous
years. In the hour before the coffee break, Preben
GrosbÜl gave an overview of the most relevant
changes that had taken place in the past year, including
new or modified procedures for the distribution,
upgrading and maintenance of MIDAS. Selected
technical aspects were then expanded upon in more
detail by other Image Processing Group (IPG)
members who had been involved in developing them.
Most of this has been or is being reported separately
in the MIDAS Courier. Therefore, this summary will
rather concentrate on the two hours' general discussion
(chaired by the undersigned).
Useful nuclei for the discussion were questionnaires
distributed in advance. From the roughly 60­70
participants, only about 10 filled­in forms were
received, and no clear common concern could be
extracted from them. The surmise that this would
indicate that the MIDAS community did not feel
plagued by very general problems was confirmed
during the discussion. It happened only very rarely
that when a question had been asked or a comment
made some other participant raised his hand and
seconded the previous speaker.
One of the concerns expressed was about sudden
and not obviously technically inescapable changes
in the syntax and names of commands or how and
where results are stored. The negative impact which
such discontinuities have on user­written command
procedures was seen as an unnecessary nuisance. A
conscientious effort should be made to minimize such
modifications. Where they really are unavoidable,
it was suggested, there should at least be a tool
which can scan procedures files for the occurrence
of commands which may cause troubles of this kind
so that the user can check what modifications may be
required. The IPG agreed that such a tool can and
will be made available.
Some debate developed around the question how
it might be possible to start a MIDAS command
procedure in background and let it continue running
even after logout from the main session. Several
people suggested recipes but no one could say for
sure whether his method would also work on other
hardware platforms. Klaus Banse of the IPG will
investigate the portability issue.
A comment that saw relatively many attendants
nodding approvingly was that often the error messages
issued by MIDAS commands are not very meaningful
or misleading. Although cases can probably be
found where single low­level commands show such
a deficiency, this problem most often arises from
the nesting of basic commands in more complex
procedures where more specific error messages would
be more appropriate. This problem is well known.
But the IPG does not see much of a chance for
improvements without heavy investment of manpower
which might be more productively used by expanding
the functionality to MIDAS. However, even more
rigorous checking by such commands of input
parameters for being of correct type and within range
should provide an improved shielding.
In connection with the new procedures for the
dissemination of patches, some discussion spawned off
about the accessibility of the ESO computers through
the networks. ESO is fully aware of this problem. Yet,
there is little that ESO can do itself to improve this
situation because ESO depends entirely on the services
available through the Deutsche Bundespost Telekom
and Deutsches Forschungsnetz. Nevertheless, ESO
will continue to make its dissatisfaction with the
present situation known.
A capability which would facilitate many data
reduction tasks would be the consistent handling of
undefined pixels. With the concept of the NULL
value keyword, this capacity can be integrated into
any applications command. However, the IPG does
not presently have the capacity to implement this
feature throughout the existing MIDAS. Within the
forthcoming CCD package this point will be given
full consideration.
Often, users find it desirable to do computations
only on the SELECTed rows of a table rather
than on all of them. In the 91NOV release, this
could still be achieved only by going through the
tedious step of first copying the selected rows to a
new table, working on that copy, and then copying
them back to the original table. The argument
against deviating from the convention that certain
commands such COMPUTE/TABLE reset the selection
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992 3
flag to ALL before execution is, of course, that
operations restricted to the selected rows could very
quickly lead to a status of the table which can
no longer be traced back to its origin. As a
compromise, from the next release on the command
COMPUTE/TABLE :column = SEL will create a
column with only 0's or 1's according to whether or
not a row is selected. This column can then be used
to restrict the effect of subsequent operations to the
selected rows only and nevertheless not generate an
undocumentable chaos. In this way, also a question
by two participants in the users meeting could be
answered.
Most of the remaining suggestions or questions dealt
with relatively well confined, individual problems
or bugs. Often they concerned various applications
packages, and it was noted that they, too, could benefit
from a systematic verification as has been carried out
for the MIDAS core commands. ESO will keep this
in mind. However, these more specialized packages
usually require a real expert for their proper testing,
and the expertise among ESO's scientific staff does not
cover all areas equally well, certainly not at all times.
In this connection, a reminder may be in order that two
years ago ESO's Scientific and Technical Committee
and Council endorsed a document outlining the
MIDAS policy which a.o. encourages the formation
of regional centers of MIDAS expertise for just this
purpose, namely to develop and maintain dedicated
packages which require a very special scientific
interest.
Finally, a participant inquired whether there will
be more HST instrument­specific data reduction
and calibration software in MIDAS. While MIDAS
provides a perfectly suitable platform for such tools,
their development or implementation clearly is outside
the scope and responsibility of the IPG. Therefore,
the question was passed on to the deputy head of
the Space Telescope­European Coordinating Facility,
Rudolf Albrecht, who was among the participants.
Although no clear answer was given, it appeared that
no major efforts are being made by the ST­ECF to
offer HST calibration and reduction software to the
European MIDAS users community.
A personal observation is that in spite of this lengthy
report, my feeling at the meeting was that the
discussion would perhaps have significantly benefited
from more and more spontaneous contributions. The
more lively and intense the interaction between
MIDAS users and IPG is, the stronger MIDAS should
grow.
The MIDAS users meeting was concluded in the
afternoon with various demos of new or much
improved packages and especially the new graphical
user interface which on a trial base has been
implemented for the echelle and general spectroscopy
packages. The demos found surprisingly much interest
even among old­time MIDAS users. Many people also
used this opportunity for numerous individual contacts
among MIDAS users or between MIDAS users and
IPG staff.
A parallel demonstration of the brand­new remote
control system for the NTT with EMMI and SUSI by
Anders Wallander was also very well attended. Apart
from being based entirely on the UNIX operating
system and X­Window techniques, it also deploys
MIDAS as the on­line data reduction system which
can be run both on the remote and the local
workstation. By using a new MIDAS feature which
permits one to run a MIDAS session on some
workstation but use an image or graphics display
window on some other terminal or workstation, even a
time­efficient alternative to transferring the complete
file is available.
a
a a
Artist impression of the MIDAS Users meeting.
Drawing courtesy of Daniel Ponz .
aaa
Newsletter of the ESO­MIDAS Users' Community

4 ESO­MIDAS Courier, Vol. 2, No. 1, June
1992
a a
General, System, and Application News
The MIDAS Group, European Southern Observatory
aa
General News
The MIDAS Sites Database
A MIDAS sites Database is now in use at ESO
IPG. The Database has been developed within
STARCAT by Miguel Albrecht from the ESO Archive
Group. Detailed site information such as hardware
configuration, MIDAS release in use, etc. is entered
into the Database. A hardware specification form
was mailed out in April of this year, to obtain the
most accurate data. Approximately 50% of all user
sites actually completed and returned this form. As
the use of this Database will play a vital role in
speeding up the distribution process and making it
more efficient, it is rather important for us to have
the correct information. We, therefore, kindly ask
all user sites who have not returned the hardware
specification form to do so. In future, we would also
appreciate receiving information concerning changes
at the MIDAS user sites.
As an additional service to the user community,
external users who have a network connection to
ESO will be able to access (through STARCAT) the
MIDAS Sites Database to obtain information about
other user sites with possible similar configuration and,
therefore similar experiences/problems. More detailed
information concerning the external user interface will
be provided in a circular letter within the next few
weeks.
MIDAS Environment document
Currently, the MIDAS Group is working on the
update of the MIDAS Environment document, which
contains all information for the development of
MIDAS application code, either in the MIDAS Control
Language, or in Fortran or C. Besides improvements
of various chapters, we will include more coding
examples. Also, to facilitate easier use, the new
version will contain an index. We hope to send
copies of the new version to all MIDAS sites around
August­September. Additional copies can be obtained
from the MIDAS Group at ESO; contact Resy de
Ruijsscher.
MIDAS installed on Silicon Graphics
At the Max Planck Institute for Extraterrestrial Physics
at Garching near Munich, MIDAS is now running
on a Silicon Graphics, Model IRIX 4D35, running
under IRIX 4.01. Minor modifications were needed
in some makefiles because of default rules of the
make utility on this machine. Also, because of a
conflict between the MIDAS oserror variable and
the system one, the MIDAS one was renamed. These
installation adjustments and their verification took a
couple of hours, after which MIDAS was successfully
installed.
System News
AGL upgrade to Version 3.6
During the past year we received a number of
requests from MIDAS users for an enhancement
of the graphics software. Speed improvements and
additional functionality were among the main wishes.
The MIDAS graphics relies heavily on the lower­
level Astronet Graphic Library software, written by
Luca Fini of the Osservatorio di Arcetri, Florence.
Therefore, we invited Luca Fini for the period
December 1991 to February 1992 to visit ESO, and to
upgrade AGL in close collaboration with the MIDAS
Group. In the three month period, Luca was able to
implement most of the requests, as well as a number
of other features in the new version. An overview of
these improvements is given in a separate contribution
elsewhere in this Courier.
Obviously, the upgrade of the AGL will only become
beneficial to the MIDAS users if corresponding
adjustments are made in the MIDAS graphics
commands. Therefore, we also upgraded the MIDAS
Graphic system to take full advantage of the enhanced
performance of the AGL software. A summary of new
features in MIDAS graphics system will be given in
the next issue of the Courier.
The new version of AGL (Version 3.6), together with
the improved MIDAS commands, is currently tested
in the MIDAS development version at ESO. The
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992 5
complete graphic package will become available in
the 92NOV release.
Application News
MIDAS photometry package
A general­purpose package to plan and reduce
photometric observations is being developed by
Andrew T. Young, who is spending a year in
Garching for this purpose. There will be two main
parts: a planning program, to choose and schedule
observations of standard and extinction stars so that
the necessary observations can be obtained in a
minimum of observing time; and a reduction program,
to remove the effects of the Earth's atmosphere, and
to convert the results to a standard system.
A major problem in this effort is to adapt the
reductions to the wide variety of data formats produced
by various data­logging systems. To help solve this
problem, it would be helpful if photometric observers
at many observatories would send a sample of their
raw data, with a brief explanation of the quantities
recorded, to Andrew Young. The aim here is to make
sure the tools provided with the package will be able
to digest as many kinds of data as possible. Observers
who would like to contribute to this effort should e­
mail a sample data file and a brief description to
ayoung@eso.org.
More information will be given in later issues of the
Courier, when the project is farther along.
New CCD package
Currently, a new CCD package for the calibration of
CCD images is under development. The package will
replace the old one that has become rather dated,
and was only available until the last VAX/VMS
version of MIDAS (88OCT). Since then, CCDs
have improved substantially, allowing a large variety
of observing modes, and consequently, producing
increasing amounts of widely diverse data. The new
package will basically take care of the pre­analysis
calibration steps like subtraction of bias and dark
frames, flat field, and illumination correction. More
sophisticated tools will be implemented at a later stage.
The package will be able to run the calibration
procedures automatically (a pipeline process) with a
minimum of user interaction, as well as in a step
by step mode where the user has full control and
flexibility over the data processing. To provide the
automatic mode, extensive use will be made of the
image descriptors written at the telescope during
data acquisition. In addition, a keyword structure
will contain all settings that control the reduction
processes.
The package, in a basic version, is planned to
be become available in the 92NOV version, or
shortly after via ftp. Suggestions for the package are
welcome. You can send them to rwarmels@eso.org.
FILTER/COSMIC: Suggestion
for improvement
Below, Hans Schwengeler (Astronomical Institute,
University of Basel) describes a proposed change to
the FILTER/COSMIC routine in MIDAS to handle a
strongly varying background.
During the reduction of CCD images of the galaxy
NGC 5236, we found that no chosen value of the
sky was really adequate as input to the routine
FILTER/COSMIC. The frame (1100x1050 pixels)
has intensities ranging from 4000 in the lower right
corner up to about 8000 in the upper left corner,
in the direction where the center of the galaxy
lies. We now did first a median filtering, i.e.
FILTER/MEDIAN N5236 MEDIAN 2,2,0.0.
Then we tried FILTER/COSMIC N5236 MEDIAN
OUT 4500.0,10.0,4.0,10.0. This resulted in
good cleaning for only the lower right corner. Using
8000.0 for the sky had the effect that some stars were
already affected by the filtering. So the need arose to
adapt the filtering to a varying background.
After some experimenting, I found out that a small
number of changes do the trick. I introduced a #local#
sky estimator into the program. Now the user does
not have to specify a sky value himself, but the
program will determine it. Since I wanted to keep the
original version for "normal" CCD frames, I made a
new command, called XFILTER/COSMIC (X stands
for eXtended). I defined the command in the file
MID_MONIT:NEWCOM.IN as @ xcosmic. The file
MID_PROC:XCOSMIC.PRG is a copy of
RCOSMIC.PRG, but without the sky parameter. Of
course, I also provided help files, XFILTE.HLC
and XFILTECOSM.HLQ, both in the MID_HELP
directory.
The program itself is a copy of RCOSMIC.FOR (in the
directory /midas/91nov/prim/general/src
(UNIX)) and named XCOSMIC.FOR. The compiled
and linked program goes into directory MID_EXE.
The code, help files, and procedure files can
be obtained from me. Please, send mail
to: schwengeler@urz.unibas.ch (internet) or
CHGATE::YOGI::SCHWENGELER (decnet).
Newsletter of the ESO­MIDAS Users' Community

6 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
New IUE context
A new context has become available to read IUE data
in GO format from tape and convert them into MIDAS
Bulk Data Frame format (BDF) on disk. MIDAS
BDF files are created for the data types RAW, PHOT,
(E)LBL and FES; tables are created for the MELO
and MEHI extracted spectra. The package was made
available by Daniel Ponz at the IUE Station in Vilspa,
Spain. More information about the context will follow
in the next Courier.
aaa
a a
AGL Upgrade to Version 3.6
Luca Fini, Osservatorio di Arcetri, Florence
aa
The graphic functions provided by MIDAS are
implemented by two levels of code:
1. an upper­level layer that is "MIDAS aware"
and knows of data structures, environment,
commands, and the like;
2. a bottom layer that provides the basic graphic
primitives (lines, text, symbols) and that interacts
with the graphic devices.
AGL (Astronet Graphic Library) is the set of
tools which underlies the graphic functions provided
by MIDAS: it provides a consistent and device­
independent access to the various graphic devices
supported, and the basic functionalities needed to
produce graphic output.
The 92NOV version of MIDAS will include a
thoroughly revised version of AGL. The revision was
planned both in order to solve some problems with the
previous version and to provide more functionalities
required by MIDAS users.
In this respect one must consider that functionality
improvements seen by a MIDAS user require two
steps: a) the graphic library must provide tools to
implement the new features; b) the MIDAS­specific
graphic code must be modified in order to use those
new features, and to provide both more graphic
commands and new capabilities of existing ones.
The following lines give an overview of point a);
a description of point b) will be given in the next
issue of the Courier.
As a first improvement, the source code of AGL
has been thoroughly restructured to allow separate
compilation of modules. In fact, the previous AGL
code structure caused the overflow of compiler internal
tables on some machines. Moreover, the AGL code
has been profiled in order to spot some efficiency
bottle necks. A re­coding of those parts has led to a
36% improvement of drawing speed (measured on a
Sun Sparcstation).
In the new version, a number of new capabilities have
been included:
1. Colour management: consistent support of named
colours and background colour control. A driver
for colour PostScript printers has also been added;
2. The axis drawing routines have been redesigned
to allow more flexibility. Arbitrarily oriented axes
can now be drawn, and labels may be written in
degrees (hours), time and arc minutes, time and
arc seconds format;
3. Support for non­linear coordinate transformations:
polar and many other sky­mapping transformations
have been included;
4. Polygon drawing and filling routines.
Some users have also requested a well­defined
programming interface, like the ST/SC or TB/TC
interfaces, to be used in MIDAS application programs.
Using these interfaces one then could provide graphic
output without using MIDAS commands. This wish
has not, as yet, been provided, because of a number
of considerations which I will try to clarify below:
1. AGL is such an interface and could in principle
be used for that purpose, but, unfortunately, it
is much too "low level" to be useful: it cannot
manage the MIDAS environment and is not aware
of MIDAS data structures;
2. The set of routines in the upper part of the
graphic code in MIDAS is regularly updated, and
is thus not suitable to be frozen into an interface
definition;
3. The upper layer of the MIDAS graphic code is
currently written in FORTRAN. Hence, it could
only provide an interface for Fortran applications,
which is a major limitation for a general interface.
A Fortran and C interface would require the
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992 7
rewriting of that layer in C, which, obviously,
is a major task.
Thus, although the issue of a graphic interface for
MIDAS programming is not closed, it must be
carefully evaluated, and a workable solution is not,
at the moment, available.
As a final remark, the version of AGL used within
MIDAS is a special customized version, which is not
suitable for stand­alone applications. Programmers
who want to use AGL outside the MIDAS environment
must get the standard version of AGL from the author
(email: lfini@astrfi.astro.it) or from:
Astronet Documentation Facility (ADOC)
Osservatorio Astronomico di Trieste
C.P. Succ. TS 5
Via G.B. Tiepolo 11
34131 Trieste, Italia
SPAN: ASTRTS::ADOC
Internet: adoc@astrts.astro.it
aaa
a a
Graphical User Interfaces:
First Experiences
Pascal Ballester, European Southern Observatory
aa
Graphical user interfaces (GUIs) have been released
for the first time in MIDAS 91NOV. In this version,
two interfaces are available: XSpectra provides
reduction procedures for long slit and 1D spectra
and Filters provides characteristic curves of ESO's
filters. A new interface XEchelle is available in the
in­house MIDAS 92N03 version and is an interface to
the Echelle package. All these interfaces are based on
the Athena widget set provided with the MIT version
of X11. These GUIs were experimented at ESO on
La Silla and Garching by internal users and visitors.
Along with these first realizations came up two
concerns: The definition of a standard presentation
(the so­called #Look and Feel#), and the optimization
of the production process. Many ESO projects involve
the realization of GUIs and the definition of an ESO
#Look and Feel# is currently discussed.
Also, different GUI builders have been compared.
GUI builders are tools which enable the interactive
development of an interface, either through a form
editor or a specification language. They present many
advantages:
1. generation of bug­free, immediately executable
code;
2. clear separation between the graphical interface
and the application code;
3. code is generated in a standard way, allowing less
personal programming style and less differences
between applications.
GUI builders can be separated between management
systems (UIMS) interpreting a specification file, and
design systems (UIDS) producing a source code which
can be implemented in the applications. UIMS permit
a quick prototyping of interfaces but their capabilities
are limited to what allows the specification language.
Because they produce source code, UIDS are more
adapted for the development of MIDAS applications.
In addition, UIDS give a complete freedom for the
specification of interfaces and are only limited by the
possibilities of the widget set.
The standard X11 environment is distributed with
the set of Athena widgets. Applications developed
with this set of widgets are fully portable. Additional
public domain widget sets are available. However
those widget sets are limited and do not constitute a
coherent package. For this reason it has been chosen
to use the MOTIF widget set for the development
of MIDAS GUIs. The existing interfaces will be
rewritten with MOTIF widgets to conform to the ESO
standard #Look and Feel#.
Using UIM/X, the GUI builder retained to develop
MIDAS application, a new interface XHelp has been
realized which will be implemented in the in±house
92N07 version of MIDAS and the next release. The
XHelp interface is based on the Motif widget set and
provides access to MIDAS help files. At creation time,
the list of all available MIDAS commands is displayed
in the text area. This list results from the MIDAS
command HELP. The help file of any command is
obtained in a maximum of two clicks. A first click on
Newsletter of the ESO­MIDAS Users' Community

8 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
a command name (e.g. CREATE) corresponds to the
MIDAS command HELP CREATE and results in the
list of all commands CREATE/... Then a click on a
complete command name provides the corresponding
help file. In the same way can be obtained the help
files corresponding to commands referred in another
help, as found in the section #See also# of the help
files. The interface runs in parallel with the MIDAS
monitor, which means that help files can be read while
MIDAS is executing a command and that the interface
survives a MIDAS exit (BYE).
Two XHelp display windows. Top: the display window containing all available MIDAS commands for which help
is available. Bottom: the help text of a specific MIDAS command after two clicks in the XHelp window, or after
entering the command in the type­in area (area left to the #Main# button).
aaa Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
9
a a
1 a5 a0 a E aS aO a± aM a I a
D aA aS a U as ae a r a
A ag a r a
e ae am ae an a t a
s a
On June 1 st we have reached the next milestone in the ESO­MIDAS project: the 150 th ESO­MIDAS User
Agreement has been signed. The Agreement was signed by ESO and Instituto Isaac Newton at Santiago de
Chile. Indeed, it is a nice coincidence that the 150 th ESO­MIDAS site is located in the capital of the host
country of ESO's observatory on La Silla. We congratulate the Institute Isaac Newton and, once more, thank
the MIDAS users for their confidence in MIDAS.
aa aaa
a a
MIDAS Used for Data Acquisition of IRAC 2
Gert Finger and Peter Biereichel,
European Southern Observatory
aa
The new infrared camera IRAC 2 equipped with a
large format NICMOS3 256*256 Hg 10x Cd x T e array
detector was installed last May at the 2.2 meter
ESO/MPIA telescope on La Silla. Since no software
support was available to adapt the software of IRAC1
running on the HP1000, a new concept based on
UNIX workstations was implemented which allows
to use MIDAS for data acquisition and quick look
data reduction at the telescope.
Instrument Workstation
The general system overview can be seen in Figure
1. The central instrument control of IRAC 2 is
performed by rocky mountain basic (rmb) which
is an application running on top of the UNIX
operating system of an HP 370 workstation. Rmb
fully supports the X11 window system. The user
interface is mouse driven displaying status information
of detector, instrument and telescope by separate
pop up windows. Additional support is given to
the observer by graphics windows displaying the
measured transmission curves of the selected filters,
the wavelengths of neighbouring orders of the Fabry
Perot, the instrument efficiency as a function of
wavelength after execution of the MIDAS command
MAGNITUDE on a calibration star, transmission of the
atmosphere, etc.
The instrument functions are controlled by CAMAC
which is accessed from rmb via the GPIB bus. A
serial RS232 interface in CAMAC establishes the link
to the HP1000 telescope computer via TCINT which
is needed for getting the telescope coordinates, doing
automatic focus exposures, mosaicing of large fields
or microsteping the telescope for deep imaging.
Communication to the Detector
Preprocessor via FTP
The rmb statement EXECUTE allows access to the
underlying UNIX operating system taking advantage
of the full power of UNIX enhancements. Rmb
operation is suspended and the UNIX command
specified by rmb is executed.
The detector data acquisition system is VME based.
The controller is a motorola 68030 processor (Eltec
E6) on which the detector preprocessor software (pps)
is running under the OS9 operating system. The
communication to the pps software is established via
the ethernet 1 link. Command files for the detector
setup which have been created by rmb are transmitted
via socket communication calling a C program from
rmb. This program sends commands to the pps and
stores the replies into a file which is sent back to
the calling program via ftp. Detector images are also
transmitted from pps to the instrument workstation via
ftp. The data format is double precision.
It should be noted that pps is galvanically decoupled
by an ethernet fiber optics link in order to avoid ground
loops which is mandatory for achievement of excellent
detector noise performance. The pps system contains
also a graphics board for a real time display which
is connected to a monitor in the control room by
an rgb fiber optics video link. The fiber optics is
necessary to keep the detector electrically isolated as
just mentioned before.
Newsletter of the ESO­MIDAS Users' Community

10 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
Figure 1. Control setup for IRAC 2 installation at the 2.2±m telescope used in May 92.
Communication to the MIDAS
Host via NFS
If rmb receives an image from pps, it adds the fits
header corresponding to the instrument setup used for
the current exposure. The resulting fits file is created
on the disk belonging to the midas workstation HP
730. This disk is remotely mounted on the instrument
workstation by nfs and can be directly accessed by
rmb.
On the HP730 two parallel MIDAS sessions are
active. The monitor of the first MIDAS session is set
in the background mode using the midas command
MODE(4)=2. In this mode the midas monitor looks
in its working directory (MID_WORK) for mailbox
files to use as input. The mailbox file or send box is
created by rmb. It instructs midas to call a procedure
which does the following: It converts the fits file
created by rmb to a bdf file. It flips the image to
get the correct orientation on the sky. Optionally a
stored sky frame can be subtracted from the image.
A bad pixel mask can be applied to replace bad
pixels by neighbouring good pixels before the image is
displayed using appropriate cut levels. The bdf image
is copied to files numbered sequentially and an entry
to an image catalogue is made. Then rmb waits for
the creation of the receive box. After obtaining the
return status the receive box is deleted and a new send
box can be created for the next MIDAS command.
This background MIDAS session allows to control
incoming images without interference of the observer.
A second MIDAS session running in the foreground
gives the observer the possibility to interactively
reduce the images for a quick look during acquisition
of new exposures using the full power of the MIDAS
imaging package.
Conclusions
The described setup is provisional. It was used for
lab testing in Garching and is currently in use at
the telescope, until the final La Silla software is
available. The rmb software running on th HP370 will
be transferred to the HP730 workstation implementing
the user interface developed by La Silla staff. So
only one workstation will be needed to control the
instrument and to run MIDAS. Special rmb image
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992 11
processing routines for IRAC 2 are presently translated
into MIDAS application programs.
Installation of a new processor board (Eltec E7 68040 /
25 MHz) in the pps data acquisition system will allow
to use nfs. Thus image files will be directly written
on the disk of the HP730 instrument workstation. The
CAMAC motor controllers will be replaced by a VME
based system which can be accessed via ethernet.
The use of two separate LAN's increased the
reliability. Ethernet 1 is a completely separated
instrument LAN for undisturbed data flow between
the detector preprocessor, the instrument workstation
and the image processing workstation. Ethernet 2
communicates to the rest of the world. Before
separation of the instrument LAN from the outside
world a faulty bridge caused data communication
problems.
Two workstations reduce the load of the less powerful
instrument workstation, which in principle could
also run MIDAS. The use of two monitors is very
convenient and should be maintained in the final setup
by means of an additional X­terminal.
During the first installation of IRAC 2 in May 92 the
presented concept proved to be reliable. Each night
100 first quality infrared images have been acquired.
MIDAS was a very powerful and flexible observing
tool which could be easily adapted to our needs.
aaa
a a
ROSAT Data Analysis with EXSAS
Hans­Ulrich Zimmermann, Max Planck Institute for Extraterrestrial Physics, Garching bei M˜unchen
aa
The X­ray astronomy satellite ROSAT, since its launch
in June 1990, is performing an extremely successful
observatory mission. About 60,000 X­ray sources
(and with the XUV telescope in the order of 1000
XUV sources) have been detected in the initial All­Sky
Survey, and a similar number of sources is expected to
show up in the many images taken since the pointed
observations started 1.5 years ago. That wealth of
information # before ROSAT in the order of 5000
X­ray sources were known # offers exciting scientific
research possibilities that can be explored effectively
only in connection with appropriate analysis tools.
With EXSAS # the EXtended Scientific Analysis
System developed with an effort of 20 man years
by the ROSAT Scientific Data Center at the Max­
Planck Institut f˜ ur Extraterrestrische Physik (MPE,
Garching) # a comfortable system for the reduction
of data from the ROSAT X­ray and XUV instruments
has been made available. EXSAS comprises a
large collection of application modules as typically
required in analyzing data of this wavelength regime,
and runs as a specific context in the ESO­MIDAS
environment. Building EXSAS on top of MIDAS
using already existing applications (mainly for graphic
and image presentation) facilitated to a large extent
the development of the package and makes it easy
to install and use the software at the many sites
already familiar with MIDAS. EXSAS, completely
written in FORTRAN 77, takes full advantage of all
the standards used in MIDAS and therefore reflects
the same portability. To maintain independence from
the specifics of different operating systems also on the
data input side all ROSAT data are distributed in the
widely accepted FITS format.
EXSAS analysis modules are grouped into 4
application packages dealing with Data Preparation,
Spatial Analysis, Spectral Analysis and Timing
Analysis. A special EXSAS header, read and updated
by each application, maintains general information on
the origin, the history and the parameter space of the
data stored in tables and images.
The basic data set for all analysis is the so­called
Photon Event Table (PET). This is a space­ and
time­ordered list of all the X­ray photons collected
during a specific observation, each entry containing
the detector and (projected) sky coordinates plus the
amplitude (energy) and the time of the event. The Data
Preparation package is used to select data from this file
according to spatial, spectral and timing criteria, and
bin them into images, spectra or lightcurves. At the
same time, or as a separate task, instrument corrections
like telescope vignetting, filter transmission, detection
efficiencies and point­spread­function dependencies
can be applied. Additional modules allow the user to
manipulate spectra in different ways and prepare them
(by background subtraction and error calculation) for
spectral fitting procedures.
Newsletter of the ESO­MIDAS Users' Community

12 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
The Spatial Analysis section contains a sophisticated
package for detection of point­like and moderately
extended sources in X­ray images. Besides other
methods, a maximum­likelihood algorithm that takes
into account full information of all individual
photons, is used to evaluate best estimates for the
position, extent, and intensity, and corresponding
errors of individual sources. Other functions deal with
analyzing point­spread functions, extended structures
or offer coordinate manipulation and presentation
utilities.
For the Spectral Analysis branch, an instrument­
independent scheme has been implemented that
fits observed X­ray spectra with a multitude of
standard and user­defined source models, including
different error­calculation methods. The package is
complemented with utilities to determine fluxes and
luminosities, assuming different cosmological models,
and to present spectral fitting results.
In the Timing Analysis package the observer can
select power­spectrum calculations, auto­correlation
or cross­correlation methods, period folding with
barycentric correction of arrival times, or perform
different statistical tests on source variability.
A number of predefined procedures for standard
analysis tasks facilitate the reduction of data by
astronomers not intimately acquainted with all the
details of X­ray data evaluation. A comprehensive
EXSAS User's Guide, as well as the internal help
facilities in MIDAS style, and some tutorials, comprise
the internal documentation of the system.
The EXSAS software package can be requested
directly from the ROSAT Scientific Data Center at
MPE (by email to MPE::EXSAS or exsas@mpe.mpe­
garching.mpg.de). The only prerequisite is
access to a MIDAS installation. At MPE, EXSAS
is presently used by about 70 scientists. In addition,
the software has been distributed to 38 institutes all
over the world. Some astronomers also come directly
to the Data Center to use the system there. For this
purpose a number of image processing workstations
and some visitor service are available.
The Data Center will also in the future provide full
support and service for the package, updating and
extending the functionality at regular intervals.
aaa
a a
Expeditions into the MIDAS Jungle:
Images and Descriptors
Klaus Banse, European Southern Observatory
aa
In this second tour we explore the internal structure
of MIDAS images and their descriptors. We also
take a closer look at the possible operations involving
descriptors.
The data structure for images consists of three main
parts: a Frame Control Block (FCB) of 512 bytes,
the image data (pixels) and the descriptors containing
the descriptive information about the image. Among
other internal variables, the FCB contains the MIDAS
version with which the image was created, the number
of pixels and their format, as well as pointers to the
descriptor and pixel sections of the image.
The detailed layout of the FCB is not documented
to allow for upgrades of the internal structure of
images. Therefore, we strongly recommend accessing
MIDAS images and descriptors only via the ST or SC
interfaces in an application program.
If you really need to get at a MIDAS
image (or parts of it) directly, you should
look into the file fcbdef.h which is in the
directory $MIDASHOME/$MIDVERS/incl (UNIX)
or MID_DISK:['MIDASHOME'.'MIDVERS'.INCL]
(VMS). Together with a dump of the image file, this
should help you in locating the address of the data
you are interested in. Examples of this kind of image
access are in the IDL procedures, written by the ST­
ECF at ESO, to convert MIDAS images to IDL in
a restricted but quick way. But beware, there is no
guarantee that this method will work from one MIDAS
release to the next; if the internal format is updated,
your program has to be changed as well!
The FCB is followed by space for an initial set of
descriptors, then comes the space reserved for the
pixels of the image. Additional descriptor data is
appended to the end of the image. By the way,
MIDAS tables also use this data structure.
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992 13
The number of pixels of an image and their internal
format is fixed after the creation of an image, so if you
want to change the size or data format of an image
you have to copy the image to another one and rename
it back to the original name.
The descriptors of an image contain all the additional
information related to the image data, and correspond
to the KEYWORDS of the FITS files (an in­depth
article about the MIDAS FITS­reader/writer and its
handling of descriptors is planned for a future issue of
the MIDAS Courier). Each descriptor has an entry in
a ‘descriptor directory' containing the descriptor name,
type, size and a pointer to the actual descriptor data.
This directory itself is a normal MIDAS descriptor
named DIRECTORY.MIDAS with restricted access,
though.
Descriptors may be expanded after their creation, but
the maximum size of a descriptor is currently 32767
elements. You may also add descriptors at will to
an image, but since the descriptor directory has the
same limits in size as the other descriptors, and 30
bytes are used for each entry in the directory at
the moment, only 1092 descriptors can be created
currently. Each image needs a minimum set of
descriptors, the standard descriptors, which contain
the information needed to interpret the data correctly,
like number of axes, number of pixels in each axis,
etc. See the MIDAS Environment document for more
information about that.
Other descriptors are used by MIDAS applications,
e.g. the descriptors STATISTIC and HISTOGRAM,
to save results from calculations on an image and
therefore have predefined usage. And for the ESO
Archive we are currently implementing hierarchical
descriptors.
Creating, updating and expanding descriptors
is done by the command WRITE/DESCR. To
get rid of descriptors, use DELETE/DESCR.
READ/DESCR and SHOW/DESCR display the
contents of the descriptors and the descriptor
directory. PRINT/DESCR not only prints the
descriptor contents, but may also be used to convert
descriptor data and store it in an ASCII file (using
ASSIGN/PRINT FILE myascii_file before).
And since MIDAS tables also have descriptors,
all these commands apply to tables as well; just
add the file type ‘.tbl' to the table name.
E.g., SHOW/DESCR mytable.tbl will display all
descriptors of the table mytable.tbl.
The COPY/... commands copy descriptors to
keywords and images, and vice versa. Furthermore,
descriptor data may be manipulated directly via
the substitution mechanism provided by the MIDAS
Command Language, e.g.:
COMP/KEY inputd(1) = {myfile,step(1)}
puts the x­step of image myfile.bdf into keyword
inputd(1).
PLOT/DESCR and OVERPLOT/DESCR draw the
contents of a descriptor. All descriptors are saved
in FITS keywords when writing an image on tape via
the OUTTAPE/FITS command.
Soon 1 , you can edit the contents of a descriptor via
the application procedure dscedit.prg.
Normally, all (sensible) descriptors of an input image
are copied to the result image at the end of an
operation on images. This can be changed via the
command SET/MIDAS_SYSTEM DSCCOPY=... 1 .
Finally, we want to draw your attention to
the descriptor HISTORY, where all the MIDAS
commands applied to a given image/table are recorded
automatically. If that is not exhaustive enough for you,
how about adding another descriptor to individually
log what you did?
a 1 from the 92NOV release on
aaa
a a
D ae aa ad a l a
i a
n ae a f a
o
a r
a
t a
h ae a n ae ax a t a
E aS aO a ­ a
M
a I a
D aA aS a C ao au a r a
i a
e a r a
The next issue of the ESO­MIDAS Courier, Volume 2, Number 2, will be prepared and printed in December
1992. The issue will contain information about the MIDAS 92NOV release: upgrades of existing software, as
well as new packages. In order to improve the communication between the MIDAS Group and the MIDAS
users, the latter are encouraged to make contributions to the Courier. Please, send your contributions, preferably
in computer­readable format, to the MIDAS account or to the editor (see the back cover for the mail addresses)
before December 1, 1992.
aa aaa
Newsletter of the ESO­MIDAS Users' Community

14 ESO­MIDAS Courier, Vol. 2, No. 1, June
1992
a a
MIDAS Questions and Answers
The MIDAS Group, European Southern Observatory
aa
This column presents answers to questions which the
MIDAS Group encounters frequently. If you have
your own bag of questions which are asked very often
at your site, or any other suggestion or commentary,
please send them to us.
This time we will discuss some questions concerning
graphics. Because of the upgrade of the underlying
AGL software (see elsewhere in the Courier), the
MIDAS graphics system will be upgraded in the
coming 92NOV release. In that release you will find
additional functionality and more tools to compose
your graphics layout. The upgrade will be mainly on
the command level; the way the graphics system work
(concept) remains unchanged.
The problems discussed here are partly related to
the system as a whole, and partly related to the
functionality of the individual commands. As far as
the concept is concerned, the answers will also apply
to the coming MIDAS releases. Answers to individual
commands will also preview some of the new features
in the new release (just to give you some taste). In the
next issue of the Courier a larger article will discuss
the upgrade of the graphic system.
Problem: I want to plot a line (row or column) of an
image. The plot I get contains a coordinate box but
no data are plotted.
Answer: As in the answer in the last Courier
concerning the command LOAD/IMAGE, the problem
is probably the wrong setting of the cut values. The
graphic system looks at these cut values to determine
the range of coordinate axes. Obviously, with wrong
cut values the coordinate box has a wrong range
and the data values may, partially or completely, fall
outside the box. The solution is to set the cut values
(in the descriptor LHCUTS), correctly yourself, using
the command CUTS/IMAGE. In the previous issue of
the Courier you will find more information about the
descriptor LHCUTS and the meaning of its values.
A second reason might be that you have preset the
graphic axes with the command SET/PLOT 2 . This
will have a similar effect: the data values may fall
outside the defined graphic coordinates and will not
be plotted. A similar situation will occur if you plot
a
2 from the 92NOV release onwards all qualifiers /PLOT will be
replaced by the qualifier /GRAPHIC
table data which range falls outside the preset graphic
area coordinate range.
Problem: How can I make more than one graph on
one sheet of paper (or in one graphics window)?
Answer: Currently, in the 91NOV release, the
possibilities are rather limited. The only way of doing
it is to use the command OVERPLOT/AXES where
you can specify the ranges and sizes of the axes
and the location where the graph is to be drawn.
Subsequently, one can use e.g. the OVERPLOT/ROW,
OVERPLOT/COL, OVERPLOT/TAB commands to
overplot data.
In the coming 92NOV release, tools will be available
to enable easy overplotting. All PLOT commands
can then also be used in overplot mode, i.e. without
erasing the screen first. In addition, offset parameters
can be used to position the graphs.
Problem: Can I use a nice quality font to make
publication quality output?
Answer: Yes, you can, though the possibilities are
limited again. There is only one quality (roman) font
available. This is enabled if you set PMODE=3 in the
command SET/PLOT.
In the new release this restriction will be lifted and
more (six) fonts will be available. The font setting
will then also be decoupled from the layout setting
(PMODE).
Problem: When I have two graphs in the graphics
window I obtain wrong numbers when running the
GET/GCURSOR command on the first graph.
Answer: All essential information of the last graph
drawn , e.g. axes, boundaries, and data, is stored in
the MIDAS keyword structure. Hence, when making
a new graph with an OVERPLOT command, the
information about the first graph is overwritten and
cannot be retrieved when running the graphics cursor.
Similar, when drawing a new graph in another graphic
window, the keywords will only contain information
about that last graph.
Also, in overplotting data on an existing graph the
information, is updated. The values returned by the
graphics cursor will refer to the data plotted last.
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992 15
Problem: When I copy a square graph to the printer
the frame isn't square anymore; its dimensions have
changed completely.
Answer: All graphics output produced is stored in
a plotfile, the metafile. The information in that
file is based on so­called normalized coordinates:
the physical area of the assigned graphics device is
mapped onto a coordinate system that runs from 0,0
for the lower left position to 1,1 for the top right
position. Hence, if one sends a metafile, produced
together with a graph in the graphics window, to
an A4 format laser printer, using the command
SEND/PLOT 3 , the normalized coordinates in the
metafile will now be mapped onto the (new) physical
a
3 in the new 92NOV release the command SEND/PLOT will be
replaced by the command COPY/GRAPHICS
dimensions of the receiving laser printer. Since the
laser printer has a larger output area, the graph will
be stretched in both x and y directions.
To solve the problem one should first do
the assignment of the output graphics device
(ASSIGN/GRAPHICS with possibility the NOSPOOL
option). Then, one can execute the PLOT/... and
OVERPLOT/... commands. Finally, dump the
graph on the graphic device with the SEND/PLOT
command.
Clearly, if you create a graphic window that has the
same dimension ratio as a DIN A4 piece of paper,
sending a square graph to the printer will produce
a square graph. Only the size of the graph will be
scaled, but that scaling will be the same for the x and
y directions.
aaa
Newsletter of the ESO­MIDAS Users' Community

16 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
17
a a
The MIDAS Patch File Summary
(from the anonymous ftp account)
aa
The following text contains the header summary of the patch file 91NOV.0001 that can be obtained by anonymous
ftp (from 134.171.8.4). Please refer to the ESO­MIDAS Courier Vol. 2 Number 1 for a detailed description.
Archive­name: P91NOV.01
Submitted­by: cguirao@eso.hq.org (Carlos Guirao)
This is official patch 01 for MIDAS 91NOV.
Please apply it by:
% cd $MIDASHOME (eg. /midas)
% patch ­N ­p < P91NOV.01
NOTE: "patch" is public domain software that can be very easily installed
in any UNIX machine. There is a compressed tar file called "pub/patch.tar.Z"
available on our anonymous "ftp" account in "ftphost.hq.eso.org"
(134.171.8.4). Follow README for installation details.
NOTE: This patch file should be applied on the original release MIDAS 91NOV.
If you have already modified files by hand,the command "patch" could
fail, leaving the file partially or completely unpatched. A correct
execution should print only "Hunk" messages like:
Hunk # succeeded at .
NOTE: After applying all the patches, you will have to run the "update MIDAS"
procedure, for the modifications to take effect:
% cd $MIDASHOME/91NOV
% ./config
............
Select: 7 (update MIDAS)
...........
Do you want to continue [yn]? (y): y
...........
Select: q (quit config)
Fixes:
­ Creates the file "patchlevel.h" with this header.
­ In Fortran_to_C interface: Temporary strings need to contain
original Fortran strings.
­ Bug fixed for COPY/II with delete flag set.
­ For VMS: Plotting is now a bit faster.
­ AGL: The cursor does not misbehave on Tektronix compatible
devices (only on BSD platforms)
­ Number of points that can be plotted is increased from 10000 to
100000
­ Mathematical library ­lm is needed on IBM6000.
­ Undefined errors when compiling module "ost.c" now solved.
­ Tapeserver returns a proper message when DEVCAPFILE not defined.
Newsletter of the ESO­MIDAS Users' Community

18 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
­ AGL makefile excludes the installation of Versatec commands.
­ Double definitions in "cc.c" are better treated.
­ Some modifications in "osu.c" to deal with SunOs 4.1.1.
­ Corrections to the FITS reader.
­ Corrections to the command COMPUTE/PRECESSION
­ Corrections to the command CREATE/TABLE with a format file.
­ Corrections to the command COMPUTE/BARY & COMPUTE/AIRMASS
­ Corrections to the command COMPUTE/UT & COMPUTE/ST
­ Elimination of two extra planes in the graphic output when plotting.
­ Disabling catching float exceptions (For DecStations)
­ AGL3CONFIG definition has been corrected for Bourne­Shell users.
aaa
Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
19
a a
Problems found in the 91NOV Release
aa
The following text contains a listing of problems found in the 91NOV release. Some of these problems are detected
on both VMS and UNIX systems, some only on one of these two. If you find one of the problems described below,
check with your local MIDAS site manager to get the solutions implemented.
The previous issue of the Courier (Volume 1, Number 2) contains the problems found until December 1991. The
present listing contains the problems found between December 1991 and June 1992.
VMS and UNIX Systems
1. The source code for the perspective plotting will produce two extra planes in the graphics output. The
problem is caused by comments in a number of lines in the subroutine NXTVU in the file plper.for in
the directory /midas/91nov/prim/plot/libsrc (UNIX) or
[midas.91nov.prim.plot.libsrc] (VMS). These comments, those in lines starting with ‘CC',
should be taken out. The line numbers are 545 and 547, 570 and 572, 586 and 588, 634 and 636, 651
and 653, 687 and 689, 704 and 706.
2. A bug was found in the command SKY/ROMAFOT in the context ROMAFOT. In the source file
rfotsky.for in the directory /midas/91NOV/contrib/romafot/src (UNIX) or
[midas.91NOV.contrib.romafot.src] (VMS) the variable AMA is declared as INTEGER. However,
AMA like AMI should have type REAL.
3. In the command GROUP/ROMAFOT in the context ROMAFOT the minimum allowed value for the
background intensity was 1.0. This value caused severe problems for lower background levels. The restriction
was therefore removed. The source file is/midas/91NOV/contrib/romafot/src/rfotgroup.for
(UNIX), or [midas.91NOV.contrib.romafot.src]rfotgroup.for (VMS). The lines 327 to 329
should be commented out.
4. In DAOPHOT an integer overflow may occur due to normalizing if the variable ISEED in the routine
SEED3 is in the range 1±1073741823. The overflow then causes a crash in the routine DAORAN
when computing the value of IDUM. The problem can be fixed by changing the value 1073741823
to 524288 in the file /midas/91NOV/contrib/daophot/libsrc/addstar.for (UNIX), or
[midas.91NOV.contrib.daophot.libsrc]addstar.for (VMS).
5. On systems having not enough memory or swap space the array dimensioning in both ALLSTAR and
DAOPHOT may be too large. One can change the dimensions with the parameter MAXPIC in
/midas/91NOV/contrib/daophot/src/daophot.for (UNIX) or
[midas.91NOV.contrib.daophot.src]daophot.for (VMS),
and the parameter MAXFRM in allstar.for in the same directory.
UNIX Systems
1. In the file /midas/91NOV/libsrc/os/unix/ost.c the following lines should be deleted.
line 100: static jmp_buf env;
line 102: VOID ostalarm();
line 676:#if ((OS_ULTRIX||OS_BSD) && TIMEOUT) /* ALARM function to stop */
line 686:#endif
The next line included:
line 136:VOID ostalarm();
line 137:static jmp_buf env;
Do modify this file according to the above lines, and if you have not yet installed MIDAS, just follow the
installation notes. If you have already installed MIDAS, you will need to update the installation:
Newsletter of the ESO­MIDAS Users' Community

20 ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
% cd /midas/91NOV/install/unix
% config
...........
Select: 7 (update MIDAS)
Do you want to continue [yn]? (y): y
...........
Select: q (quit config)
2. Tapeserver installation (only on UNIX systems). In the file /midas/91NOV/system/tapeserv/inmtaped.c
line 472: l = getdev(name);
should be changed to:
line 472if ( (l = getdev(name)) < 0) {
line 473: P_ERROR;
line 474: (void)ret_client(­1,oserror,osmsg());
line 475: return(­1);
line 476: }
Then run the script: /midas/91NOV/system/unix/makemidas
to generate the /midas/91NOV/system/exec/inmtaped.exeand follow the documentation for a
complete installation.
3. This patch applies only to DecStation/Ultrix running the Fortran compiler f77 release 3.0.2. In the
file /midas/91NOV/libsrc/st/scst.c
line 189: osscatch(SIGFPE,interf);
should be put in comments:
line 189: /* osscatch(SIGFPE,interf); */
In the file /midas/91NOV/local/make_options (after "5 ­ preinstall MIDAS") or
/midas/91NOV/install/unix/systems/decstation/make_options(before "5 ­ preinstall
MIDAS"):
add the line:
F_OPT=­fpe1
Do modify this file according to the above lines, and if you have not yet installed MIDAS, just follow the
installation notes. If you have already installed MIDAS you will need to update the installation:
% cd /midas/91NOV/install/unix
% config
...........
Select: 7 (update MIDAS)
Do you want to continue [yn]? (y): y
...........
Select: q (quit config)
VMS Systems
1. In the context DAOPHOT on some VAX/VMS systems ALLSTAR produces an #access violation#. It appear
that the subroutine WRARRAY in the DAOPHOT library is called with two constants instead of two variables.
To correct the code, include before line the call to WRARRA (line 1488) the lines LX=1 and LY=1, and
change the call to WRARRAY accordingly.
aaa Newsletter of the ESO­MIDAS Users‘ Community

ESO­MIDAS Courier, Vol. 2, No. 1, June 1992
21
a a
ESO­MIDAS TM Request Form
aa
This is a request for the latest release of the ESO­MIDAS 1 system. To obtain a new release, please complete
this form and return it to us (see reverse). Material will be shipped only to users with a valid ESO­MIDAS
User Agreement. If you still have a tape or a tape mailing box from the last release, you MUST return these
before we send the new release.
1. ESO­MIDAS User Agreement No. 2
aaa
2. Technical Support Coordinator
aaa
3. E­Mail
aaa
4. Tape format
VAX/VMS backup format, density: O 1600 bpi O 6250 bpi
Tar format, density: O 1600 bpi O 6250 bpi
O QIC­24 O exabyte 8mm
O DDS/DAT 4mm
O ftp
5. Documentation requested
(one set per site) O MIDAS User Guide 91NOV
O MIDAS Environment Ver. 1.1
O IDI­routines
O AGL Reference Manual
6. Computer facilities
Computer model:aaaaaaa, RAM (Mb):
a, Disk (Mb):aaa
Operating system:
aaaaaaaaaaaaa, Version
aaa
Display systems:
aaa
Graphic terminals:
aaa
Plotters:
aaa
Modem connection: Baud rate
aaaaaaa, Telephone
aaa
Network address: EARN
aaaaaaaSPAN aaa
Internet
aaaaaaaOthers aaa
Date:aaaaaaaaaaaaaaaaaaa Signature:
aaa
a
1 ESO­MIDAS TM is copyright protected software developed by the European Southern Observatory for the purpose of Image Processing of Astronomical
Data.
2 for new user agreements t.b.a.
Newsletter of the ESO­MIDAS Users' Community

22 ESO­MIDAS Courier, Vol. 2, No. 1, June
1992
a a
The ESO­MIDAS Courier is published twice per year (June and December) by the MIDAS Group at the
European Southern Observatory.
Contributions as well as suggestions and comments are invited and can be sent to the editor. In particular,
authors of MIDAS application software that would be of general interest for the MIDAS community, are invited
to make this software available.
If you are not on the mailing list and want to receive future issues of the Courier, contact the editor. More updated
information about MIDAS can be obtained from the MIDAS Bulletin Board, accessible via anonymous ftp.
aa
The ESO±MIDAS TM Courier:
Editor: Rein H. Warmels
Typeset by The Publisher TM
Published by the European Southern Observatory
Karl­Schwarzschild­Strasse 2,
D 8046 Garching bei M˜unchen,
Federal Republic of Germany.
ISSN 1018±3051
The following MIDAS support services are available
for sending suggestions, comments or to obtain help
when problems arise:
Telephone: +49­89­32006456
Telex: 5282822 eso d,
attn. MIDAS Courier
Telefax: +49­89­3202362,
attn. MIDAS Courier
SPAN: ESO::MIDAS
EARN: MIDAS@DGAESO51
Eunet: midas@eso.uucp
Internet: midas@eso.org
anonymous ftp: ftphost.hq.eso.org
(134.171.8.4)
aaa
Newsletter of the ESO­MIDAS Users‘ Community