Introduction

For most of the last few years the efforts of the JCMT software group have been fairly tightly focussed on the on-line control and data-acquisition systems, and particularly on development on the Vax computers. However, in the last year or so our focus has begun to widen. The installation of Unix systems, which began in 1992, is now well underway. With increased software support, we now have the ability to address other areas, such as off-line utilities and enhanced data reduction facilities. Although the basic framework for the on-line systems is now relatively stable, it's limitations are becoming more obvious, and we are starting to consider some more major design changes in order to provide a more flexible and efficient system.

Individual areas which will have specific impact on Users are described in the following articles. The purpose of this article is to give a broad overview of the main areas of work currently being undertaken by the JCMT software group, and a general indication of our plans for the future.

Main Areas of Work

The efforts of the group falls into four main areas:

* continued development of the on-line system (support for new instrumentation, support for engineering work, new observing modes).

* development of facilities for visiting astronomers, including support of utilities and `wishlist' items and on-site data reduction.

* investigations and developments of new systems (e.g. the database).

* maintenance (bug fixes, ADAM and Starlink Upgrades, etc).

The distribution of effort between these categories is such that approximately half our effort goes into continued development of the existing system, approximately 10% into utilities and wishlist items, and around 30% into new systems. The number of known bugs in the on-line system is now rather small but new ones continue to surface from time to time. These, together with the usual maintenance requirements takes the remainder of the efforts of the group.

The following sections explore the three main areas in more detail.

Continued Development of the On-line System

As noted, this continues to take the bulk of the effort of the software group. The obvious major item is preparation for new instrumentation - RxB3, RxW and SCUBA. JAC staff are working closely with each of the instrument groups to ensure that the integration of these instruments into the JCMT control system goes as smoothly as possible. An area which has grown in importance as the telescope matures is to provide computing support for new engineering projects. For example, software development has been required for the acquisition and reduction of carousel strain gauge data, and we are now working with astronomers and engineers on the new inclinometry system. Support for the instrument group has also increased in the past year or so; with projects such as improved ICL procedures to simplify the measurement of receiver efficiencies. Interferometry is another project which has required considerable software effort.

Finally, we are continuing to make significant developments to the observation control system. A prototype scheme for heterodyne rastering with the DAS is under development. We plan to make more general improvements to the DAS software (improved calibration techniques, more flexible configuration software) in the fall, in collaboration with Hans van Someren Greve, who will be visiting from Dwingeloo. We also continue to investigate methods of reducing the observing overheads. We have recently moved a Vaxstation 4000/60 to the summit, to take some of the general system overheads off MWTVAX, and we are considering replacing MWTVAX with a model 4000, if the work required to install the appropriate SCSI interfaces is not prohibitive.

Facilities for Visiting Astronomers

Apart from actually improving the software involved in the real-time data acquisition, we are also trying to improve the ancilliary service provided to visiting astronomers.

One of the most visible manifestations of this has been the provision of Unix workstations at the summit, at Hale Pohaku and in Hilo. These facilities are briefly described in a following article by Remo Tilanus. As noted by Remo, the Unix versions of SPECX and JCMTDR should be available soon; many of the most popular alternative data reduction packages are also available (with data input via FITS). Although CLASS might be the most common alternative to SPECX, we have already had visiting astronomers reduce their data with IRAF - complete with graphical displays labelled in Angstroms! Along with the introduction of the Unix systems, we have installed an improved guest account allocation procedure, which will give visiting astronomers access to a private account accessible at all sites on both Unix and VMS systems.

We are also starting to devote some time to software items which do not have a major impact on the observing system, but which make life a little bit easier or simpler for visiting astronomers - the `wishlist' and improved support for `utilities'. Wishes are simple upgrades to the on-line system which do not require a serious specification or project plan. Utilities on the other hand are generally off-line packages which have been written outside the software group (either by local support staff or other establishments) and which we are now helping to support.

As a starting point for the wishlist, we have collected all the outstanding requests from local support staff, along with input from visiting astronomers into a single document - `The software wishlist' (MTUN/154). This summarises existing requests, and their current status - either in progress, pending due to lack of effort, or rejected as being either too ambitious or not considered to be worth the effort. This latter determination has of course been made after consultation with support scientists, and we will continue to have regular meetings of the JCMT science group to assess the relative merits of proposed work. Examples of recently or almost completed `wishes' include the addition of phase and transmission data to UKT14 data-files, and the provision of spectral-line fivepoints. Users who would like to look at the current wishlist should use anonymous FTP to get the file

\tt/pub/jac\_docs/jcmt\_docs/mtun154.tex

As noted in MTUN/154, visiting astronomers are encouraged to e-mail to Mary Fuka any requests which they might have for these sorts of small improvements.

Until recently, the utilities were not supported by the software group. Although we certainly don't wish to imply that only members of the software group are competent to write software, there have been problems in the past with multiple divergent copies of the same item, and with packages being `broken' either by upgrades to the on-line system or simply the genuine passage of time (e.g. changing ephemerides). Starting with the most important items, our plan is to move the software into a single central location, identical over sites and as far as possible operating systems, and clearly identify the support scientist(s) and member of the software group who will take responsibility for continued development and correct installation.

Finally, in all of this work we hope to significantly improve the documentation available to users. We will be working closely with Henry Matthews in this area to make documentation more accessible (both physically and in terms of content). We expect to make much of this documentation available on-line via the World Wide Web, and encourage users to explore this system.

Recent and Proposed New Developments

One of the largest new developments which has come to fruition in the last year has been the installation of Unix systems in Hilo, at JACH and at the summit. Although these are currently being used mainly for data reduction and new applications such as the database, we do plan to slowly migrate the real-time systems away from the Vaxes. The first instance of this will be SCUBA, which will use the Unix version of ADAM.

The relational database management system itself is described by Remo Tilanus in an accompanying article. As well as forming the basis for the astronomical data archive, this will provide us with powerful tools to manage instrument and other engineering data.

We have in the last few months started considering some more major re-designs of the on-line systems. Although we now have a stable, relatively bug-free and reasonably flexible implementation of the observing schemes as first envisaged 5-10 years ago, as the system has matured this very process has caused it's shortcomings to become more obvious.

We are continuing to modify the observation control task (CONTROL) to provide additional observing modes and improved efficiency, but it is clear that CONTROL is nearing the end of its useful life. One of our primary goals therefore is to replace this task (and associated ICL procedures) with a more efficient and flexible system which also provides the additional features missing from the current system. An upgrade or replacement for the existing GSD data-storage system, which is also failing to meet current demands, will probably be done at the same time.

Apart from the CONTROL task, at a lower level we wish to upgrade the software associated with controlling the telescope. The actual telescope servo is closed in a PDP which is unique at the JCMT, is poorly documented and contains custom interfaces. The fact that it has performed as well as it has is a tribute to the original design, but it really needs to be replaced by a more maintainable system. At the same time, control of a lot of the the existing hard-wired analog electronics (interlock monitoring systems and so forth) could be moved into the new telescope computer. This will allow us to implement some much-requested improvements to the operator interface. Finally, we would move some of the near real-time processing currently performed by the Vax into this telescope computer.

SCUBA will be delivered as a Unix ADAM system, and will come with a user interface based on a queueing system executing pre-defined observation definitions . Although initially this system will be separate from the heterodyne system, we will be working towards providing a consistent interface to the two systems. The current JCMT instrument development programme includes the delivery of a heterodyne array receiver (initially 16 elements) and matching multiple-input DAS in 1998. This will have control requirements far in excess of the existing heterodyne equipment. Finally, we wish to generally improve the user interface, and particularly improve error reporting and logging facilities.

In performing these upgrades, we will be constrained by the requirement that any new system will have to interface to a significant amount of existing software; we are certainly not going to re-implement everything from scratch. However, we are investigating whether there are tools other than, or in addition to ADAM which it may be appropriate to use. For example, the Tcl command language, and its graphical interface Tk has become a widespread de facto standard in the scientific community. It has already been interfaced to the ADAM message system at ROE and may well be used for the SCUBA queueing system.

A second possibility that we are considering is the use of the AAO real-time system, DRAMA, for new developments to the control system. Developed at the AAO as a result of experience gained with ADAM, it is portable across operating systems, has an efficient message system, and interfaces well to modern standards such as X and Tcl/Tk. An enormous attraction of DRAMA is it's ability to interface to an existing ADAM system, both from the point of view of exchanging messages, and in terms of the similar paradigm. The possibility that the appropriate parts of DRAMA might be easily ported to run under OS/9 is something which we intend to investigate further.

Given the potential impact of Gemini and UKIRT development at the JAC, we have not ruled out the possibility of using EPICS for parts of the system. Once again, this is a package which is generating widespread interest in the astronomical community. It might be particularly attractive for the `PDP replacement' for example where it would be well suited to the monitoring of motor drive currents, antenna leg temperatures, general weather parameters, emergency stop interlocks and so on. It is not clear yet however whether the benefits of EPICS outweigh it's costs, especially when a GUI competitive to that provided by EPICS exists in TCL/Tk.

Final Remarks

The requirements for new instrumentation are rather well defined and efforts to prepare for these will be given the highest priority. Our plans for the support of utilities and wishlist items are also well in hand. The route forward on major new upgrades is somewhat less clear - for example, we are currently awaiting with interest the outcome of the ADAM Review, and are also monitoring closely the changing plans of the Gemini Controls Group. At the moment we are using the current system along with a variety of other input to provide the baseline specification for future upgrades, and are still exploring the potential of new tools such as TCL/Tk. However, we hope to start work in earnest on some of these developments in the next six months.

I hope that this article has given a flavour of where we are going. If anyone would like more information, or has any comments or suggestions they should feel free to contact me.

Richard Prestage, JAC