Report of the IRAF Users' Committee -- March 1996

Report of the IRAF Users' Committee

March 1996

The IRAF Users' Committee (IUC) met at NOAO on February 6. We would like to thank the IRAF staff for their informative presentations. We realize that preparations for the meeting and attending the meeting are both time consuming and we appreciate their efforts.

Current members of the IUC (all of whom were present at the meeting) are:

Jeff Pier (chair) 1993-95  U S  Naval Obs
Peter Eisenhardt  1993-95  JPL        
Andrea Prestwich  1995-97  CfA        
Bill Romanishin   1994-96  Univ  of Oklahoma
Bill Sparks       1994-96  STScI      
Steve Walton      1994-96  Cal State Northridge

Want to get Involved?

Peter Eisenhardt and Jeff Pier are due to rotate off the committee this year. If you are an IRAF user and are interested in serving on the committee, or if you know of someone else who you think would be a good representative of the user community, please contact Jeff Pier or George Jacoby.

A Plea

The IUC is charged with representing the IRAF user community interests in future IRAF developments and in advising the IRAF project on priorities and schedules of releases. In order to be truly representative, we are seeking input from the IRAF user community. PLEASE feel free to contact any committee member and let the member know your concerns about IRAF, your hopes, your frustrations, what you like about it, what you don't like, what you'd like to see it do that it can't, etc., etc.

The whole idea of this process is to make sure that IRAF is responsive to the needs of the community. Your input does make a difference!

Staffing: Kudos and Concerns

The overriding impression of the committee is that the IRAF project is seriously under-staffed. We believe that the problem is becoming critical. 1995 saw a loss of 1-1/2 FTE's from the IRAF staff. Only five years ago there were 9 FTE's. There are now only five full-time IRAF staff members who are supplemented by fractional time from another three NOAO staff members (whose contributions total less than one additional FTE). Meanwhile, the requirements placed upon the staff, as determined by any of a number of metrics (e.g., number of users, number of sites supported, number of packages/tasks, lines of code, complexity of code, number of architectures supported) continue to grow apace. We see no let up.

Upcoming software development demands to support Gemini and the Infrared detectors will have a significant impact upon IRAF staff requirements, and the IUC wants to stress that these projects must provide the IRAF project with additional staffing.

The committee can't help but be disappointed in the release schedule realized for IRAF upgrades and in non-delivery of such enhancements as the X11 Graphical User Interface for science applications. The root cause of this is insufficient staffing, combined with unrealistic expectations of what this staff can reasonably accomplish. The committee is very impressed with the caliber and talent on the staff. They are all extremely hard-working, dedicated, and conscientious. They are performing Herculean tasks and have made significant strides this year (see below). But the committee is concerned that the effort they have already invested in some areas (e.g. on the GUI's) may be wasted, because this work had to be set aside before release to address various urgent demands, and such demands seem to be on the increase. Either additional staff must be added, or promises of what the IRAF group will deliver must be scaled back.

The good news is that this small core of dedicated programmers continues to have a positive outlook. They gave the committee the impression that they love their jobs and eagerly look forward to continuing to provide the astronomical community with quality software to meet ever increasing challenges.

RECOMMENDATION: IRAF staffing needs to be maintained AT LEAST at the level of eight FTE's. We STRONGLY recommend that two additional software positions be filled immediately: one a specialist in systems software, the other in science applications software.

General Comments

IRAF continues to be the most heavily used resource provided and maintained by NOAO, and it is certainly the most cost effective in terms of the service it provides to the astronomical community per dollar expended. Although exact counts are elusive, best estimates are that IRAF is installed at approximately 1500 different sites around the word and that roughly 5000 users see a "cl> " prompt on their computer monitor on a regular basis.

Although IRAF is now well over a decade old, it continues to keep abreast of advancing computer hardware and new astronomical detectors and techniques. IRAF remains at the forefront of software/algorithmic development for astronomical data analysis and reduction.

It is worth noting that a sizeable fraction of the IRAF team's effort is directed not just towards "Reduction and Analysis" of astronomical data, but also towards providing valuable support for instrumentation and for the acquisition and archiving of data. Some examples:

Highlights of 1995

Despite staffing limitations, a rather large number of significant accomplishments were realized in 1995. A few of the highlights:

IUC Priorities for 1996

  1. Release V2.11 The first priority, from the committee's perspective, is to get the new version of IRAF (V2.11) out to the community in the first half of 1996. This release is long overdue! The community has been hearing for some time about some of the nice new features which will be available "in the next release" and we fear that their patience is being tried rather severely. This release will include the FITS image kernel, WCS function drivers, and the SLALIB library.

    The committee is very concerned that if V2.11 is not released before summer, demands upon the IRAF staff to provide support for specifying and developing software for the new CCD Mosaic detector (due for on-telescope engineering tests this summer) will push the V2.11 release efforts to the back burner (AGAIN!).

  2. Platform Support The committee notes that a number of previously "supported" platforms seem to be have been orphaned in recent years. Specifically, AIX, HP-UX, DEC Ultrix, and SGI IRIX have received scant attention from IRAF. Releases of patches and X11IRAF have been made available largely only for the most heavily used IRAF platforms (Suns running both SunOS and Solaris) and newer platforms (DEC Alpha OSF/Digital Unix and PC Linux).

    Meanwhile, a substantial fraction of the user community is still running IRAF essentially as it existed in 1993 or 1994, still dealing with xterm and its hang-up bug while running plot tasks, etc. Furthermore, the plans for release of V2.11 are to port first to Suns, DEC Alphas and PCs, those very platforms which are already most up-to-date in terms of IRAF support. Ports to the other platforms would follow "in a few months time" -- placing this community of users even further behind the times. Of course, we see the logic in providing new releases first to the most heavily used platforms but we hope that the minority of orphaned users will not have to wait much longer.

    We feel that IRAF should make a sincere effort to support platforms which it has led the community to believe it will support. If they are unable to do so, they should announce that support for such-and-such a platform will be dropped -- allowing ample time for users to plan their future computer hardware/software needs well in advance. It is simply not fair to neglect users who bought their present computers at least partly because of IRAF platform support. Perhaps even worse is the perceived uncertainty about continuing support for some platforms for those who are contemplating their next computer upgrade.

              V2.11 should be released for ALL supported platforms.
              X11IRAF should be released for ALL supported platforms.
  3. Open IRAF The IRAF team has long recognized that it cannot be insular. The software framework developed and maintained by IRAF provides a wonderful vehicle for other software developers to take advantage of and to build upon. The amount of software running within the IRAF environment, but developed outside of the IRAF group (though with their help and cooperation) is now nearly as substantial (in terms of lines of code) as IRAF itself.

    The ability to customize/enhance the IRAF environment will become much easier with the advent of "Open IRAF." This initiative will make it easier for a user to (1) invoke individual tasks from a shell, (2) combine IRAF with non-IRAF software, and (3) more easily develop software (in other languages such as Fortran, C or C++) and incorporate it into the local IRAF environment.

    The committee strongly supports the Open IRAF Initiative, and believes that it must proceed rapidly to ensure IRAF's viability in today's rapidly changing computing environment.

  4. Graphical User Interface (GUI) Considerable headway has been made towards making IRAF accessible through GUIs in addition to the traditional command line interface. X11IRAF has been released for some platforms, and provides a very nice image display (XImtool), a significantly enhanced xterm/graphics window (XGterm), a tape monitoring window (XTapemon), and a library/toolkit and "widget server" for developing GUIs.

    A lot of effort has also gone into the development of prototype GUI interfaces to some of the IRAF packages: SPECTOOL, XAPPHOT, XHELP and XRV. Unfortunately, this development effort has had to be put on the back burner because of staffing shortages and higher priority demands. It is a shame that the user community is denied the use of these user friendly interfaces, especially after so much effort went into their development. We urge the staff to complete their initial efforts here and make these nifty interfaces available to the user community.

  5. Management Plan The committee is very favorably impressed with the IRAF management and we feel that significant improvement has been attained in the past couple of years in the area of project management. Nonetheless, the IUC has been pressing for some time to see a realistic, comprehensive management plan. As we pointed out last year, it is very difficult for the committee to assign realistic priorities unless we have some estimates of level-of-effort required for each project. We once again request the IRAF management to present a management plan at next year's meeting.

  6. Science Applications In previous years, reports of this committee have assigned priorities to science application software needed/desired by the community. This year, due to the rather overwhelming constraints of staffing, many of these priorities have received less attention (both by the IRAF staff and by the IUC) than they deserve. We hope that additional staffing will remedy this situation so that higher priorities may once again be assigned to science application software.