I plan to discuss a subject that everyone has had experience with, and apparently thoroughly understands. I will discuss the subject anyway. My subject could be called Engineering, Design, Optimi-zation, or the game of life. It is a completely open-ended subject - no beginning, no end, no simple solutions, many paradoxes.

The general form of the panel discussion was to consider several topics regarding the gap, or lack of communication, between the systems engineer and the lens designer. The first topic was that of trying to define the gap between the lens designer and the engineer and the nature of some of the problems. In this review, rather than try to ascribe a particular comment to an individual, we will simply summarize some of the pertinent points that were brought out.

A new medical instrument has been developed, which combines a surgical optical microscope with television and stereoscopy to provide a Surgical Stereo-Video Microscope. This instrument has many advantages over previous methods of viewing surgery and also provides an excellent means of medical education.

Solar energy is attracting considerable national attention as a potential new option in the search for energy alternatives for the future. The dream of the use of solar energy is not new. Since solar energy interest has grown and waned in the past one immediately wonders what might be new this time that would lead to success. Do the problems encountered in the past still hold today? If not, what is new that could make solar energy a practical reality? I would like to address this question in some detail under the title of this session: Optical Interfaces; since there obviously is a very important role for optical technology in solar energy, ranging from optical collectors to solid-state physics, but first, some background should be presented.

Astrometry deals with the space-time behavior of celestial bodies. Important areas within this classical field of astronomy include the determination of stellar distances and motions, studies of star clusters, and construction of photographic star catalogs.

I was asked to discuss interface problems related to the design of optical equipment for the Army as seen from the designer's point of view. Before getting involved in this discussion, I think it appropriate to provide some background information regarding the type work we are involved in at Frankford Arsenal.

The Los Alamos Scientific Labora tory (LASL) has participated in five airborne total solar eclipse expedi tions designed to measure the temperature, sensity, structure, motions, and composition of the sun's corona. This information is being gathered as part of a study of the sun's behavior throughout its eleven, year cycle. The advantage of these airborne observations over earthbound ones is that they are made above eighty percent of the atmosphere and its clouds and water vapor. The consequent reduction in atmospheric turbulence, scattered light, and water-vapor absorption greatly improves visibility. Because the plane moves with the eclipse shadow, the time available for making the observations is up to one half longer than a ground observer's time.

A camera systems design and optimization computer program has recently been developed (Ref. 1). At present the program is set up to search for camera systems which operate over maximum altitude range, have minimum volume, and still achieve some specified ground resolution. The program can be easily modified to optimize most any other parameter, including cost, or to maximize performance under a volume constraint, etc.

A successful beryllium reflector applicable to land-, sea-, air- and spaceborne surveillance systems is the culmination of a joint effort between the Customer and the Manufacturer. It involves interfacing contributions from Engineers in a number of disciplines such as systems, optics, stress analysis and, most importantly, machining. It requires the knowledge of the characteristics and the experience in the special handling technology of beryllium. It demands that the Manufacturer has an appreciation of the End-User's system requirements and that the Customer has an awareness of the Producer's expertise and potential in order to achieve a feasible production design specification of a metal optic. The average mirror density within the envelope of a mass relieved beryllium optic is about 0.6 g cm^-3. The essential steps from concept through design analysis and manufacture of a typical stiff, beryllium, low inertia, lightweight 1 kilogram (2.2 lb.), active plane mirror 35.6 cm x 21.6 cm x 2.7 cm (14" x 8-1/2" x 1-1/16") are presented.

The development of an optical system to be used with the imaging science payload of an interplanetary space probe requires that the optical engineer reconcile two rather diverse sets of requirements. One set is provided by the Imaging Science Team - the ultimate customer and represents the objectives and goals for which reason the mission is undertaken. The second set is provided by the Engineering Team - the ultimate employer and represents the practical constraints of the people responsible for the successful achievement of this mission objectives.

In designing optical systems, most designers at some point are con-fronted with the problem of specifying optical components that lie outside their area of expertise. The common practice in these circumstances has been to specify the "best" (i.e. most expensive) component. This attitude often leads to consid-erable waste of time, effort and money. We illustrate our point of view by considering in detail many of the common examples of unduely stringent and often inconsistent specifications for optical components. Some of the examples are: Specification of 1/200th wave flat mirrors for Fabry-Perot interferometers without consideration to other limiting deformities caused by coating, mounting, etc. Specifications calling for bulky prism systems (usually made of high quality quartz) when inexpensive mirrors could do the same job. Specification of interferometric quality glass for optical windows with quarter-wave flatness and parallelism, when selected float - glass could be used at much reduced cost. Specification of an expensive laser collimator and spatial filter for use in holography, when $50 worth of surplus optics plus a pinhole would do as well. Specification of a high degree of mechanical parallelism between faces of optical components without consideration of optical parallelism (as in large discs of laser glass). The discussion presented also includes a number of techniques in design and specification of optical components which can ultimately result in improved cost effectiveness.

In order to obtain the lens you want, or to design the lens desired by a customer, whichever end you're on, the lens must be thoroughly specified. Not only that, but the specifications must be understood the same by all parties involved. Unfortunately, even though most terminology related to lens specifications and tolerances is well defined, there is still much ambiguity and lack of care in the use of these terms. In addition there are many terms that are not universally used with a single meaning. This paper discusses some of the ambiguities and inconsistencies, with the aim of emphasizing the need for adequate definitions and standardized usage of optical terminology.

One observation one might readily make about some optical systems of recent manufacture is that their imagery performance is appreciably superior to that obtainable from outwardly similar designs originating in the 1960's. Many camera buffs for example, have had the recent experience of acquiring a varifocal (zoom) lens which replaces (and possibly out-performs) anywhere from two to four of the fixed focus lenses purchased only a few years ago. Though it is not appropriate for the purposes of this presentation to delve into the specifics of the pertinent performance criteria, it is clear to most lens designers that the improvements in performance have oc-curred for several reasons. In addition to the fact that the level of sophistication of the average designer has probably increased, the software currently available is certainly more versatile and dependable than that of years past. It is also true that many modern optical systems are sufficiently complex (and expensive) that designers are, in some cases, becoming less reluctant to take advantage of the extra degrees of freedom afforded through use of the more exotic glasses, and even aspheric surfaces. Ever increasing availability of awesome computing power at reasonable prices has probably been at the real root of the dramatic performance improvements, however. It is probably safe to guess that the average lens design specialist today employs roughly ten times the computing leverage as might have been applied only five to seven years ago.

The development of an optical instrument involves many technological and project trade-offs. These become particularly important to a scientist, planning an experiment for a space mission. Typically, the scientist customer is accustomed to making the technological trade studies, but the wide variety of project options is a confusing subject. This review of some of the major options and their interaction with the engineering aspects of an instrument development program, should provide new insight for the optical customer.

Itek has worked on four major NASA programs. I have personally worked on two of the programs as the system optical engineer I am going to describe some of the work, keynoting "co-munication", one important aspect of how the optical engineer works with the rest of the project team.

During the last decade the United States and others have orbited a variety of comparatively small exploratory instruments intended for limited use in astronomy. These instruments have demonstrated beyond any doubt the critical advantages of the space observatory working above the earth's atmos-phere they have the capability for much increased spectral range and angular resolution.

General performance specifications are described for photographic and electrooptical multispectral cameras to be used for remote sensing purposes. The effect of natural limitations on the specification for the smallest reflectance difference detectable is discussed. It is shown that reflectance differences of 0.01 can be detected from orbit by a nearly diffraction-limited P3.5 system of 1-m focal length in a spectral interval from 600 to 700 nm and with an effective instantaneous field of view of 30 m, using a solid-state array of detectors having a noise equivalent signal of 1 μJ/m².

SEOS is an acronym for the Synchronous Earth Observation Satellite. As a geostationary system, it provides unique possibilities for Earth Surveillance - it is ideally placed for monitoring short lived and dynamic phenomena over any and all areas on a near hemispheric size field. This implies the capability for continuous monitoring, timely observations between periods of cloud cover and signatures associated with sun angles. For SEOS, applications (missions) are primarily in the areas of Earth Resources and Meterological Disaster warning "Applications" as used herein refer to missions with eco-nomic and/or social benefits.

Five years ago the performance of discrete detector sensors with which the writer had contact was specified by setting essentially an acceptable noise-equivalent target (NET), line-of-sight (LOS) accuracy, frame-time (FT), and data rate (DR). In addition, target intensity-time profiles, back-ground data, and other available information were given. Often, object detection probability, false track rate, tracking accuracy, object reporting time, and other specifications were set as well. It will be noted that the above specifications are more than a minimum complete set. Thus, there was a potential for inconsistency. More important, the mission specifications (principally object detection probability, false track rate, tracking accuracy, and object reporting time) were sometimes missed even though NET and FT values were met because the object-intensity and the background models used to connect them with NET, LOS accuracy, and FT were inaccurate. The sensor (or its subsys-tems) was either overspecified, making it overly expensive, or it was underspecified, making it in-capable of performing its mission adequately. No criticism is intended for the workers who made specifications for previous programs. They did the best they could with the information they had available. However, it was clear at that time that a better procedure was needed if optimum system performance was to be achieved for minimum cost.

Spaceborne sensors for today's technology requirements must meet stringent specifications with regard to faint object detection and observation. The optical designer is faced with the challenge of furnishing a design that is fast and well corrected, has wide field coverage, and can be used to view faint objects when bright sources, such as the sun or the sunlit earth, are just out of the sensor field of view.

Early in November 1973, the Mariner Venus/Mercury spacecraft was launched from Complex 36B at Cape Kennedy on top of an Atlas /Centaur vehicle. It was the first dual-planet mission, the first spacecraft to explore Mercury, the first close-up television observation of Venus, and the first spacecraft to approach the Sun so closely. Along with two television cameras, six scientific instruments returned interplanetary and planetary data during Mariner's five-month primary mission.

The man who creates TV cameras for space is bad news to the lens de-signer. He asks for a high quality lens or telescope with painful design constraints, and can only offer a need for very low volume production. On top of that, he wants a lens tommorrow, not next week, because he is faced with an unslippable spacecraft schedule. The requirements vary with the type of mission and kind of space-craft, but invariably impose physical or optical difficulties beyond those normally experienced in lens design. The purpose of this paper is to discuss these constraints as related to typical types of space missions, the kind of optical design problems they have created, and what can be done to arrive at an acceptable interface between optical designer and camera design engineer.

Image quality in periscopes is generally limited by the aberrations of field curvature and secondary spectrum. It will be shown how these aberrations are related to the other optical and physical parameters of a periscope so that effective design trade-offs can be made. Specifically, this paper examines the relationship of secondary spectrum and field curvature to periscope length and diameter and their relationship to the relay lens system configuration, field of view, relative aperture, and image relative illumination.

Specification and evaluation of electro-optical systems which detect and locate moving weak point sources in the presence of background noise are the subjects of this paper. Our intent is to present a design procedure for a large aperture sensor utilizing several discrete detectors. A step-by-step design procedure will be discussed, with emphasis on the optics and the focal plane detectors. Much of the procedure has been implemented by automatic computation, but some is left to experienced judgement.

In choosing a configuration for an optical image-forming system, one is usually given several prescribed parameters, such as focal length, field, aperture, or magnification, and the required performance. The required performance is usually specified as some tolerance on a quality factor which ultimately depends on the aberrations of the system. In this paper we deal only with simple systems where the performance is limited by third-order aberrations, although the approach outlined is not restricted to such systems.

Once the requirements for a diffraction-limited imaging optical system have been established, the aperture needed to permit sufficient energy to reach the image and the focal length required to meet the specified resolution can be determined. As anyone who has spent even a little time with the problems of optical design knows, this upper bound can be achieved in only the rarest of circumstances. The real system must be manufactured to a practical set of fabrication tolerances. The design, therefore, must be sufficiently better than the requirements so the realities of execution will permit a system to be fabricated and assembled to yield satisfactory performance. This paper describes a procedure whereby the designer can not only predict the average performance of a proposed system, but can also establish some bounds on its expected performance.