United States. Department of Defense

“ By the type of a data object is meant the set of objects themselves, the essential properties of those objects and the set of operations which give access to, and take advantage of, those properties. The author of any correct program in any programming language must, of course, know the type of all data and variables used in his programs. If the program is to be maintainable, modifiable and comprehensible by someone other than its author, the types of variables, operations, and expressions should be easily determined from the source program. ”

“ There is no need, however, that constant expressions impose run time costs for their evaulation. They can be evaluated once at compile time or, if this is inconvenient because of incompatibilities between the host and object machines, the compiler can generate code for their evaluation at load time. In any case, the resulting value should be the same (at least within the stated precision) regardless of the object machine (see D2) . Allowing constant expressions in place of constants can improve the clarity, correctness, and maintainability of programs and does not impose any run time costs. ”

“ Software, particularly the maintenance portion, is becoming much more important than the hardware. Hardware costs are falling at such a rate that the change in technology over just a few months can decrease the effective cost of memory enough to offset any loss in efficiency, while the cost of going to machine code might result in a much longer delay in developing the program. ”

“ When the language contains all the essential power, when few can tell the difference between the base language and library definitions, and when the introduction of new data types and routines does not impact the compiler and the language standards, then there is little incentive to proliferate languages. Similarly, if type definitions are processed entirely at compiler time, and the language allows full program specification of the internal representation, there need be no penalty in run time efficiency for using defined types. ”

“ Language commonality is an essential step in achieving software commonality, but the real benefits accrue when projects and contractors can draw on existing software with assurance that it will be supported, when systems can build from off the shelf components or, at least, with common tools, and when efforts can be spent to expand existing capabilities, rather than building from scratch. Support of common, widely used tools and aids should be provided independent of projects if common software is to be widely used. ”

“ As with the descriptions of digital computer hardware, the semantics and syntax of each feature must be defined precisely and unambiguously. The action of any legal program will be determinable from the program and the language description alone. Any computation which can be described in the language will ultimately draw only on capabilities which are built into the language. No characteristics of the source language will be dependent on the idiosyncrasies of its translators. ”

“ The characteristics of translator implementations will not be dictated by the language definition or standards.
The adoption of a common language is a commitment to the current state of the art for programming languge design for some duration. It does not, however, prevent access to new software and hardware technology, new techniques, and new management strategies which do not impact the source language definition. In particular, innovation should be encouraged in the development of translators for a common language, providing they implement exactly the source language as defined. ”

“ It is not suprising that the implementation is not as powerful as that which could be obtained if 10 or 100 times more effort could be devoted to its development and to the tools and techniques for its use. Such powerful implementations are very good and optimize better than the average programmer. Assuming a very minimal number of well-supported languages for the Department of Defense, one could devote sufficient effort to provide the tools and compilers to produce very efficient object code. ”

“ By key words of the language are meant those symbols and strings which have special meaning in the syntax of programs. They introduce special syntactic forms such as are used for control structures and declarations, or they are used as infix operators, or as some form of parenthesis. Key words will be few in number because each new key word introduces another case in the parsing rules and thereby adds to complexity in understanding the language, and because large numbers of key words inconvenience and complicate the programmer's task of choosing informative identifiers. ”

“ Every translator for the language will implement the entire base language. There will be no subset implementations of the language.
If individual compilers implement only a subset of the language, there is no chance for software commonality. If a translator does not implement the entire language, it cannot give its users access to standard supported libraries, or to applications programs implemented on some other translator. Requiring that the full language be implemented will be expensive only if the base language is large, complex, and nonuniform. ”

“ The ability to initialize variables at the time of their allocation will contribute to program clarity, but a requirement to do so would be an arbitrary and sometimes costly decision to the user. Default initial values, on the other hand, contribute to neither program clarity nor correctness and can be even more costly at run time. It is usually a programming error if a variable is accessed before it is initialized. It is desirable that the translator give a warning when a path between the declaration and use of a variable omits initialization. ”

“ Each of the above points can contribute to program clarity. The use of free format, mnemonic identifiers and conventional forms allows the programmer to use notations which have their familiar meanings to put down his ideas and intentions in the order and form that humans think about them, and to transfer skills he already has to the solution of the problem at hand.
A simple uniform language reduces the number of cases which must be dealt with by anyone using the language. If programs are difficult for the translator to parse, they will be difficult for people. ”

“ One must remember, however, that the coding portion of a large project normally represents only ten or fifteen percent of the total time involved and the promotion of writing must, therefore, be weighted with other portions of the program production cycle. A major fraction of this cycle, sometimes up to half, is involved with debugging and checking and, therefore, the thrust to more reliable code must be stronger.
Specification and design is a similarly large fraction of the effort, and factors such as modularity which simplify the design are also important. ”

“ It is desirable that the number of primitive control structures in the language be minimized, not by reducing the power of the language, but by selecting a small set of composable primitives which can be used to easily build other desired control mechanisms within programs. This means that the capabilities of control mechanisms must be separable so that the user need not pay either program clarity or implementation costs for undesired specialized capabilities. ”

“ When the object machine does have an operating system or executive program, the hardware/operating system combination will be interpreted as defining an abstract machine which acts as the object machine for the translator.
A language definition cannot dictate the architecture of existing object machines, whether defined entirely in hardware or in a hardware/software combination. It can provide a source language representation of all the needed capabilities and attempt to choose these so they have an obvious and efficient translation in the object machines. ”

“ The large systems can support cross compilers which are much more powerful than those self-hosted on a smaller target machine. While this is not yet a universal procedure, it certainly provides a further opportunity to bring to bear tools to promote efficient programs. In such a case, the requirement on the high order language is to give as much information to the compiler as possible without overly restricting its options. ”

“ Translators will continue processing and checking after errors have been found, but should be careful not to generate erroneous messages because of translator confusion. The translator will always produce code to cause a run time exception condition upon any attempt to execute those parts of the program. Warnings will be generated when a source language construct is exceptionally expensive to implement on the specified object machine. A suggested set of diagnostic messages, provided as part of the language definition, contributes to commonality in the implementation and use of the language. ”

“ The language must not just be a combination of existing features which satisfy the individual requirements, but must be held together by a consistent and uniform structure which acts to minimize the number of concepts, consolidates divergent features, and simplifies the whole. ”

“ It should be remembered, however, that interfaces cannot be validated at program assembly time without some equivalent of their source language interface specifications, that object modules are machine-dependent and, therefore, not portable, that source code is often more compact than object code, and that compilers for simple languages can sometimes compile faster than a loader can load from relocatable object programs. ”

“ When a program does not use a feature or capability, it should pay no run time cost for the feature being in the language or library. when the full generality of a feature is not used, only the necessary (reduced) cost should be paid. ”

“ Any good high order language should lend itself to efficient compilers for most any machine. With appropriate linking mechanisms programs from various languages could be put together as required. Unfortunately, the practical world intervenes and the production of good compilers for every language and every machine is not a reasonable expectation, nor is sufficient familiarity on the part of a majority of programmers with a large number of languages. ”

“ The source language will provide encapsulated access to machine dependent hardware facilities, including machine language code insertions.
It is difficult to be enthusiastic about machine language insertions. They defeat the purpose of machine independence, constrain the implementation techniques, complicate the diagnostics, impair the safety of type checking, and detract from the reliability, readability, and modifiability of programs. ”