LooPy: interactive program synthesis with control structures

Live programming is an activity in which the programmer edits code while observing the result of the program. It has been exercised mainly for pedagogical and artistic purposes, where outputs of a program are not straightforwardly imagined. While most live programming environments so far target programs that explicitly generate visual or acoustic outputs, we believe that live programming is also useful for data structure programming, where the programmer often has a hard time to grasp a behavior of programs. However, it is not clear what features a live programming environment should provide for such kind of programs. In this paper, we present a design of live programming environment for data structure programming, identify the problems of synchronization and mental map preservation, and propose solutions based on a calling-context sensitive identification technique. We implemented a live programming environment called Kanon, and tested with 13 programmers.

Programming language definitions assign formal meaning to complete programs. Programmers, however, spend a substantial amount of time interacting with incomplete programs - programs with holes, type inconsistencies and binding inconsistencies - using tools like program editors and live programming environments (which interleave editing and evaluation). Semanticists have done comparatively little to formally characterize (1) the static and dynamic semantics of incomplete programs; (2) the actions available to programmers as they edit and inspect incomplete programs; and (3) the behavior of editor services that suggest likely edit actions to the programmer based on semantic information extracted from the incomplete program being edited, and from programs that the system has encountered in the past. This paper serves as a vision statement for a research program that seeks to develop these missing semantic foundations. Our hope is that these contributions, which will take the form of a series of simple formal calculi equipped with a tractable metatheory, will guide the design of a variety of current and future interactive programming tools, much as various lambda calculi have guided modern language designs. Our own research will apply these principles in the design of Hazel, an experimental live lab notebook programming environment designed for data science tasks. We plan to co-design the Hazel language with the editor so that we can explore concepts such as edit-time semantic conflict resolution mechanisms and mechanisms that allow library providers to install library-specific editor services.

Data transformation is a laborious and time-consuming task for analysts. Programming by example (PBE) is a technique that can simplify this difficult task for data analysts by automatically generating programs for data transformation. Most of the previously proposed PBE methods are based on search algorithms, but recent improvements in machine learning (ML) have led to its application in PBE research. For example, RobustFill was proposed as an ML-based PBE method for string transformation by using long short-term memory (LSTM) as the sequential encoder–decoder model. However, an ML-based PBE method has not been developed for tabular transformations, which are used frequently in data analysis. Thus, in the present study, we propose an ML-based PBE method for tabular transformations. First, we consider the features of tabular transformations, which are more complex and data intensive than string transformations, and propose a new ML-based PBE method using the state-of-the-art Transformer sequential encoder–decoder model. To our knowledge, this is the first ML-based PBE method for tabular transformations. We also propose two decoding methods comprising multistep beam search and program validation-beam search, which are optimized for program generation, and thus generate correct programs with higher accuracy. Our evaluation results demonstrated that the Transformer-based PBE model performed much better than LSTM-based PBE when applied to tabular transformations. Furthermore, the Transformer-based model with the proposed decoding method performed better than the conventional PBE model using the search-based method.

We present a method for example-guided synthesis of functional programs over recursive data structures. Given a set of input-output examples, our method synthesizes a program in a functional language with higher-order combinators like map and fold. The synthesized program is guaranteed to be the simplest program in the language to fit the examples. Our approach combines three technical ideas: inductive generalization, deduction, and enumerative search. First, we generalize the input-output examples into hypotheses about the structure of the target program. For each hypothesis, we use deduction to infer new input/output examples for the missing subexpressions. This leads to a new subproblem where the goal is to synthesize expressions within each hypothesis. Since not every hypothesis can be realized into a program that fits the examples, we use a combination of best-first enumeration and deduction to search for a hypothesis that meets our needs. We have implemented our method in a tool called λ2, and we evaluate this tool on a large set of synthesis problems involving lists, trees, and nested data structures. The experiments demonstrate the scalability and broad scope of λ2. A highlight is the synthesis of a program believed to be the world's earliest functional pearl.

Several studies provide the questions developers ask during software evolution tasks, providing foundations for subsequent work. Nevertheless, none of them focus on Live Programming environments that gain in popularity as they are perceived to have a positive effect on programming tasks. Studying the impact of a Live Programming environment on software development activities is thus the goal of this study. In a partial replication of the study by Sillito et al., we conducted 17 software evolution sessions in a Live Programming environment and report 1,161 developer questions asked during these sessions. We contrast our results with the results by Sillito et al., focusing on the question occurrences, question complexity and what information participants used to gain a required knowledge. We report eight new questions and observe that the Live Programming facilities do have an impact on the way developers ask questions about source code and use tools to gain corresponding knowledge.

Learning how to program can be quite difficult. Beginners face two obstacles to learning to program: developing a mental model of the steps of computation and learning the syntax of a programming language. Because programmers are prone to make mistakes that introduce errors into their code, it is important that they understand what each line of code is doing in their program. I propose that a system be built in which writing code and running code happen simultaneously. This system shows the results of code to the user immediately after it is typed; this type of system is called a live programming environment. The environment's features allow the programmer to highlight individual lines of code to see their result. The programmer can walk through loops and explore the details of a function call. The programmer can unit test a function or class. The environment displays variables declared in the code and their values and allows the programmer to create a visualization for class objects. The environment appears to be beneficial for beginners based on the results of two surveys with a total of 27 participants. The results of this thesis show that live programming environments should be developed to help the beginner learn to program.

The software pipeline technique is an effective approach to optimizing loops in array processor programs, but existing methods are of high complexity and the results may not be satisfactory.This paper introduces the URPR algorithm, an extension of the microcode loop compaction algorithm URCR. Firstly, unroll the loop(the number of unrolled loop bodies relies on the inter-body data dependency); secondly, pipeline the unrolled loop bodies one by one; and finally, a new optimized loop body is obtained after rerolling. Preliminary tests indicate that URPR produces favorable results with lower complexity.

Program Synthesis, which is the task of discovering programs that realize user intent, can be useful in several scenarios: enabling people with no programming background to develop utility programs, helping regular programmers automatically discover tricky/mundane details, program understanding, discovery of new algorithms, and even teaching.

While more and more data are available through web services, it remains difficult for end-users to create web applications that make use of these data without having to write complex code. We present Gneiss, a live programming environment that extends the spreadsheet metaphor to support creating interactive web applications that dynamically use local or web data from multiple sources. Gneiss closely integrates a spreadsheet editor with a web interface builder to let users demonstrate bindings between properties of web GUI elements and cells in the spreadsheet while working with real web service data. The spreadsheet editor provides two-way connections to web services, to both visualize and retrieve different data based on the user input in the web interface. Gneiss achieves rich interactivity without the need for event-based programming by extending the 'pull model' of formulas that is familiar to the spreadsheet users. We use a series of examples to demonstrate Gneiss's ability to create a variety of interactive web data applications.

As a culture, object-orientation encourages programmers to create objects, both short- and long-lived, without concern for cost. Excessive object creation and initialization can cause severe runtime bloat, which degrades significantly application performance and scalability. A frequently-occurring coding pattern that may lead to large volumes of (temporary) objects is the creation of objects that, while allocated per loop iteration, contain values independent of specific iterations. Finding these objects and moving them out of loops requires sophisticated interprocedural analysis, a task that is difficult for traditional dataflow analyses such as loop-invariant code motion to accomplish. Our work targets data structures that are loop-invariant, and presents a static type and effect system to detect loop-invariant data structures. For each loop, our analysis inspects each logical data structure in order to find those that have disjoint instances per loop iteration and contain loop-invariant data. Instead of automatically hoisting them to improve performance (which is over-conservative), we report hoistability measurements for each disjoint loop data structure detected by our analysis. Eventually these data structures are ranked based on these measurements and are presented to the user to help manual tuning. We have performed a variety of studies on a set of 19 moderate/large-sized Java benchmarks. With the help of hoistability measurements, we found optimization opportunities in most of the programs that we inspected and achieved significant performance improvements in some of them (e.g., 82.1% running time reduction).

We look at program synthesis where the aim is to automatically synthesize a controller that operates on data structures and from which a concrete program can be easily derived. We do not aim at a fully-automatic process or tool that produces a program meeting a given specification of the program’s behaviour. Rather, we aim at the design of a clear and well-founded approach for supporting programmers at the design and implementation phases. Concretely, we first show that a program synthesis task can be modeled as a generalized planning problem. This is done at an abstraction level where the involved data structures are seen as black-boxes that can be interfaced with actions and observations, the first corresponding to the operations and the second to the queries provided by the data structure. The abstraction level is high enough to capture intuitive and common assumptions as well as general and simple strategies used by programmers, and yet it contains sufficient structure to support the automated generation of concrete solutions (in the form of controllers). From such controllers and the use of standard data structures, an actual program in a general language like C++ or Python can be easily obtained. Then, we discuss how the resulting generalized planning problem can be reduced to an LTL synthesis problem, thus making available any LTL synthesis engine for obtaining the controllers. We illustrate the effectiveness of the approach on a series of examples.

Current techniques for prefetching linked data structures (LDS) exploit the work available in one loop iteration or recursive call to overlap pointer chasing latency. Jump pointers, which provide direct access to non-adjacent nodes, can be used for prefetching when loop and recursive procedure bodies are small and do not have sufficient work to overlap a long latency. This paper describes a framework for jump-pointer prefetching (JPP) that supports four prefetching idioms: queue, full, chain, and root jumping and three implementations: software-only, hardware-only, and a cooperative software/hardware technique. On a suite of pointer intensive programs, jump pointer prefetching reduces memory stall time by 72% for software, 83% for cooperative and 55% for hardware, producing speedups of 15%, 20% and 22% respectively.

Polikarpova et al. have recently proposed a method for synthesizing functional programs from specifications expressed as refinement types, and implemented a program synthesis tool Synquid. Although Synquid can generate non-trivial programs on various data structures such as lists and binary search trees, it cannot automatically generate programs that require auxiliary functions, unless users provide the specifications of auxiliary functions. We propose an extension of Synquid to enable automatic synthesis of programs with auxiliary functions. The idea is to prepare a template of the target function containing unknown auxiliary functions, infer the types of auxiliary functions, and then use Synquid to synthesize the auxiliary functions. We have implemented a program synthesizer based on our method, and confirmed through experiments that our method can synthesize several programs with auxiliary functions, which Synquid is unable to automatically synthesize.

Pipelining allows the execution of loop iterations with cross-iteration dependences to overlap in time, provided that the loop body is partitioned into stages such that the data dependences are not violated. Then, the stages are mapped onto threads and communication and synchronization between stages is typically achieved using queues. Pipelining techniques that rely on static scheduling perform poorly for load-imbalanced loops. Moreover, previous research efforts that achieve load-balancing are restricted to work-stealing and imply high overhead for fine-grained loops. In this article, we present URTS, a unified runtime system with compiler support that provides a lightweight dynamic scheduler by combining mapping, communication and synchronization algorithms with a suitable data structure and an efficient ticket mechanism. Particularly, URTS shows that it is possible to combine the efficiency of static scheduling with the load-imbalance tolerance of work-stealing by using a unified design that exploits the properties of a novel data structure. The evaluation on 8- and 32-core machines shows that URTS implies low overhead, of the same order as a static scheduler, for a set of benchmarks chosen from widely-used collections. URTS is a scalable solution that performs efficient dynamic scheduling for fine-grained loops, i.e., a class of interesting loops that is poorly handled by the state-of-the-art due to high overhead.

This paper presents an algorithm for synthesizing recursive functions that process algebraic datatypes. It is founded on proof-theoretic techniques that exploit both type information and input–output examples to prune the search space. The algorithm uses refinement trees, a data structure that succinctly represents constraints on the shape of generated code. We evaluate the algorithm by using a prototype implementation to synthesize more than 40 benchmarks and several non-trivial larger examples. Our results demonstrate that the approach meets or outperforms the state-of-the-art for this domain, in terms of synthesis time or attainable size of the generated programs.