In 2017, the Fifth eCAADe1 Regional International Symposium (RIS) was held at the Welsh School of Architecture at Cardiff University, co-chaired by Dr Wassim Jabi and I. The symposium is part of a series supplementing the eCAADe annual conferences which promote and are concerned with computation in the realm of design, architecture and space in education and practice. Like the annual conferences, the Regional International Symposia rotate between various Universities within Europe but have rather a workshop character in a more familiar setting. Their results are typically published as a Workshop Proceedings Booklet and are systematically catalogued on CUMINCAD.2 During this fifth instance at Cardiff University, titled ‘The virtual and the physical’ – homonymous to this special issue – European researchers and academics came together to present their latest research discussing issues ‘between the representation of space and the making of space’. As the symposium title suggests the contributions at the symposium explored a wide territory spanning from investigations of virtual spaces and their theoretical reflection over performance-based approaches to building design to the production of space and materialization of architecture by robotic fabrication methods. The increasing merger of the virtual and physical world and its impact to the design of architecture, the production of buildings and the role of the architect was a major motivation for the symposium and the subsequently emerged special issue you are holding in your hands. The symposium was structured in four different thematic strands: Spatial Representation, Computational Thinking, Building Performance Computation and Computational Fabrication, which subsequently provides the basic structure for this special issue. Computation is involved in many aspects of the design of architecture and its production: from the representation of design concepts in virtual and augmented realities to the physical making of architectural objects and spaces through robotic fabrication. Furthermore, the social and political dimensions of architectural space are addressed by computational means and become parameter in the variety of computational performance criteria. It is the privilege and obligation of academia to rigorously and fearlessly explore new territories but also to challenge supposedly safe grounds, which the hereby-presented work does. This Special Issue includes three selected and extended papers from the eCAADe RIS 2017 complemented by invited papers from the larger academic community working in the realm of computational methods in architecture. The complete work presented at the eCAADeRIS 2017 symposium is published as complete proceedings (Spaeth and Jabi, 2017). The first article contains an overview on significant approaches to computation in architecture, reflecting on the development of design computation in three stages: Caetano and Leitão in ‘Architecture Meets Computation’ identify the embryonic stage of architectural computation in the 1960s, followed by a ‘post-folding period’ from the 1990s to the 2000s, to merge into the second generation of computational designers. Based on an extensive literature review, the authors provide a comprehensive overview of computational tools and corresponding architectural theories and respective architectural innovation. The complex interaction between architectural theory, production and computational design advances is critically reflected and uncovers that the causality of events, theory and tools appear not necessarily direct and linear. In the second article titled ‘Animation as a Computational Framework for Architectural Design Composition’, Dounas manifests the computational design methods through dynamic grammars analysing andbeing applied to a constructivist design reflecting onHiller and Leaman’s theory of transformation, repetition and transformation of repetition on higher level principles leading to the construction of structures and space. Computational methods lend themselves almost naturally to the technical utilization for building performance simulation. The challenge, which the presented papers are focused on is the integration of building performance criteria into computationally driven design processes. Alnusairat and Jones propose in their article ‘Ventilated Sky Courts to Enhance Energy Savings in Highrise Office Buildings’ an integrated design approach using a combination of Building Energy Simulation (BES) and fluid dynamics to develop enhanced passive strategies for sky courts in high-rise office buildings. The authors conclude that through the computationally developed sky courts, consideringorientation, height, floor area andothermorphological parameters in the design system the cooling load could be reduced by 50% when free-cooling was adopted. Like energetic simulations, the simulation of daylight situations in buildings is characterized by their time complexity, which results in the need of relatively high computing power and subsequently computing time. The design process, which requires immediate response from the design system to the designer, hardly tolerates delayed interaction. In the fourth article, Lorenz, Spaeth, Bleil de Souza and Packianather explore in the ‘Parametric Design for Daylight using Neural Networks’ the viability of neuronal networks and their ability to accurately predict spatial daylight conditions in early design stages. The definition of relevant design parameter and the setting of the training simulations appear to be crucial to the process. Successfully implemented – Lorenz et al.

中文翻译：

---

社论

2017 年，第五届 eCAADe1 区域国际研讨会 (RIS) 在加的夫大学威尔士建筑学院举行，由 Wassim Jabi 博士和我共同主持。该研讨会是 eCAADe 年会的补充系列的一部分，这些年会旨在促进和关注设计领域的计算，建筑和教育和实践中的空间。与年度会议一样，区域国际研讨会在欧洲各大学之间轮流举行，但在更熟悉的环境中具有研讨会性质。他们的结果通常作为研讨会议事录小册子出版，并在 CUMINCAD 上系统地编目。 2 在卡迪夫大学的第五次实例中，题为“虚拟与物理”——与本期特刊同名——欧洲研究人员和学者齐聚一堂，展示他们最新的研究成果，讨论“空间再现与空间制作”之间的问题。正如研讨会标题所暗示的，研讨会上的贡献探索了广泛的领域，从虚拟空间的研究及其对基于性能的建筑设计方法的理论反思，到空间的生产和通过机器人制造方法的建筑物化。虚拟世界和物理世界的日益融合及其对建筑设计、建筑生产和建筑师角色的影响是研讨会的主要动机，随后出现了您手中的特刊。研讨会由四个不同的主题组成：空间表示、计算思维、建筑性能计算和计算制造，随后为这一特刊提供了基本结构。计算涉及建筑设计及其生产的许多方面：从虚拟和增强现实中设计概念的表示到通过机器人制造的建筑对象和空间的物理制造。此外，建筑空间的社会和政治维度通过​​计算手段解决，并成为各种计算性能标准的参数。严谨而无畏地探索新领域，同时挑战所谓的安全领域，是学术界的特权和义务，特此介绍的工作所做的工作。本期特刊包括来自 eCAADe RIS 2017 的三篇精选和扩展论文，并补充了来自在建筑计算方法领域工作的更大学术界的邀请论文。在 eCAADeRIS 2017 研讨会上展示的完整作品作为完整论文集出版（Spaeth 和 Jabi，2017 年）。第一篇文章概述了建筑计算的重要方法，反映了设计计算在三个阶段的发展：Caetano 和 Leitão 在“建筑与计算的相遇”中确定了 1960 年代建筑计算的萌芽阶段，然后是“后“折叠期”从 1990 年代到 2000 年代，合并为第二代计算设计师。根据广泛的文献综述，作者全面概述了计算工具和相应的架构理论以及各自的架构创新。建筑理论、生产和计算设计进步之间复杂的相互作用得到了批判性的反映，并揭示了事件、理论和工具的因果关系不一定是直接和线性的。在题为“动画作为建筑设计构成的计算框架”的第二篇文章中，Dounas 通过动态语法分析和应用于建构主义设计体现了计算设计方法，反映了希勒和李曼在更高层次原则上的转换、重复和重复的转换理论导致结构和空间的构建。计算方法几乎很自然地适用于建筑性能模拟的技术应用。所呈现的论文所关注的挑战是将建筑性能标准整合到计算驱动的设计过程中。Alnusairat 和 Jones 在他们的文章“通风空中庭院以提高高层办公楼的节能效果”中提出了一种综合设计方法，结合建筑能源模拟 (BES) 和流体动力学，为高层办公楼的空中庭院开发增强的被动策略建筑物。作者得出结论，通过计算开发的天庭，在设计系统中考虑方向、高度、建筑面积和其他形态参数，采用自然冷却时，冷负荷可降低50%。与能量模拟一样，建筑物中日光情况的模拟具有时间复杂度的特点，这导致需要相对较高的计算能力和随后的计算时间。设计过程需要设计系统立即响应设计师，几乎不能容忍延迟交互。在第四篇文章中，Lorenz、Spaeth、Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。建筑物中日光情况的模拟具有时间复杂性的特点，这导致需要相对较高的计算能力和随后的计算时间。设计过程需要设计系统立即响应设计师，几乎不能容忍延迟交互。在第四篇文章中，Lorenz、Spaeth、Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。建筑物中日光情况的模拟具有时间复杂性的特点，这导致需要相对较高的计算能力和随后的计算时间。设计过程需要设计系统立即响应设计师，几乎不能容忍延迟交互。在第四篇文章中，Lorenz、Spaeth、Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。这导致需要相对较高的计算能力和随后的计算时间。设计过程需要设计系统立即响应设计师，几乎不能容忍延迟交互。在第四篇文章中，Lorenz、Spaeth、Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。这导致需要相对较高的计算能力和随后的计算时间。设计过程需要设计系统立即响应设计师，几乎不能容忍延迟交互。在第四篇文章中，Lorenz、Spaeth、Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。Bleil de Souza 和 Packianather 在“使用神经网络的日光参数化设计”中探讨了神经元网络的可行性及其在早期设计阶段准确预测空间日光条件的能力。相关设计参数的定义和训练模拟的设置似乎对该过程至关重要。成功实施 – Lorenz 等人。