Chair for Information Architecture | FS2009 Elective …...Transport by bus 13.00 Meeting point at...

Introduction to Information Architecture

Prof Dr Gerhard SchmittFebruary 16, 2009

Chair for Information Architecture | FS2009

Elective Course Information Architecture

Donnerstag, 19. Februar 2009

Introduction Information Architecture

Overview

– Chair for Information Architecture– INFORMATION Architecture– ARCHITECTURE Information– INFORMATION ARCHITECTURE– Introduction Exercises



Overview



Chair for Information ArchitectureProf Dr Gerhard SchmittDr Remo BurkhardDipl Ing Jan HalatschDipl Ing Antje KunzeDr Stefan Müller ArisonaMSc Christian SchneiderDipl Des Sandra WipfliAdmin Ass Martina Sehmi


Overview Lectures FS2009


Lecture | 13 - 14 p.m. | HIT H42Participation 80% = 1 ECTS

Information ArchitectureProf Dr Gerhard Schmitt051-0724-09 V


General Information Lecture


Exercise Line I | 9 - 12 a.m. | HIT F22 / H31 Participation 80% + accepted Project = 3 ECTS

Topic: New Methods in Urban Simulation:“Collaborative City Design” | 051-0726-09Jan Halatsch, Martina [email protected], [email protected]


General Information Exercise


mailto:[email protected]




Exercise Line II | 14 - 17 p.m. | HIT H31 Participation 80% + accepted Project = 3 ECTS

Topic: Visualize Urban Complexity | 051-0726-09Christian Schneider, Sandra Wipfli, Lukas Treyer


General Information Exercise


Seminar | Thesis Elective/Diplomwahlfacharbeit 14 - 15 p.m. | HIT H31Participation 80% + Presentation + Thesis (6-8 pages) = 5 ECTS

Topic: New Methods in Urban Simulation:“State of the art in city modeling” | 063-0724-09Jan Halatsch, Martina [email protected], [email protected]


General Information Diplomwahlfacharbeit







Schedule

16.FEB 23.FEB 02.MAR 09.MAR 16.MAR 23.MAR 30.MAR 06.APR 13.APR 27.APR 04.MAI 11.MAI 18.MAI

seminar week

introduction exercises

spring break

final critique

biblioteca oechslin

16.FEB 02.MAR 09.MAR 16.MAR 30.MAR 06.APR 20.APR20.APR20.APR20.APR20.APR 27.APR 04.MAI 11.MAI 18.MAI 25.MAI

guest lecturers


L1L2L3L4L5L6L7L8L9L10L11

16-0223-0202-0309-0316-0330-0306-0427-0404-0511-0518-05

Introduction Information ArchitectureExcursion Biblioteca Werner OechslinHistory of Information ArchitectureThinking Information as a raw MaterialInformation. The 5th Dimension in ArchitecturePush.Click.Touch. Visiting the Value LabSmart EnvironmentsComputer Supported Collaborative WorkNew Methods in Urban SimulationGuest lectureGuest lecture


Overview Lectures


L1 Introduction to Information Architecture



Overview Lectures


L2 Excursion to the Biblioteca Werner Oechslin

Monday 23rd of February, 2009 | 13:00 - 17:00

Transport by bus

13.00 Meeting point at the orange sign ‘SCIENCE CITY

INFOSPOT’ outside the HIL-building.

Contact:

Sandra Wipfli, [email protected], + 41 44 633 29 02


Overview Lectures




L3 History of Information Architecture

– From Cardboxes to Web 2.0– Principles – Evolving Trends


Overview Lectures


L4 Information as a raw Material

– Data, Information, Knowledge– New Methods in the architectural Design Process

– Reading– Writing– Calculating– Designing– Modelling


Overview Lectures


L5 Architecture and Dimension

– Dimensionality in Architecture – Information as raw Material– Information and Sustainability


Overview Lectures


L6 Push.Click.Touch. Visiting the Value Lab

– New Methods in – reducing Complexity– elaborating Knowledge– comparative Studies– strategic Planning

– Opportunity Diplomwahlfacharbeit


Overview Lectures


L7 Smart Environments

– Principles– Technologies pervading Space– Quality Characteristica– The changing Relationship with the Environment


Overview Lectures


L8 Computer Supported Collaborative Work (CSCW)

– CSCW-Tools– CSCW in Teaching– CSCW in Research– CSCW in Practice


Overview Lectures


L9 New Methods in Urban Simulation

– Urban Design Indicators– Urban Patterns– Digital City Structures– Simulation Methods


Overview Lectures



Overview



Jantar Mantar, astronomical Observatories of Jai Singh II, 1727-1734

INFORMATION Architecture or ARCHITECTURE Information?


Information Architecture


http://www.jantarmantar.org/

http://www.jantarmantar.org/


Information Architecture

INFORMATION Architecture Metaphors and principles of physical architecture applied to digital data and information creates an architecture of information, using information as raw material

ARCHITECTURE Information Making the invisible visible in the form of digital information applied to physical architecture to better understand and design physical architecture


ArchitectureRequires thoughtful analysis to manifest thoughtful synthesis (design).

The Architecture of InformationConstruction of a structure or the organization of information.


INFORMATION Architecture


The Architecture of Information

‚INFORMATION Architecture‘ expands the definition of architecture to areas outside the traditional realm.

‚INFORMATION Architecture‘ uses architectural metaphors for better structuring, understanding and communicating information


INFORMATION Architecture



Overview




ARCHITECTURE Information

‚ARCHITECTURE Information‘ uses simulation for more than creating images or artefacts based on geometric constraints, rules, or cases.

Rather, non-geometric factors such as light, energy, structure, behaviour or systems knowledge become available for integrated direct modelling.

Formalise and generalise design principles



Overview



Knowledge

Architecture

Culture

RES

EARCH

SYN

THES

IS

INFORMATION ARCHITECTURE

Architecture and Information

Information

Data



Creative Industries



The City - Data, Information, Knowledge, Architecture

First zero-carbon, zero-waste city in Abu Dhabi. Masdar City, Foster & Partners 2007


http://www.jetsongreen.com/2007/05/foster_partners.html

http://www.jetsongreen.com/2007/05/foster_partners.html



First zero-carbon, zero-waste city in Abu Dhabi. Masdar City




First zero-carbon, zero-waste city in Abu Dhabi. Masdar City


Value Lab

Projects

Reduce Complexity


Projects

Tools for collaborative planning

Future City Designer


CityEngine

Projects

Tools for planning future cities


Projects

Bridging the gap between digital and physical

The sensitive Tapestry



Overview

– Elective Course ‚Information Architecture‘– INFORMATION Architecture– ARCHITECTURE Information– INFORMATION ARCHITECTURE– Introduction Exercises


Workshop | 9 - 12 p.m. | HIT F22 / H31 New Methods in Urban Simulation:Collaborative City DesingJan Halatsch, Martina [email protected], [email protected]


Information Architecture - Praxis I







New Methods in Urban Simulation

Design your Science City Dübendorf withlatest tools and technologies together with your peers



New Methods in Urban Simulation

Design your Science City Dübendorf withlatest tools and technologies together with your peers




051-0726-09

New Methods in Urban SimulationCollaborative City Design

(pictures courtesy of foster+partners)

BUT: Need for collaborative design and variant generation





Main Goals to think about

Zero CarbonZero Waste









(examples from HS2008)














New Methods in Urban SimulationState of the Art in City Modeling

“You want to know what will be “in” tomorrow? Experience new technologies in City Modeling, Urban Simulation and Visualization and for the Planning of Future Cities”

D. Fellner & S. Behnke / Modeling the Appearance and Behavior of Urban Spaces

Figure 10: Interactive reconfiguration of urban lay-outs [ABVA08]. (Top) Satellite image of an original andmodified urban layout (Istanbul). A residential zone has beenconverted to an industrial zone. (Bottom) The process con-sists of recomputing the topology of the affected area to ac-commodate parcels of a new zoning type, and then copyingselected tiles from the industrial zone of the city (red) to thepreviously residential area (blue).

card smaller tiles of pixels during rasterization. The bookby Möller and Haines has a good overview of the currentstate of the art [MH02]. Another important feature are oc-clusion queries that can test bounding volumes for occlu-sion. A major challenge in this context is to hide the la-tency of the occlusion queries and to utilize temporal co-herence [BWPP04,GBK06,MBW08]. In all cases, when theviewpoint is far above the city occlusion culling has oftenlittle impact and other strategies need to be used that focuson simplifying the environment.

Figure 11: Continuous Model Synthesis. The modelsshown on the left are automatically generated from the ex-ample model on the right. Different textures are applied tothe buildings, but the shape of each building resembles theshape of the input. The images are courtesy of Paul Merrell,University of North Carolina.

An interesting challenge is to compute geometric levels-of-detail (LOD) for urban models. There are many strongmethods for general mesh and object simplification. A goodintroduction is the book by Luebke et al. [LRC!02]. Twofundamental early papers introduce operations to collapseedges [Hop96,GH97] in triangle meshes. While these meth-ods work well for objects that contain smooth surfaces mod-eled with many triangles there are substantial obstacles whenapplied geometric level-of-detail methods to urban environ-ments. Usually, traditional level-of-detail techniques do notwork well with disconnected polygons that are frequent inurban environments due to plant leafs in urban vegetation, orplanar structures with many sharp corners dominant in archi-tecture. Some special purpose methods have been inventedfor building footprints. For example, Chang et al. [CBZ!08]describes a clustering based simplification method for urbanspaces inspired by Kevin Lynch’s Image of the City book,but in general a high quality automatic simplification of ur-ban environments is an unsolved problem. The main aspectof urban environments that can benefit from level-of-detailis the terrain. Early approaches, e.g. [LKR!96], built on theassumption that triangle rendering is expensive and there-fore the algorithms were quite sophisticated. Current algo-rithms try to incorporate the fact that larger batches of trian-gles need to be rendered at once and focus on algorithm sim-plicity [LH04] and large memory management [GMC!06].For architecture and vegetation, simple alternatives to auto-matic simplification are to create multiple versions of eachbuilding or tree procedurally [MWH!06] or by hand. Fur-ther, procedural modeling can help simplification, by pro-viding semantic data. For example, it is very helpful to knowwhat the actual facade planes are and what geometry be-longs to one facade. Finally, since man-made structures arehard to simplify with geometric methods, another option is

c! The Eurographics Association 2009.








mation about geographically weighted statistics at severalscales concurrently.

Roman et al. [RGL04] presented an interactive systemfor constructing multi-perspective images from sideways-looking video captured from a moving vehicle. The inputto their system is a set of video frames with known cam-era pose. Their system automatically computes an addi-tional cross-slits camera between every pair of adjacent user-specified cameras leading to a smooth interpolation of view-point in the final multi-perspective image. Multi-perspectiveimage of a whole city block can be created in a few min-utes. The goal of this work is to simultaneously view real-world urban scenes that cannot be captured in a single pho-tograph, rather than to visualize the simulation data of an ur-ban space. New techniques could be explored that combinea multi-perspective approach for data visualization.

To date, simulation systems such as UrbanSim have beenrelatively limited in their scope of visualization, in spite ofproviding a sophisticated economic and behavioral simula-tion engine to model the location and travel choices of mil-lions of agents in the system. A typical scenario is that man-ual post-processing of simulation results must be done bya model user to extract summary indicators from the results,export them from the simulation environment into a GIS sys-tem, establish relational joins of the indicators to existingGIS layers, and then manually rendering thematic or choro-plethic maps to render the spatial variation in the resultingindicators (Figure 13). As used in the planning literature,an indicator is a variable that conveys information on thecondition or trend of one or more attributes of the systemconsidered. The work of Schwartzmann and Borning [SB07]developed a web-based indicator system for UrbanSim andevaluated design techniques including Value Sensitive De-sign, paper prototyping, and frequent user testing.

4.3. Bridging the Gap between Simulation andVisualization

This process of simulation and then visualization has manylimitations, not the least of which is the level of effort. Asa result, too little visualization is actually done in practice,and this leads to diminished access to the simulation results,and reduced diagnostic capacity to determine when there areproblems in the simulation. One option being recently ex-plored is to more tightly integrate visualization efforts withthe simulation process, achieve a beneficial and symbioticrelationship, and ultimately lead us to a more desired andintegrated approach to urban modeling.

A recent approach proposed by Vanegas et al. [VABW08]uses urban layouts for the visualization of urban simula-tion results. This work builds upon existing visualizationtechniques of urban simulations and extends them by auto-matically inferring new urban layouts for any time step ofthe simulation sequence, by considering both the values of

the state variables of the simulation model, and the originalstreet network, parcels, and aerial imagery of the simulatedcity (Figure 14). The inference algorithms gather stochas-tic data of the original urban layout and use the simulationstate values to obtain a plausible urban layout, consistingof new parcels, streets, and imagery (e.g., vector and imagedata). Altogether, this approach allows for traditional visual-izations as well as that of new content. It has been applied tovisualize a 16,300 km2 urban space.

Another recent approach is that of Weber etal. [WMWG09] who proposed to combine proceduralmodeling techniques with urban simulation to obtain three-dimensional models that change over time (Figure 15). Thesystem includes a street expansion algorithm to place newstreets, a land use simulation, and a traffic simulation asmajor building blocks. The simulation is interactive and theuser can make modifications during the simulation, such ascontrolling the major growth areas, editing road segments,editing land use, and changing parameters used to controlthe simulation. The goal of this work is to provide a genericframework that can be configured for different types of landuse categories.

Both of these works make the initial strides toward theultimate goal of a more integrated approach of urban model-ing, simulation, and visualization.

Figure 15: Interactive Geometric Simulation of 4DCities [WMWG09]. Two time series are shown in thecolumns. The left column shows the transition from low den-sity to high density in the city center. The right column showsa transition of a city based on sustainable development withsufficient green areas.



Figure 13: (Left) A screen snapshot from the Indicator system supported by UrbanSim. (Right) A standard choropleth maptraditionally used to visualize the results of an urban simulation.

Figure 14: Visualization of Simulated Urban Spaces [VABW08]. (Left) An overview of the simulated region (borrowed fromGoogle Maps) along with the parcel geometry inferred from the urban simulation data for a subset of that region. (Center) Aclose up of a part of the city where new developments are predicted by the simulation after a 30 year period, and their geometryinferred by our system. (Right) A close up of a different part in which parcel subdivision was indicated by the simulation.

5. Conclusions, Challenges and Open Problems

Providing realistic and plausible models of dense urbanspaces is a challenge that requires knowledge from severaldisciplines. The pursuit of the accurate modeling of urbanspaces is of significant interest today to urban planners, toemergency management, and to visualization efforts. In re-cent years, the data collected via several forms of acquisi-tion is available, through the Internet, to a widespread audi-ence and has fomented significant activity and applications.Several geometric modeling methods have focused on urban

spaces in order to improve their flexibility and efficiency. Si-multaneously, simulation models are becoming increasinglysophisticated and better able to represent the complex pro-cesses occurring in urban spaces. These simulations serve tobetter understand urban spaces and to help guide 3D model-ing and rendering efforts to produce more realistic and inter-active imagery.

In this state-of-the-art report, we have attempted to helpguide future efforts in urban modeling to have a better un-derstanding of the multiple aspects of this challenge. We



Figure 9: (Left and Center) Example-based urban layout synthesis [AVB08]. A new urban layout is generated by extracting andreproducing the structural attributes of the example fragment and reusing aerial-view imagery. (Right) Interactive proceduralstreet modeling [CEW!08]. The concepts of flow fields and tensor fields are used to model the street layout of a city.

els with a few planar polygons and most details stored intexture maps.

Besides the well established and easily implementabletechniques of back face culling and view frustum culling,occlusion culling can provide speed-ups of several ordersof magnitude for most large and dense urban scenes. Thisis because an urban environment usually lies on a locallyflat surface and the nearby urban structures easily fill thefield-of-view and prevent observing distant structures. Oc-clusion culling either can rely on a precomputation or canbe computed online. To precompute visibility, the naviga-ble space is broken down into smaller volumetric view cells.Researchers have tackled the question of how to efficientlycompute visibility for volumetric view cells and proposedmultiple algorithms [SDDS00, DDTP00, WWS00, LSCO03,SN04, WWZ!06]. The advantage of precomputation is that

occlusion culling needs very little runtime overhead and thatthe results of precomputation can be used for other prepro-cessing algorithms, such as level-of-detail selection and ren-dering time estimation. The disadvantage are the complexityof the algorithms and the long preprocessing times. Onlinevisibility typically computes visibility for each frame of aninteractive simulation from the current view point. In a sem-inal paper Greene et al. [GK93] introduce the hierarchicalz-buffer algorithm. The algorithm allows the identificationof large parts of the scene that are guaranteed to be occludedand that is especially suited to rasterization based rendering.Currently, a complete implementation of the hierarchical z-buffer is not supported by graphics hardware, but the gapis closing. Current graphics hardware includes several use-ful features for occlusion culling, including pixel level oc-clusion culling to avoid expensive shading operations anda hierarchical z-buffer with a few levels (e.g. three) to dis-



Figure 9: (Left and Center) Example-based urban layout synthesis [AVB08]. A new urban layout is generated by extracting andreproducing the structural attributes of the example fragment and reusing aerial-view imagery. (Right) Interactive proceduralstreet modeling [CEW!08]. The concepts of flow fields and tensor fields are used to model the street layout of a city.

els with a few planar polygons and most details stored intexture maps.

Besides the well established and easily implementabletechniques of back face culling and view frustum culling,occlusion culling can provide speed-ups of several ordersof magnitude for most large and dense urban scenes. Thisis because an urban environment usually lies on a locallyflat surface and the nearby urban structures easily fill thefield-of-view and prevent observing distant structures. Oc-clusion culling either can rely on a precomputation or canbe computed online. To precompute visibility, the naviga-ble space is broken down into smaller volumetric view cells.Researchers have tackled the question of how to efficientlycompute visibility for volumetric view cells and proposedmultiple algorithms [SDDS00, DDTP00, WWS00, LSCO03,SN04, WWZ!06]. The advantage of precomputation is that

occlusion culling needs very little runtime overhead and thatthe results of precomputation can be used for other prepro-cessing algorithms, such as level-of-detail selection and ren-dering time estimation. The disadvantage are the complexityof the algorithms and the long preprocessing times. Onlinevisibility typically computes visibility for each frame of aninteractive simulation from the current view point. In a sem-inal paper Greene et al. [GK93] introduce the hierarchicalz-buffer algorithm. The algorithm allows the identificationof large parts of the scene that are guaranteed to be occludedand that is especially suited to rasterization based rendering.Currently, a complete implementation of the hierarchical z-buffer is not supported by graphics hardware, but the gapis closing. Current graphics hardware includes several use-ful features for occlusion culling, including pixel level oc-clusion culling to avoid expensive shading operations anda hierarchical z-buffer with a few levels (e.g. three) to dis-











CoordinatesCourse:Series:Topic:When:Where:Who:Contact:

051-0726-09 (Elective Course Exercise, 3 ECTS)

New Methods in Urban SimulationCollaborative City DesignMondays, 09 - 12 a.m.HIT H31 (iA computer lab)Jan Halatsch, Martina [email protected]@student.ethz.ch+41 44 633 79 62 & HIT 32.3






Workshop | 14 - 17 p.m. | HIT H31 (3 ECTS)Visualize Urban Complexity | WorkshopChristian Schneider, Lukas Treyer, Sandra Wipfli051-0726-09


Information Architecture - Praxis II




How did Zurich grow? How sustainable is it? What path did my breakfast take this morning? What is the journey of rubbish? What flows in and out of the city? Where do how many people live on how much space and why? How would a subway change the appearance of Zurich?

Visualizations can do more than just show data.Tell a story! Interact!

You will:

- adopt Information Visualization skills- learn about Computational Information Design- learn to code with Processing- create and present your interactive, visually

appealing Visualization about Zurich in the Value Lab

You will be creating something which may look like....







http://hindsight.trulia.com/


http://hindsight.trulia.com/map/#lat=36.233&lon=-115.242&zoom=11&mix=0.585

http://hindsight.trulia.com/map/#lat=36.233&lon=-115.242&zoom=11&mix=0.585

http://livepage.apple.com/

http://livepage.apple.com/



- Schedule:

- Team: ‣ Christian Schneider, [email protected]

‣ Sandra Wipfli, [email protected]

‣ Lukas Treyer, [email protected]

16.FEB 23.FEB 02.MAR 09.MAR 16.MAR 23.MAR 30.MAR 06.APR 13.APR 27.APR 04.MAI 11.MAI 18.MAI16.FEB 02.MAR 09.MAR 16.MAR 30.MAR 06.APR 20.APR20.APR20.APR20.APR20.APR 27.APR 04.MAI 11.MAI 18.MAI

seminar week

processing exercisesprocessing for beginnersvisualization coaching

spring break

start of individual projects in groups

bibliotheca oechslin

final critique

25.MAI









Podcast Information Architecture

http://www.ia.arch.ethz.ch/teaching/fs2009-lecture/




18 http://image55.webshots.com/155/7/29/80/2864729800035413848pEXZAa_fs.jpg

26 - 28 http://www.flickr.com/photos/onikasi/1075044888/sizes/o/ 29 - 32, 34, 36 Chair for Information Architecture, ETHZ37-39 Chair for Information Architecture, ETHZ41 http://users.design.ucla.edu/~akoblin/work/faa/


Sources


http://image55.webshots.com/155/7/29/80/2864729800035413848pEXZAa_fs.jpg




http://www.flickr.com/photos/onikasi/1075044888/sizes/o/

http://www.flickr.com/photos/onikasi/1075044888/sizes/o/

http://users.design.ucla.edu/~akoblin/work/faa/

http://users.design.ucla.edu/~akoblin/work/faa/

Date post:	23-Aug-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Chair for Information Architecture | FS2009 Elective …...Transport by bus 13.00 Meeting point at...

Documents