Dear Yodan and all,
Yodan, I think you are quite right in arguing that the flow of space is an
interesting concept to consider, and in fact many of the modelling
techniques we use (visual graph analysis, axial lines) are based on this
idea of interconnected units, or flows, how you call them.
I would argue that there are several issues to reflect on: first of all,
what is the research question that needs to be addressed by the
analysis? Depending on what I am interested in I might choose different
representations of space, for example if I want to study pedestrian
movement patterns I can include shortcuts into my spatial model (whether
line based or visibility based) that wouldn't make sense for vehicular
traffic.
Secondly you are mixing up polygons and convex spaces. While polygons are
necessary to run certain computional analysis as Jim argued (since the
computer needs to read a floor plan as connections of areas, and typical
architectural drawings are messy in this sense; they might have small
holes that seem to suggest connections to the machine, which in fact may
not be connections at all). As such polygons can be very complex and
large, the only 'rule' being that they consist technically of one line
that connects the end of the line to its beginning (like a casing or
shell), so that the computer can read this. In contrast, convex spaces are
well defined as areas where you can see all points within the convex space
from any chosen point. (See: The Social Logic of Space).
Last but not least, whether a building is easily represented by a series
of clearly defined convex spaces or rather a 'flow of spaces' may depend
on the building you choose to study. The National Gallery, for instance is
easily perceivable as a series of convex spaces, while you may find it
more difficult to define convex spaces for the Tate Modern, and therefore
may choose to use a different type of analysis that focuses on the 'flow
of spaces' like visibility graph analysis.
Cheers,
Kerstin
Yodan Rofe wrote:
> A question from the sidelines:
> But isn't almost the whole point of modern architecture to blur these
distinctions between rooms as convex polygons. Isn't this what the "flow
> of space" all about? So, when you identify particular spaces as closed
polygons you're interpreting the building in only one of the ways that
it could be interpreted?
>
> On the other hand, doesn't that makes buildings more like cities, where
there are convex spaces, but really most of the analysis is of the axial
> lines and isovists?
>
> Yodan
>
>
> On Mon, Oct 26, 2009 at 6:14 PM, Jim Turner <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Hello Alireza. I am the original inventor of the "Syntax2D"
application at the University of Michigan. I no longer work on it
but I monitor the messages of the space syntax group. Your request
is the only one like it I've seen. I offer this abridged response:
>
> It seems to me that most syntax people ignore a very important step
in the process of applying syntax algorithms to architectural floor
plans. In other disciplines it is assumed that any data being used
for computational purposes must first be "cleaned." Much of the
cleaning (removing of bad data, duplicates, ...) can be automated,
but some of it cannot. Often, an expert must inspect the data
carefully before approving it for analysis.
>
> In the world of "building space syntax" well-defined spaces seem to
me to be necessary. A set of floor plans from AutoCAD in DXF format
usually is not in a state where it can be used directly by space
syntax software. I wrote Syntax2D to require the rooms to be closed
polygons (with multiple circuits for holes); and this is one of the
complaints of those attempting to use the software.
>
> No one wants to spend the time to create a new layer and trace wall
lines to make closed POLYLINES. I'm very fast at the task, but of
course, it's my program. Without room polygons it is impossible to
determine if a person is inside or outside a space; areas are
impossible to calculate as are the tens of interesting and useful
polygon measures that are trivial if one has polygon data. The
convex hull is a nice attribute to work with. Subdivision of space
into triangles, convex spaces, ... is impossible. One cannot even
draw spaces in different colors.
>
> Granted there are a few functions that can take a set of unconnected
lines and intelligently yield a polygon by matching endpoints,
trimming overlapping lines and extending non-overlapping lines; I
have written many of these myself over the years. This is the
dilemma: Should the SS application assume that the DXF file contains
room polygons or should it provide options to attempt to compute the
polygons from LINES and open POLYLINES?
>
> The first option puts the onus on the program user; the second
option puts its trust in a programmer to provide automatic polygon
extraction.
>
> I have not seen a discussion of this topic. It's no wonder you and
many others request building/room data "even if there is the complex
plans with out any extra information." The extra information will
not help with SS analysis; architects draw lines, dimensions, notes,
title blocks; they do not care about polygons.
>
> If you wish, I can send you some reports I wrote regarding this
issue, including the first version of Syntax2D.
>
> I am passionate about this cleansing of data having observed fellow
researchers try to import a full DXF file and find the layers that,
they assume, will provide them with correctly formatted data for
lighting, acoustic, thermal, structural, ..., analysis.
>
> Now, where did I put my Primo-line-to-polygon-conversion
application.
>
> Jim Turner
> Professor Emeritus of Architecture
> University of Michigan
>
>
>
>
> --On Saturday, October 24, 2009 1:39 PM +0330 Alireza Kharazmi
<[log in to unmask] <mailto:[log in to unmask]>>
wrote:
>
>
> Dear All,
> I am studying some architectural plans of well-known city hall
complexes , as a sample study in architectural scale. it is the
study of
> justified graphs and visibility in city hall plans. I will do
it for 2
> samples. but I cant find complete plans(images or DWG files). it
will be
> appreciated if you inform me about important points of space
syntax studies in this kind of buildings and send me please any
useful
> documents about it,if it is possible. (even if there is the
complex plans
> with out any extra information).
> thanks for your favor in advance.
> Alireza
>
> --
> Regards,
> Alireza Kharazmi Nezhad.
> M.Arch Stu.,Qazvin Azad University
> http://qazviniau.academia.edu/AlirezaKharazmi
> [log in to unmask] <mailto:[log in to unmask]>
>
>
>
>
> --
> Dr. Yodan Rofe` - Senior Lecturer
> Desert Architecture and Urban Planning
> J. Blaustein Institutes for Desert Research
> Ben-Gurion University of the Negev
> Sde boqer Campus, Israel 84990
> Tel. 972-8-6596884 Fax. 972-8-6596881
> http://www.bgu.ac.il/CDAUP/yodan-rofe.html
--
Kerstin Sailer
Lecturer in Complex Buildings
The Bartlett School of Graduate Studies
University College London
1-19 Torrington Place
London WC1E 6BT
United Kingdom
(mobile) 075 45211657
(www) http://www.aas.bartlett.ucl.ac.uk/
(www) http://www.space.bartlett.ucl.ac.uk/people/kerstin/
|
