 |
 |
Well, Bill, there was mathematics before space syntax... and graph theory was not less fashionable in mathematics or computer sicence than it is now -most of the recent developments are in physics, which is different from mathematics... I know it's hard to conceive that anyone else has done any work, but you may want to check the work of the great Hungarian mathematician Paul Erdos (who published more than 1,500 papers). Without him, we wouldn't have the recent developments in network science. And I 'm afraid he didn't wait for the Social Logic of Space to start working... Good to see that the interest in physics doesn't die -keep in mind, though, that physicists quote every other disciplines becase that's the way we work. I have never read the Social Logic of Space myself! Best, Rui ___________________________________________ Dr. Rui Carvalho School of Mathematical Sciences Queen Mary, University of London Mile End Road, London E1 4NS, UK http://www.ruicarvalho.org/ On Wed, 22 Oct 2008 23:38:22 +0100, Professor Bill Hillier <[log in to unmask]> wrote: >Dear Lucas - when we did the research that led to The Social Logic of >Space graph theory was much less developed and much less fashionable >than it is now. It was regarded as a backwater in mathematics, and >serious mathematicians - with outstanding exceptions such as Harary - >did not think of it as a field where a good mathematician could make >a living. High speed computers changed all that. But at the time >there was not even agreement on basic terminology: arcs, vertices, >nodes, edges, links and so on. The basic texts all talked about the >Konigsberg Bridge problem and the travelling salesman problem. Claims >were advanced about how graph theory could solve applied problems, >but there was little except preliminary results which seemed to fall >well short of the claims. And there were relative few new theorems. A >great deal of graph theory has in fact developed in the last 25 >years, and of course a key part of this has been in 'experimental >mathematics' - where you see what happens if you do this, this and >this, rather than prove theorems. In fact, some suspect that graph >theory has found a real forte in experimental mathematics. > >In the situation that prevailed then, we did take a deliberate >decision to focus on terms which would immediately communicate - for >example, connectivity rather than degree for the number of nodes >connected to a particular node. At the time, but the way, 'valency' >was used for connectivity more often than 'degree'. Did you know that >? But from our use of it 'connectivity' entered into architectural >parlance, as did less specific terms such as 'permeability'. We >believe this was achieved without loss of mathematical rigour, >because the basic idea of space syntax was to apply simple graph >measures to discrete geometrical objects such as convex spaces or >lines. This again has now become a large field in mathematics call >discrete geometry. But again discrete geometry has expanded and >matured mainly in the last 25 years. > >If you want to do some real history on this, you should look back on >what was around in the 1970s. Freeman's work was of course >remarkable, but its impact was made over a comparatively protracted period. > >But a word of advice. If you do offer a general academic criticism of >others, you must immediately give chapter and verse, so people can >check what you say for themselves. If you don't, you may persuade >those who feel themselves less knowledgeable than you of your case, >but those who prefer evidence to rhetoric may recoil somewhat. - Bill > >At 18:29 22/10/2008, you wrote: >>2008/10/22 Professor Bill Hillier <[log in to unmask]>: >> > Dear Lucas - You only partially quote what is said in The Social Logic of >> > Space on p 273. What it actually says is 'All mathematical formula are >> > original, as far as we know, with the exception of ...' >> >>".. of the formula for ringness which is well know" >> >>I think you missed the point of my e-mail. I agree that the >>normalisation mechanisms are not only original, but very innovative, >>as well as the measure of control. >> >> > I think what you should have said was that measures of >> depth/accessibility - >> > and betweenness - were already in use well before SLS, but our adaptations >> > and interpretations of them for patterns of real space were not. - Bill >> >>As you correct me here, for instance, the mean depth is closeness >>centrality, connectivity is degree. In other words, there is enough >>material to make comparisons. >> >>My point is that, at least in my opinion, there is no effort to make >>such comparisons. Since all following research is based on these early >>books/papers, there we go... more and more hermetic. >> >>At the same time that this simplifies things for not creating a side >>argument to the main focus of analysis, it complicate things for >>outsiders. It creates a huge barrier for them. >> >>This is my opinion and it has been for some years. >> >>Best Regards, >>Lucas >> >> > More substantively: >> > - the two normalisations of depth measures that we used in SLS >> (RA and RRA) >> > are not in earlier literature as far as I know - if they are, no one has >> > even pointed them out - it is now incumbent on you to do so. Please do not >> > instance Phil's book 'Architectural Morphology' as there it is quite clear >> > that Phil is explaining our normalisation - and again not referring to any >> > others. Can you find these in Buckley and Harary's compendious 'Distance in >> > Graphs' for example ? There are of course other ways of normalising >> > depth-type measures, but the way we did it reflected the intuitive picture >> > you could get from the justified graph - another useful innovation in SLS - >> > and this made it intuitively clear and accessible >> > - the 'control' measure is original, as far as I know - again if you know >> > it is not, please point it out. >> > - other measures of line patterns in Chapter 3 are not in earlier >> > literature, with the exception acknowledged on p 273 >> > >> > Our version of the 'betweenness' measure, which we called choice was not >> > developed by us until after SLS, but it is calculated on quite a different >> > way to Freeman's 1977 measure, and is, I still believe, a better and more >> > accurate (in some cases, especially small systems) - though computationally >> > more laborious - way to measure the same thing. >> > >> > I think what you should have said was that measures of >> depth/accessibility - >> > and betweenness - were already in use well before SLS, but our adaptations >> > and interpretations of them for patterns of real space were not. - Bill >> > >> > >> > >> > At 17:17 22/10/2008, you wrote: >> > >> > Hello Ozlem, >> > >> > 2008/10/22 Ozlem Sahbaz <[log in to unmask]>: >> >> Have any of you come across a reference that discusses the commonalities >> >> and differences between Space Syntax methodology and traditional network >> >> analysis ? >> > >> > It depends which kind of traditional network analysis you refer to. >> > Networks are everywhere, in Transport Planning, Geography, >> > Mathematics, Computer Science, Social Sciences, and recently, a huge >> > flurry in Physics (mostly Statistical Mechanics) and Biology - often >> > called 'Network Science'. >> > >> > My thesis compares in details the 'configurational analysis' with two >> > of those traditions, Quantitative Geography and Network Science, also >> > mentioning transportation models. But, unfortunately, it will be >> > publicly available only next year. >> > >> > On this new network science, you may also have a look in may last >> > paper ( http://eprints.ucl.ac.uk/2694/). Hoon's papers, sometimes >> > Sheep's too, also make use of this literature. But no direct >> > comparison, I am afraid. >> > >> > The problem is that our 'configurational studies' evolved along the >> > years into an hermetic field that ignores other literature. This was >> > also discussed here. Alan says that the intention was to facilitate >> > things for architects, I would argue that the effect was opposite, >> > complicating things for outsiders. >> > >> > Doing the review for my thesis, I discovered that this is a problem >> > since its inception. If you get your copy of 'The Social Logic of >> > Space' (1984) and look at p. 273, it is mentioned there that 'all >> > mathematical formulae is original'. It is not, as we know, and most of >> > them was used in the same form in other disciplines, mostly in social >> > networks (early 70's). >> > >> > In the same page there is a brief mention to Kruger's work (1979) as >> > if it were not relevant (because he does no solve the problem of >> > representing space), when it is. It does not mention other >> > graph-theoretic uses in architecture, such as Matela and O'Hare (1979) >> > or P. Steadman (1973). >> > >> > If graph-theoretical approaches are not considered relevant in >> > Architecture, imagine those of other disciplines.... >> > >> > Worse still, Kruger himself (the two papers I read in EPB) does not >> > make a huge review of other disciplines, citing briefly Kansky and >> > Harary. It seems that no-one was ever interested in making such kind >> > of comparison. >> > >> > The payback is that people ignores what is being done here, and they >> > do it on purpose. >> > >> > To complicate matters, this body of research had a 180 turn and is now >> > closely related to environmental psychology. The idea of a 'network >> > analysis' that measures arrangements of objects is now secondary >> > because, in practice, the angular-segment or the visual analysis is >> > 'distance model', as any other such as shortest-paths or travel time. >> > >> > This is even explicitly acknowledged by several authors as you may >> > find in Batty's papers (distance in space syntax) and in the Place >> > Syntax accessibility model. >> > >> > As a newcomer I find all of this totally inconsistent and laborious to >> > deal with. The only positive thing I see in this 'hermetic approach' >> > is that it opens space for criticism, so one has a subject to write >> > papers about. >> > >> > Best Regards, >> > Lucas Figueiredo >> >> >> >>-- >>Lucas Figueiredo >