sorry...

the links should be these:
http://en.wikipedia.org/wiki/Spatial_network_analysis_software
http://www.awe-communications.com/Propagation/Urban/Comparison/index.htm
http://www.spacesyntax.org/software/index.asp

Sounds really interesting. The descriptions are really vague though. I'm not quite sure I understand how the programs work. Do you input data, or is it read from a dataset file? What kinds of things is it used for specifically? Is it just more specific GIS, or is there another way to use it? It could be a really useful tool, but it also seems like it may just be some fluff.

In Depthmap, info is read from a DXF file and the visibility analysis is rendered from that. What gets me is that it has a pedestrian path prediction as well, and I'm not sure how these pedestrian "agents" are modeled. They don't seem to have a goal in mind as a human pedestrian would...they just seem to wander aimlessly. I'm not a programmer, but it seems like there should be some way to get the "agents" to react to certain factors in the design. For example, some sites and elements of a cityscape will have more magnitude than others, and in contrast, some parts of a city will repel people. These elements of a city will alter pedestrian paths and consequently alter development, and this doesn't seem to be addressed in this software.

I think my descriptions might be vague because I'm still trying to grasp what exactly this is and what it can do.

here are some more links on the pedestrian part of it:

http://www.vr.ucl.ac.uk/research/evas/
http://www.vr.ucl.ac.uk/publications/turner2002-000.html

That's much more specific. It does seem a little one dimensional, though. you can almost do that sort of thing with some MEL flocking scripts. I think animators are a little farther advanced in determining human motion behavior. What are you looking to do with this? I am certainly intrigued by the concept. I think I might look into the Axess plugin to ArcGIS.

Yes! Thank you! I didn't even think to look at animation scripting as a similarity. Can those be manipulated to attract/repel from a certain object? In other words, can they modify a pedestrian path according to the environment. If so, then the animators win this round.

As of now, I'm looking at this software to attempt to develop a pedestrian/transit oriented community. Depthmap (or another one of the programs) could deploy 100 or so people/agents from the transit stops and I could see where they go and decide where should be developed. In my mind it works great, but the mindless program "agents" aren't helping.

try this out:

http://web.tiscali.it/maya_tutorial/

There are some that are probably more advanced and others that are dumber, depending on your level of MEL knowledge, but this one will get you started.

Christian Derix at Aedas / some school in Zurich / University of East London is your man.

we use a software called legion

legion

and we are doing large scale transit, is this a school project?

though less graphically sophisticated than the examples your showing up there, i have done some work in the past with a free program called NetLogo. Developed by researchers at Northwestern, it is very easy to use and as open ended and flexible as you could need.
if you have any programming experience... even just flash programming you would likely be able to build your own rudimentary models and display their effect on up to thousands of agents independently...depending on the quality of your processor, the sky is the limit.

check it out, though im not sure its exactly what your looking for, i found it both easy to learn and fascinating. I could easily see it being applied to pedestrian traffic as its easy to set up positive feed back loops and even swerve effect models which could give you some really interesting emergent paths.

NetLogo

ArT.: Thanks. I don't know anything about MEL or Maya for that matter (I've always been a 3dmax guy...I guess its time to change). I haven't looked at it in depth yet, but it seems very useful. I'm probably saying this prematurely, but I don't know if it will be able to analyze the context of the site. In a way, it seems exactly the OPPOSITE of Depthmap's agents. Where Depthmap has focused agents with no brains/direction, the MEL scripting looks like it creates particles with direction, but no focus...if that makes sense.

PsyArch: I just googled Christian Derix and he seems very interesting. I'll look him up.

aseid: Legion looks like what I need! Too bad its not downloadable from their site. This is for a thesis project on the development between two rail transit systems (one urban, one suburban) and the periphery.

lletdownl: I just downloaded NetLogo... now to see what its like.

Thanks everybody!

its an interesting program that can give you some good diagramtic stuff at the start of projects.

i made a quick model to show the dispersing of many small community parks was more effective than a small number of large parks. i did this roughly by correlating the size of the park to the sphere of influence that park had. Assumed a variable density level of the surrounding population which randomly generated values (within a given range) for each space in the field. depending on the number generated for that space, the presence of a park would have more or less influence (for instance a 1 being a single person alone through 6 a family with 4 children... in that way i said a higher number = more attracted/influenced by parks) , and that influence would be passed on creating a limited positive feedback loop as might loosely be considered the impact of a park on say parents with children... through which they introduce others nearby to the park. I messed around with the values of these until i got a diagram that suited me... so in that regard it was nothing profound, but because the logic was relatively sound, it proved a valuable presentation tool/concept booster.

there were others in my studio who, using this program, spent the entire semester writing a program and refining values and relationships that in the end, began to approximate the logical dispersement of people around the resources of a city... so its quite diverse... kind of the 'flash' of agent based modeling in that way

Oh yeah, NetLogo. A classmate of mine looked into that. I don't think he could get it to work for what he needed, but he wasn't very patient. It does seem interesting, although Legion is definitely what you need. Seems way to legitimate (=expensive) for your purposes though. You're probably right about the flocking scripts. they certainly lack brains since their behavior is based on a few simple rules:

http://www.red3d.com/cwr/boids/

dunno 'bout all this, but on the main spaceSyntax mailing list folks couldn't agree for ages, and still can't imho, what represents a "street". I'd take all of this with a HUGE grain of salt and not rely on any "analysis" the software comes up with.

had a seminar once with Alain Chiaradia from Ssyntax and it was a bad joke. it all appears half-baked theory, at best. and Ssyntax AFAIK being a privately held firm with univ. ties does not help their credibility, as they market services that "have" to work. so, criticism of what other scientists see as their main appraoch is violently repelled.
now feel free to bash me on all of these points if you know better..

"the presence of a park would have more or less influence (for instance a 1 being a single person alone through 6 a family with 4 children... in that way i said a higher number = more attracted/influenced by parks) , and that influence would be passed on creating a limited positive feedback loop [...] I messed around with the values of these until i got a diagram that suited me... so in that regard it was nothing profound, but because the logic was relatively sound, it proved a valuable presentation tool/concept booster."

well - and no personal attack intended, don't we all need profound diagrams? - isn't this just flawed? you get credibility from using some software that is "science" and thus you're able to underpin your "findings" wiith "data". while in fact, they only thing you're doing is modelling events that suit your preconceived idea.
let's face it, there is no "proper" behavioural model currently available that would give you accurate predictions on flow this that etc.

Its good to hear about a meeting with someone from Space Syntax. I'm not sure how their joint venture with the UCL works, but I'm usually not a fan of professors taking students ideas and selling them as their own. Then again, I don't know if this happens there.

Just as a reminder that this software, whether from Space Syntax, Legion, or NetLogo, still only develops a theoretical model. No software (yet...read Kurzweil) can exactly predict what a group of pedestrians will do, but some can make a really good guess. And where this gets interesting is when this "flawed" software is applied to practices other than architecture. It'd be great to see vehicles take this software into a real-time on-board program as to analyze the context and avoid hitting pedestrians/cyclists/other vehicles.

noci-
no personal attack taken... but i think its as valid a diagramtic tool as anything else... how is developing an agent based model demonstrating your concept within logical and rational boundaries any less valid than a sketch, text, or any other means of conveying an idea? in the end its all subjective... there is no singular event to be shown

the difference between a sketch and an agent based model isn't great *if* the constraints and limitations of said model are made clear in the process & presentation.
Insofar you might say the problem lies more with the handling of the idea and data than it does with what it actually says in the end.
There is an uncanny intersection of "science" and "subjectivity" going on. As a practice this is not sustainable, since scientific demands of methodology are way more strict than, say, artistic deductions or stipulations- and there is no valuation of one over the other implied in this statement whatsoever.
Creating a crossover between those two zones will damage independent efforts on both fronts and further move the architect towards a position of a babbling pseudo-scientist who can't differentiate between art and science.

science is art.

it truly is. just ask watson and crick.


however, as far as it goes these sorts of programs face the same problem associated with ALL computer models, even the ones that model the origin of the universe (created to test the role of dark matter and dark energy, etc)...namely that a model can be manipulated to produce the results that "feel" correct, regardless of any actual connection to reality.

The response to this issue is to describe and admit it at the outset. Then test the results in real life and see if anything good comes of it. If it does then there is something there. I not, then...not.

unfortunately, much of the work tends not to be implemented in real world, though i have to admit my info on the subject is mostly limited to a lit review of bill hillier's work for my phd. i eventually gave up on it and switched to morphological analyis...ie, changed my topic to something easier to measure...;-)

"science is art."

well, it depends on which strata of science you're talking about.
read as a generalized statement, I respectfully disagree with you.

but I have an intuitive feeling as to where that theorem originates- namely in the realm of intuitive decision-making that is a crucial element of any scientific endeavour. without it, there'd be no progress, yes. at the same time though, its external over-pronounciation as part of the process mystifies it.

on a hermeneutic and empiricist level, science is detached from art insofar as i.e. hermetic postulations, constructivists assumptions and culturally symbolic expression of "results" - all of those possible modes of artistic expression - are invalid when it comes to rationally underpinning a scientific hypothesis. (*1)

touching on the dreaded subject of "subjectivity" - it is true that anything is "subjective", as the prime ontological question of origin - from where a universally valid frame of reference could be constructed - has eluded philosophers throughout the ages, and still does. so in the "hard" science we are talking about, "objectivity" was replaced with "intersubjectivity", i.e. a social contract that regulates what is assumed the status of knowledge at some point. as one can clearly see, "subjective" from that vantage point does not mean "free for all to say whatever". it is often being read as such, though, leading to all kinds of misunderstandings.

(*1) naturally, this is only true in that science where a rationalist / empiricist model serves as a common ground plane. it might not be true in, say, ethnology.

very interesting discussion, by the way. thank you.

there is a standard brief by John Fruin regarding pedestrian flocking and analysis that is boilerplate for transit projects, make of that what you will

http://elevatorbooks.stores.yahoo.net/pedplanandde.html

the question is what is the "level of service" you are designing to

may have a digital copy around here somewhere

also try this site, they have many publications, of particular use is the Report "Transit Capacity and Quality of Service"

http://www.trb.org/news/blurb_detail.asp?id=2326

some of my colleagues find the TRB publication to be more reliable, however, i am not totally understanding as to why that is, ill find out though

the "Pedestrian Planning and Design" book looks like it could be useful. But, no ISBN and it doesn't show up in my school library search. If you have a digital copy, that would be great. The TRB book looks to be very in depth and very applicable to this project.

So what I'm finding from all this info and insight is extremely helpful and the links lead to some amazing things. Thank you to everyone. Now time for me to get down and really research all of this.

sorry, i should be clear. science is like art; for just the reasons you describe.

science is limited in its extrapolations for the reasons you describe too. Perhaps I am flawed, but to me the art of science is not in its execution so much as its interpretation. One of my pet interests is in work with dark matter and dark energy, so give this example from the new york times...the science related to this is in the strictest sense very hard and empirical, but the interpretation is not, as the article makes clear. Scientists have to do their research in context, and give it meaning, if they want it to be read (which is very important). One thing i have learned in higher academia is that nothing gets lost faster than research without a hooky narrative. this is as true for particle physics as it is for morphological analysis in urban planning...

and is where the work by hillier and so on gets dodgy. except they also have some serious issues with methodology. but then again hillier is not a particularly great researcher in my books.

I don't think your point on "interpretation" is flawed at all.
in fact, imho you hit the nail right on the head.

what you describe is at the borderline of a set of two subsequent "interpretations": one that is still reasonably "within" science and part of its process, and another that is taking place "outside". the former allows for intuition, creativity and all the stuff we discussed. the latter opens up an extraneous reading or manufacture of "results", or even a hijacking if you are so inclined, by agents and institutions that may be alien to science itself. in methodology, this last "opening" is a point of worry, as illicit feedback effects might happen that influence results, or their reading, when still in the process of, say, fundamental research.

now this was a negative illustration; it can be reversed just as well: institutions or any kind of initiative can provide fertile ground for research- *if* they keep their hands out of it. just as well, you might want to keep them entertained- as far as "hooky narratives" go. true...

in any case - and this has spooked 'round in my head since my last post - we're in a inherently difficult positions, as architects in academia. we're operating at the intersection of art and science, and their cross-fertilization. our ground is shaky, and rightfully so- don't we acknowledge our existence based on such playfulness?
anyways, this leads me off track.
but it's clear that a thoroughness of method ("science") and playfulness of thought ("art") must both be cared for. it is there that we operate.