On Thu, 8 Jun 2006 09:05:46 +0100, Professor Bill Hillier
<[log in to unmask]> wrote:
>Dear Yixiang Long - the discussion of this issue in the 'Natural
>movement' paper is still probably the best thing on this problem.
Hello Yixiang,
Actually, you may also want to check our paper in EPB last year.
As you are well aware, there is no rigorous definition of axial lines, so
one can only claim to develop algorithms inspired by those. But anyone who
claims to be able to generate axial lines may as well be selling you snake
oil -to this day, they remain algorithmically undefined.
We had to approach the problem of how to trace axial lines at the border
of the domain, and as you will see from the paper that the solution in the
SLoS is not correct -actually, there are two figs in the book and it
looks "as if" the layout of Gassin was cut in one to make all lines
intersect.
See my web page for the paper.
All the best,
Rui
The
>main thing is to make the system analysed sufficiently large the
>remove any edge effect from the areas of focal interest.
>
>At the stage of the 'Natural movement' paper, we did not of course
>have the concept of radius restricted measures. In fact in axial
>analysis using Axman, 'local integration' analysis at radius-3 does
>not show signifcant edge efect in most systems, and nor does a more
>global analysis using what came to be called 'radius-radius' (i.e.
>setting the radius of analysis at the mean depth of the system from
>the most integrated line).
>
>It is also worth noting that edge effect is much less of a problemn
>with segment angular analysis which is now routine with the use of
>DepthMap in urban syntactic studies.
>
>>Today, in fact we routinely use whole city maps even when we are
>>studying particular areas. We now have (thanks to the efforts of
>>Valerio Medeiros) a data base of 158 cities (and counting) which
>>should be available to the whole research community soon. - Bill
>
>
>
>
>
>>Hello, all
>>I remembered that there are some papers that had some discussion on the
>>boundary effect on space syntax value (how far beyond the study area the
>>axial map should be drawn). I wonder if someone can give me the
>>information about it. Thanks a lot.
>>
>>Regards
>>Yixiang Long
>=========================================================================
|