On 2-3-2013 2:50, Christian Froeschlin wrote:
> Thomas de Groot wrote:
>
>>> You have the data for where the buildings are but how do you translate
>>> that into a ground texture pattern?
>>
>> I have no idea presently.
>
> Depending on how the data looks like
>
> 1. If you have some sort of geometric data (corners of the boxes
> for the building base or suchlike) you can use it to create somewhat
> larger boxes for use in an object pattern.

In this case the buildings are randomly dispersed within a pre-defined 
object. Writing away the geometric data would be possible, without 
knowing what the final union of buildings would look like however.

>
> 2. If you have some sort of floor plan as an image_map or similar,
> you can use 2d image processing methods to produce an image where
> the building locations are expanded by some amount (e.g. using a
> minimum/maximum filter, depending on the software you have you can
> also simlpy apply some smooothing).

The block object defining the contours of the building union can be 
drawn according to an image_map of street patterns on the underlying 
height_field.

Thomas

Post a reply to this message

Illustrating my answer to MichaelJF in p.b.s-f (Note 2); here using a 
different set of basic building elements (on its way to Gancaloon use).

T

Post a reply to this message

Attachments:
Download 'urbanism2_test.png' (482 KB)

Preview of image 'urbanism2_test.png'

This is an improved version. Note that a next step would be to clean up 
the inner spaces of the houses ;-)

Improved code in p.b.s-f

Thomas

Post a reply to this message

Attachments:
Download 'urbanism2_test.png' (394 KB)

Preview of image 'urbanism2_test.png'

> This is an improved version. Note that a next step would be to clean up
> the inner spaces of the houses ;-)
>
> Improved code in p.b.s-f
>
> Thomas

Now, add the possibility to have non-square buildings and some with 5, 
6, 7 or even more sides. Appen all the time when the neibouring 
buildings are not aligned to some orthographic gird.


Alain

Post a reply to this message

On 03/03/2013 7:25 PM, Alain wrote:
> Now, add the possibility to have non-square buildings and some with 5,
> 6, 7 or even more sides. Appen all the time when the neibouring
> buildings are not aligned to some orthographic gird.

And three sided buildings. Okay maybe, actually four sides but the 
fourth side is very small. This sometime happens when they are built at 
a Y junction.

Or a hypercube.
Only kidding, Thomas. ;-)

-- 
Regards
     Stephen

Post a reply to this message

On 3-3-2013 20:25, Alain wrote:
> Now, add the possibility to have non-square buildings and some with 5,
> 6, 7 or even more sides. Appen all the time when the neibouring
> buildings are not aligned to some orthographic gird.

I'll think about it. The buildings themselves are not the problem and 
can be added by anybody presently. However, if the code has to be 
adapted to their use, that would be a major overhaul and one I am not 
yet prepared to do because it would need a different approach. Here, the 
locations are chosen randomly but apart from testing for too much 
overlap, there is no control towards proper shape intersection.

Thomas

Post a reply to this message

On 3-3-2013 22:16, Stephen wrote:
> And three sided buildings. Okay maybe, actually four sides but the
> fourth side is very small. This sometime happens when they are built at
> a Y junction.

I would love to do that but am presently unable to implement something 
like that :-)

>
> Or a hypercube.
> Only kidding, Thomas. ;-)

Now, this I am taking seriously :-) Hypercubes in a spatial arrangement...

Thomas

Post a reply to this message

Thomas de Groot wrote:

>> In real life, you can have cases when you have prety small elevation
>> differences between roofs or floors, especialy when they belong to
>> different buildings.

but I think its rare within a building to have something like 0.5m
height difference between floor corners, leading to all the non-level
baselines. If its a steep slope it may get an additional floor (2-3m?)
but otherwise I'd expect the floor to be roughly levelled. So the
architecture shapes the ground to some extent.

Post a reply to this message

On 4-3-2013 12:38, Christian Froeschlin wrote:
> but I think its rare within a building to have something like 0.5m
> height difference between floor corners, leading to all the non-level
> baselines. If its a steep slope it may get an additional floor (2-3m?)
> but otherwise I'd expect the floor to be roughly levelled. So the
> architecture shapes the ground to some extent.
>

You'd be surprised... ;-)

Thomas

Post a reply to this message

Le 04/03/2013 09:08, Thomas de Groot nous fit lire :
>>
>> Or a hypercube.
>> Only kidding, Thomas. ;-)
> 
> Now, this I am taking seriously :-) Hypercubes in a spatial arrangement...

It would allow to make a lot of money selling the same house more than
one time, with doors on different side opening to each 3D cube of the
hypercube.
I wonder about the associated taxes: based on external 2D projection on
ground, or actual volumes ?

Post a reply to this message