Dear all,
I had and email from Bruce F. Donnelly saying the AA site is down.
this is not exactly true
the companion server is off line but the original and true website url is up
http://www.aaschool.ac.uk/sheep/
still points to both the companion server and the back up copy on my
CS home page. The CS home page points to the current down loadable
copy of Meanda.
How ever you might have some problems downloading from this kind of
server. Remeber you
can always email me for the latest version.
Meanda 0.9.8.8 includes some improvements over previous versions
- map naviagtion/pan like Minicad,
- a quick implementation of Steve R's Gradient measure
- an experimental fractional gradient ,
- relitivesed RRA3 and relitivitsed fractional RRA2.2 ( use the log
rra2.2 for best pictures )
- along with RRA5 ( = best pedestrian radius in London ).
- The scattergram is slightly more functional ( you can click on
things and see them selected
in the main veiw).
- A new dialog where you agree not to use this adcademic version for
commercial gain.
- There is no fore ground processing option now, all processing
happens in the background, so dispite
long processing times you can still work comfortably on your
machine with the processing
happening out of site.
-New measures are exported in the table
-If you have axman 2.6a you can export all information includeing
observations and these can be
read into Meanda directly. This simplifies processing and makes sure
the observation values
are directly connected the correct axial lines.. Axman 2.6 can also
exports the unlinks for a
map which will be automatically read into meanda ( seee unlinking below)
Old versions of Meanda will now self destruct to preserve the code base.
yet to be done
-------------
What will be available in Meanda 0.9.8.9
Line Unlinking permtting the introduction of overhead intersections.
( the menus exist but don't do anything yet).
Automatic sub-segmentaion of ultra-long axial lines. Useful for lines
which are longer than the average journey distance in a city.
Some kind of internal file format storage makeing a new file type.
PC VERSION
-----------
I have had the very occasional request for a PC version. No one
seemed much intreasted in contributing to the purchase of PC upgrade
to the compiler. The
conclusion is that PC people want a copy of axman but not enough to
stump up the $10 nessasry to create
it. How ever the digital artist Lucy Kimble (
http://www.lucykimbell.com/ ) thought the
project was intresting enough to donate the $200+ dollars necessary
to purchase the compiler
upgrade. So thanks to her there is an experimental PC version - which
is available for testing
(email me [log in to unmask] for a copy). You might want to email
her and thank her for her
contribution to syntax-kind. I have no idea if the PC version will
work or if it will produce the
same kinds of results as Axman/MacMeanda. So I would like some
feedback on this. I also have
no idea what you might want to use to create the lines on a PC how
ever it might be possible to use MiniCad/Vector works Text output
from the PC version of Minicad. Don't be supprised
if the PC version is slower than the MacPPC version.
Note for GIS people Meanda will accept fractional coordinates - how
ever you might want to make sure your coordinates are in a flat frame
of referance ( same
warning goes to Ovation users ). If you have your coordinates in
Longitute/Latitude format you will
notice lines will not cross towards the edge of the map - the is
caused by the curvature of the earth
which is significant over the size of a large city.
Fractional Radius
---------------
The biggest step forward since the last version of Meanda is the
existance of the radi measures. this was due to the theoretical
problem of what radius means. I
didn't have time to say this in my talk but a fractional radius
leads to some questions about what is radius ?
Is it a way of saying 'journeys are only this long in a city and
time is roughtly proportional to step distance'. Or is it a way of
saying 'People don't like makeing complex journeys more than this
amount'.
With factional anaylsis we are definitly saying 'people don't like
moving though more than this kind of angle'. The next problem is the
more mechanical of how
do you relitivise for graph size. For example with intergration we
apply the RRA and Dvalue
formulas to correct the intergration factor for the size of system. I have some
long standing problems with DValue ( this would be a good PHD topic
for someone), how ever
when we are looking at fractional depth the Dvalue does not make any
sense what's so
ever. Currently Meanda is computing RRA by counting the number of
lines which are less than
depth 2.2 ( or what ever the fractional radius is set to ) and using
that to compute the
RRA formula as would appear in traditional intergration computations.
It seems to work but
I am open for other ideas on how to compute this kind of
reltivisation factor ( again
another good PHD topic ).
Again I havn't had time to look at the radi measures or check them by
hand. Correlate
at your own risk.
Other software
--------------
If you are intreasted in looking at organisations I am intrested in
collaborating. I have some new software which looks at networks of
people like networks of
spaces. I have a program which computes a number of measures from a
network constructed from
the email and phone logs for an organsation. You can also type these
in by hand if
you have questionnaires If you have this kind of info then contact
me. I was especially
enthused in this direction given Bill's concluding speech at the
conferance ( the interaction
network of people bit). If you have this kind of data then email me.
thanks
sheep
[log in to unmask]
|