Christian Mayer wrote:
> What I've seen is that there are only two road types. I think this could
> be limiting in the future. (As line data can dramatically increase the
> triangle count it might be necessary to switch off the smallest visible
> roads.)
Yep, this a tribute we're currently paying to conformance with the
'historic' schema in TerraGear. The two protagonists in the area of
handmade landcover data, Ingrid/Ralf on one and Thomas (Foerster) on
the other side, favour different approaches in the handling of line
data. Both produce favourable results, as you know. This is why a more
modern approach is not yet ready (and beause of time contraints ....).
On one hand you can simply stick to an idea in the sort of what you
proposed by defining five or six categories of roads and stuff
everything into that schema. This makes it easy for everyone who aims
at _handling_ that data. The same applies to any sort of water flows.
Certainly such a schema wouldn't be hardcoded in software but instead
we'd use a template file - here's a smple of what Ralf uses:
##### Typnummern für Liniendaten ##############################################
# Art Typnummer Quelle Name AreaType Breite
l 10 strassen Autobahn Freeway 14m
l 11 strassen Schnellstrasse Freeway 10m
l 12 strassen Hauptstrasse Freeway 6m
l 13 strassen Nebenstrasse Road 5m
l 14 strassen Nebenstr_schmal Road 4m
[...]
This template file would be located at a central repository,
preferrably in a database table, and everyone dealing with landcover
data (necessarily including users of TerraGear !!!) could/should load
the content of the template into his favourite software. If something
breaks, we all know that we have to shoot the maintainer of the
template :-)
Such an approach has several advantageous side effects, as you could
easily attach some eye-candy to each class. For example you could make
good estimates on what colour a river is supposed to have. You could
assign a central reserve to every road of 11 m and more. One day you
could even estimate how many AI-cars would drive down a little by-road.
_But_ this is probably not the best approach for the group of those who
_create_ and _use_ the data. This is especially true for rivers but for
roads as well. If you stuff everything into a fixed schema of different
widths you are supposed to loose many interesting details. Variable
line width is the key here - but you are at risk to loose structural or
logical information that might be very valuable next year as
Terra-/FlightGear develop. Nonwithstanding a common, central template
would be necessary in case you use variable line widths, the discussion
is just about in which pattern to store the details/attributes.
I see an analogy to the way airport layouts are being defined. There
has been a schema that did the job for a while, thousands of hours of
workforce have been poured into this to create taxiway layouts for a
significant number of airports that please your eye. Now, after the
discussion already lasts for several years, people can't avoid facing
the fact that the schema doesn't match the needs because it is
incapable of reproducing the logical layout of an airfield (except the
main runways) and major parts of the eye-candy in the taxiway layouts
have to be redone once people have agreed on a modern schema.
I wouldn't want to make such a mistake with a new landcover data schema
and I'm happy for proposals that help closing the gap and are mature
enough to stand real-world use. I would like to dispel the assumption
that we, especially Ralf, Thomas and I have been deedless in terms of
this topic, it's just that a long-term solution is somewhat tricky.
Aaaah, I intended to post such an 'essay' to this list so many times,
but never did so. Sunday morning before breakfast is definitely the
best time for such an undertaking,
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel
