Bill,

I sort of disagree here.  It would be nice if there was a hierarchy of
points, arcs and polygons and all of those modelling niceties were there,
but these were designed back when computers had less grunt than my
calculator.  Is it a big advantage now?

Now, most of those sorts of calculations can be done by the software on the
fly, generally in less time than in the old days.

Arc's topology make it a good data capture tool, especially if you want to
avoid getting slivers etc between polygons - but if you want to follow the
same procedures, then surely there are some other good clean and build
tools.

I think that a far more important design feature is the intimate connection
between the spatial object and the data about it.  After all, arguably the
most widely accepted spatial model is Oracle Spatial, and to my knowledge,
it stores the spatial object as part of the data, and lets the software
worry about topology.

I'm sure that this connection to the data makes the whole integrated spatial
and data query tighter.

R



-------------------------------------------
Robert Crossley
Agtrix P/L Australia
 
Far Southern Queensland Office:
9 Short Street
PO Box 63 
New Brighton 2483
 
P: 61 2 6680 1309
F: 61 2 6680 5214
E: [EMAIL PROTECTED]
W: www.agtrix.com
 
Brisbane Office:
109 Milsom St
Cooparoo  4151
Queensland
P: 61 7 3843 3363
 
 
 
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Bill Thoen
Sent: 23 December 2005 10:16
To: MapInfo List
Subject: Re: [MI-L] Fixes and New MapInfo Features

Søren Breddam wrote:

>This shouldn't be hard to implement.. Topology is obviously more difficult.
>If the topology issue was implemented, we could sell our license to
Microstation ;-) And do even more interesting queries.
>  
>
Well, in the beginning, MapInfo decided to use a simple model that 
didn't need the rigorous set up that topology requires. When MI came 
out, it was trivial to make a map -- you just drew what you wanted and 
presto! You had a map. From the very beginning they were after the 
business market, not the science people. In fact MapInfo's first product 
was more like a pin-map tool to locate customers, franchises, etc. i.e. 
strictly business. It's also why they've resisted being called a GIS, 
opting instead for "desktop mapping." Business people don't know GIS, 
and don't wan to know GIS, but they're bang alongside things that are as 
accessible as a "desktop." MapInfo basically made a tool that could 
associate data with drawings, which is actually a fairly powerful concept.

The alternative was Acr/INFO. To use that, you needed to understand GIS 
at the techno-weenie level and build your map objects by first 
establishing the nodes, then snap the arcs to those (assigning to and 
from nodes) and then build polygons by assigning ids to the left and 
right sides of the arcs. To do all this properly generaly took some 
time, but in the end, you were assured that operations like dissolving 
smaller areas into larger ones, or finding adjacent polygons or 
traversing a network all became pretty straight-forward when you could 
use arc-node topological math.

Personally, I prefer ESRI's model because I like the internal 
consistency that topology adds. I also like their idea of associating 
style information with data attributes rather than making it part of the 
map data. That makes it easy to select information from the data with 
SQL whereas in MapInfo, if you want all the blue objects for example, 
you can't use SQL.  But I must admit, there are times when MapInfo's 
model is just so much simpler to implement and sometimes a map *is* just 
a drawing with some data attached and you don't need to do anything with 
topology.

My point is that I don't think MapInfo needs to be more "GISey" as much 
as they should pay attention to what their core market and focus is (or 
was.) It was business information analysis and presentation of spatial 
data. That means not only do they need good analysis tools (and data) 
but they really need to sex up their graphics presentation capability.  
They aren't aimed at doing modelling or network analysis so they don't 
*need* topology. But they do need better graphic tools so that the 
software's output can blow the collective socks off an audience.

_______________________________________________
MapInfo-L mailing list
[email protected]
http://www.directionsmag.com/mailman/listinfo/mapinfo-l



_______________________________________________
MapInfo-L mailing list
[email protected]
http://www.directionsmag.com/mailman/listinfo/mapinfo-l

Reply via email to