Re: svg comment - coordinates

Michael Gould wrote:
> 
> Hello.
> 
> Those of us who work with geographic information --and hope to view our maps
> in SVG-- would like to suggest explicit treatment of real-world coords,
> within chapter 8 of the working draft. We'd like to see the ability to
> transform outside the CSS realm (points, mm, inches) to metres, kilometres
> etc.

Are you asking for metric scaling, ie the ability to specify a physical
dimension for the map as CGM does?

> This allows users to do things like measure real-world distances
> on-screen (for example, click on origin and destination on a SVG map, and
> get the crow's-flight distance), calculate polygon areas...

Ah. To answer my own question, no you are not. I certainly the
requirement but please do not confuse map kilometers with CSS
millimetres. They are referenced in different coordinate systems.

You can certainly set up your *local* coordinate system in SVG so that
the numerical values used for coordinates correspond to real-world
values such as kilometers, decimal lattitude/longuitude, grid
references, and so on.

The transformation matrices eventually map this to an actual size on the
display device (in points, mm, percent, whatever).

The aim (for others who are reading along) is to make the numerical
values in the SVG meaningful, rather than to display the graphic at a
particular fixed size. This means that the XML DOM can be used to query
the position of a selected object and to present that numerical
information (such as  lattitude and longitude) to the user


> Maybe real-world coords, units etc. are in there somewhere, but it's not
> obvious reading over the current SVG draft.

You are right; they are implicitly there but, beinga first derft, the
specification is short on examples. It would be good to add some
examples such as engineering drawings, maps and so on which can make use
of local coordinate systems to preserve meaningful numerical data.

> 
> By the way, my colleague has had a VML-format map server up for over a month
> now:
> http://www.imapper.com/iindex.htm

Neat. Am I correct in thinking that it makes a trip back to the server
each time one zooms (in or out) or pans? I had a little difficulty
looking at the source because of the number of nested framesets, but I
found it eventually and it made an interesting real-world sample.

I would be very interested to see a brief write-up or feasibility study
on the conversion of this system to SVG - are the features you need
present in the SVG specification?

--
Chris

Received on Tuesday, 23 February 1999 14:32:50 UTC