Hi Martin,
Here below, some answers to part of your questions.
On 9/7/07, Martin Desruisseaux <[EMAIL PROTECTED]> wrote:
>
> General note: some nodes begin with a lower case, and some other nodes
> begin with a upper case. It may be worth to clarify what is the policy
> for lower/upper cases. In "GML in JPEG 2000" specification, I have the
> feeling that lower case is used for attributes and upper case used for
> classes.
>
> Page 18 said:
>
> > Figure 5 depicts the Stream MetaData schema overview. The root node
> "StreamMetaData" has
> > two children:
>
> If I'm understanding correctly the IIOMetataFormat specification, the
> root node name should be the specification name. For example the root
> name for J2SE standard metadata is "javax_imageio_1.0". For coverage
> I/O, it could be something like:
>
> geotools_imageio_1.0
Indeed, in my implementation, I used a similar named node (followed by a
"//TODO: choose a proper name") as the root node, having a StreamMetadata
child. Maybe this could seem as a redundancy... We could set a proper name
containing the "stream" word to distinguish this from the other metadata
types.
Not sure if we should insert the "stream" word somewhere or not...
Page 19 said:
>
> > the "axis" node reports the name of the bands associated to the coverage
> (for
> > instance in the above example, Wind Velocity can have the U and V
> bands).
>
> I'm a little bit surprised by the "axis" name, since it could be
> confused with CRS axis. In my reading of the above sentence, "axis"
> seems to be used in the sense of "SampleDimension" in OGC 01-004. Is the
> "axis" name part of some specification for this purpose? A quick search
> in "GML in JPEG 2000" show me "axisName" used in the usual sense of CRS
> axis...
I think Alessio "imported" this name from a WCS context, where "rangeSet"
contains several AxisDescriptions elements. I'm not really experienced on
the WCS topic :D
In effect, it could be a little confusing and a rename could be helpful.
> Page 21, figure 6:
>
> bandKeys
> --------
> See previous paragraph. I don't understand why band informations appear
> as axis properties. I would expect them as coverage-specific information
> (on "bandKeys" for each coverage), and I would expect them to be named
> "sampleDimensions" instead.
>
>
> singleValue
> -----------
> Do we really need a "singleValue" element for each vertical position?
> IIOMetadataFormat cleary allows storage of arrays. See for example:
>
>
> http://java.sun.com/javase/6/docs/api/javax/imageio/metadata/IIOMetadataFormatImpl.html#addAttribute(java.lang.String,%20java.lang.String,%20int,%20boolean,%20int,%20int)
>
> Then, why not store the vertical values straight into <values> (child of
> <VerticalExtent>) as a space-separated list? Same remarks applies to
> <timePosition>.
min,max,res
> -----------
> Could we use full names instead?
Off course. This document started as a simple draft and it has been
progressively adjusted by Alessio, me and Simone.
Refinements and improvements are welcome :D
--
Best Regards,
Daniele Romagnoli
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
