[MapScript-PHP] Class Metadata Object not suportted

[thorne@…]

Class Metadata Object not supported in PHP Mapscript.

It has been observed that the python mapscript has the ability to hold metadata
object within the class object where currently the php mapscript is not supported.

[sgillies@…]

This is an issue much on my mind lately.  See:

  http://mapserver.gis.umn.edu/cgi-bin/wiki.pl?RefactoringMetadataObject

We could avoid implementing metadata separately for Map, Layer, and Class.

[thorne@…]

I agree generally that the Metadata Object should be better defined as to its
functionality and purpose.

I would argue that any OGC Metadata be placed in a separate object called "OGC"
Object, or "MetaOGC" Object or whatever. I would also suggest for ease of
management to have the OGC Object to be able to point to a file containing the
OGC Metadata information. This would be similar to the font and symbol files.

I guess my biggest question is Does the metadata objects at the Map, Layer and
now Class object do they serve any other purpose other than a placeholder for
OGC Metadata?

[sgillies@…]

I have been using custom Map, Layer, and Class metadata for my
mapscript applications.  It is a convenient way to attach
extra information.  I started doing this while the OGC features
of mapserver were very young.

Most users, I assume, are only dealing with OGC metadata.

It would be good to read metadata from a file!  But, this is not how I would
program an application.  I would keep metadata in a relational database
and then I would load values into the Map, Layer, and Class as needed like
this

   # select * from metadata_table where name = 'foo';
   # get result set as an associative array named 'metadata_from_sql'
   for key, value in metadata_from_sql:
       layer.metadata.setItem(key, value)

In summary, I am not picky about the naming of metadata objects or the names
of metadata keys, but I do want metadata objects to be instances of a 
dictionary class that is better for programming.

Were we ever introduced to each other at the users meeting?  I don't remember.
Too many names and faces! 

[thorne@…]

I remembered that we did meet, but I cannot remember what you look like. :)

Maybe you would remember me? I was really the only one with a beard, short dark
hair.

You have a very good point and I agree totally with you about setting up a
metadata server to extract and manage information for the WMS. When I was
writing the last comment I was only thinking of instances were WMS I remembered
that we did meet, but I cannot remember what you look like. :)

Maybe you would remember me? I was really the only one with a beard, short dark
hair.

You have a very good point and I agree totally with you about setting up a
metadata server to extract and manage information for the WMS. When I was
writing the last comment I was only thinking of instances were WMS
administrators don't want to set up a metadata database. They may only want to
manage one or two published data set, which in that case is easier I think to
manage one or two files. 

[dmorissette]

I think this is going off-topic.  This bug is just about exposing class
Metadata, so if you want to revisit the way OGC services are configured then
please use another bug.

I agree in general with exposing the metadata as an object in mapscript as
suggested in the Wiki, but I have some comments and that should be the topic of
another bug as well, see bug 737.

[aboudreault]

The metadata object (hashTableObj) has been added in PHP/MapScript a while ago. See #2773 . Closing this ticket.