[Mapserver-dev] MapServer and OGC services

[Tom.Kralidis at CCRS.NRCan.gc.ca]

> -----Original Message-----
> From: Daniel Morissette [mailto:morissette at dmsolutions.ca]
> Sent: Monday, April 21, 2003 6:16 PM
> To: Tom.Kralidis at ccrs.nrcan.gc.ca
> Cc: Mapserver-dev at lists.gis.umn.edu
> Subject: Re: [Mapserver-dev] MapServer and OGC services
> 
> 
> Tom,
> 
> I would tend to agree with you on the fact that there is an 
> issue there,
> but unfortunately I don't think the solution is simple.  When I
> implemented the WFS it really bugged me to have to create duplicate
> metadata names for things like wms_title and wfs_title, and I thought
> about coming up with something more generic such as ows_title (OGC Web
> Service Title), etc.  Perhaps I should have done it, but I wasn't
> convinced that this was good enough to justify breaking with 
> the past so
> I just ended up going the easy route and went with wfs_* 
> metadata names.
> 
> Note that in some cases someone might want to use different values for
> some parameters depending on server types.  This may not be 
> very common
> for layer names and titles, but for instance the SRS don't 
> work the same
> way in WMS and WFS, so someone might want to set different values for
> wms_srs and wfs_srs.
> 
> Perhaps the best would be to allow a combination of both, i.e. if you
> set "ows_title" then that applies to any OGC web service, but you have
> the option to set server-specific values using the service name as
> prefix, e.g. "wms_title" or "wfs_title" in this case.
> 
> So to reuse your example, you could have:
> 
>  LAYER
>   NAME roads
>   METADATA
>    "ows_title" "road network"
>    "wms_abstract" "Abstract that applies to WMS only"
>    "wfs_abstract" "Abstract that applies to WFS only"
>   END
>  ...
>  END
> 
> 
> What do you (and others) think?
> 

Sounds like a plan.  Just wondering whether / why someone would actually
generate different type of metadata, etc. about the same data binding.

> Another question: do we need the "ows_" prefix or not for the generic
> metadata value?  It may be nice to have to indicate that this metadata
> applies to OGC web services, but OTOH we may end up with the same
> problem later on when we decide to support other uses for this
> metadata.  I'm really not sure what's best.
> 

I suggest sticking to a nomenclature which means something to MapServer, so
that the mapfile is abstract, but as interface specifications increase in
usage, code can farm out the information given the spec (i.e. don't look for
"wms|wfs" type tages, yet publish generic tags structured as per the spec).
This also refers to non-OGC services, such as ISO, etc., so something like
"ows_" may be a bit specific.

The bottom line is keeping it simple for the user and not have them update
all their information holdings in a mapfile once the new spec of the
month/year is supported.  Just a thought.

..Tom


> Daniel
> 
> 
> 
> Tom.Kralidis at ccrs.nrcan.gc.ca wrote:
> > 
> > Hi,
> > 
> > I'd like to provide a few comments w.r.t. MapServer current 
> and forthcoming
> > support of OGC services, and development of them.  This 
> comes after I set up
> > a WMS with MapServer, and was looking to set up WFS as well.
> > 
> > Currently, MapServer provides WMS publishing by means of 
> directives through
> > the mapfile, whose structure is basically "wms_<parameter>" "value".
> > Populating these fields produces the appropriate 
> information to respond to
> > an OGC WMS GetCapabilities request.
> > 
> > According to the following wiki:
> > 
> > http://mapserver.gis.umn.edu/cgi-bin/wiki.pl?WFSMapServer
> > 
> > ..one can have WFS functionality also by using a similar 
> nomenclature, i.e.
> > "wfs_<parameter>" "value", in various areas of a mapfile.
> > 
> > Having said this, I find myself adding directives with the 
> same information
> > in order to publish as WMS and WFS, i.e.
> > 
> > LAYER
> >  NAME roads
> >  METADATA
> >   "wms_title" "road network"
> >   "wfs_title" "road network"
> >  END
> > ...
> > END
> > 
> > While this may not be a huge deal, given the limited OGC 
> services supported
> > by MapServer, what happens when MapServer supports other 
> OGC services,
> > current and / or forthcoming?
> > 
> > Perhaps an abstracted model and approach should be put 
> forth in the mapfile?
> > i.e.
> > 
> > LAYER
> >  NAME roads
> >  METADATA
> >   TITLE "road network"
> >   ABSTRACT "This is a road network"
> >   ...
> >  END
> > ...
> > END
> > 
> > This would streamline the approach for those populating 
> mapfiles; also the
> > software can fork out the information, depending on the 
> type of request it
> > gets (i.e. services=wms|wfs|...).  I guess what I'm saying 
> is that thought
> > should be given to MapServer working with this information 
> independently,
> > and not having to add parameters according to specs, yet 
> publish information
> > as per the spec using the internal model of the mapfiles.
> > 
> > Comments?
> > 
> > ..Tom
> > 
> > =================================
> > Tom Kralidis
> > Systems Scientist
> > Canada Centre for Remote Sensing
> > Tel: (613) 947-1828
> > http://www.cgdi.ca/
> > =================================
>