[Geodata] meeting narrative

[Jeroen Ticheler]

Dear Jo,
Good to hear from you!

On Jun 17, 2007, at 8:52 PM, Jo Walsh wrote:

> dear Jeroen,
>
> I've read about MEF in the geonetwork server reference, but confess i
> skipped over it, seeing it as something for internal use only.

That's how it started for sure, providing a way to back up and move  
data between GN nodes :-)

>
> On Fri, Jun 15, 2007 at 12:42:39AM +0200, Jeroen Ticheler wrote:
>> We're working on packaging data and metadata that than can be
>> directly imported in e.g. a GeoNetwork catalog. We've defined a
>> packaging format, very light weight, that contains some meta-meta
>> information, the actual metadata, preview images, basic privileges
>
> This makes a lot of sense up to "basic privileges". Can you expand
> this? Are you using the MEF to pass user credentials between
> geonetwork instances? Or do you mean more like license constraints?

Credentials between systems. It's pretty basic: you can define if the  
metadata you import should be immediately visible on the Intranet and/ 
or Internet, if the download and interactive map buttons should be  
visible for those groups. Just those things. It can save you time  
because you don't have to go to the web interface to set those  
privileges by hand. Still, an internal workflow process in an  
organization may require someone to for instance do quality checking  
before a data is visible on the Internet.
License constraints should be written/ referenced in the metadata  
itself where needed, MEF doesn't deal with that.

>
>> We're also discussing with the gvSIG folks to develop a bot that will
>> search a directory, generate metadata and create MEF files that can
>> than be harvested by GN
>
> So gvSIG will consume these data packages, and emit them.
> When you say, 'then be harvested', what interface are you using?

The current idea is to do that from WebDAV folders, possibly people  
want to use the abstract framework available in GN to implement  
harvesting from FTP folders or other types. There's a GUI in GN 2.1  
that allows you to schedule such harvesting from a (remote or local)  
URL. No rocket science really. No CSW or so as that would not work I  
think (!?).

> Will this become the 'best' way to submit metadata to a GN node via a
> machine interface?

I think yes, as it contains all that's needed to publish a data set.  
It probably needs some expanding to for instance automatically deploy  
a data file on GeoServer, MapServer or another map server. I imagine  
for instance that SLD documents can be uploaded with the file that  
can be used by a map server for the symbology. Also the file types  
uploaded and the way they are packaged within the MEF need some best  
practice approach so GN and other apps would know what to expect.

>
>> The format is not the solution to everything, but for our cases it
>> works. It's been described in our documentation and we're open to
>> further improve/extent it.
>
> It is great to see something in this space emerge from succesful  
> usage :)
>
>> Maybe something to consider for the data dump bucket!?
>
> Definitely, especially if more clients are consuming the format by
> then. What's holding you back from socialising this more, e.g. to uDig
> or QGIS - waiting for proof of concept first? or more data out there?

Nothing is holding back, its just those projects uptake and probably  
us telling more about it/ get more people involved like we did with  
gvSIG.

Ciao,
Jeroen

>
> cheers,
>
>
> jo