From: <pgr...@uw...> - 2006-01-26 16:43:21
|
Roberto, I wish I had the most current schema or DTD. I don't have a schema, and the DTD I've been using so far was the ESRI DTD from http://www.esri.com/metadata. That one, however, is not the best one either, because it has some of ESRI's own customizations. I don't have the Altova tools, so it would be a major chore to create a schema from scratch. I used a small java program to validate the XMLs before import, so that's how the languageCode issue came up. The other issues such as 'status' that should be 'idStatus' came up when I was manually editing the XSLT stylesheet for conversion. I abandoned that effort because Jeroen posted his XSLT, so now mine would be redundant. If I come up with something worthwhile, I will be glad to share it. For now, I suppose we have to use what's there. Cheers, --Paul Roberto Giaccio <r.g...@it...> Sent by: geo...@li... 2006-01-26 04:39 AM To pgr...@uw... cc geo...@li..., geo...@li..., Jeroen Ticheler <Jer...@fa...>, Marco Marsella <m.m...@it...> Subject Re: [Geonetwork-devel] Re: ESRI to Geonetwork XML transformation utility In the languages codelist the second "fr" between "mo" and "ms" should be "mr" (Marathi). This was fixed on GeoNetwork 1.x, but somehow came out again in the schema file used in version 2.x, which was generated from the old DTD; I am going to update the file on CVS. If you send us your updated schema file we may use it in the next version; however, we would also have to provide some migration application to update the metadata of the existing sites, which takes some time. Roberto On 25 Jan 2006, at 21:34, pgr...@uw... wrote: > > Well, I came across an Aug 2002 document that shows the corrections > to the standard, where onLineMed was recognized as an error and > changed to offLineMed. I think you are using an old version of the > DTD. Another issue is the presence of 'fr' twice in the codelist > for languageCode within the DTD. This prevented me from validating > the XMLs. I don't know how long it's going to take for the > official standard to be ready, so perhaps it is possible to update > geonetwork? Any comments? > > --Paul > Paul Grzeszczak > GIS Analyst / Programmer Dept. of Earth Sciences > University of Windsor > > > > > pgr...@uw... > Sent by: geo...@li... > 2006-01-25 10:17 AM > > To > Jeroen Ticheler <Jer...@fa...> > cc > geo...@li..., Marco Marsella > <m.m...@it...>, Giaccio Roberto <r.g...@it...> > Subject > [Geonetwork-devel] Re: ESRI to Geonetwork XML transformation utility > > > > > > > Hi, > I was able to export metadata using Template Exporter which is > sample code provided with ArcGIS ( http://edndoc.esri.com/ > arcobjects/9.0/Samples/Metadata/Exporters/Template_Exporter/ > Template_Exporter.htm ). All you need is an XSL. > > The utility you provided did not initially work for me. It did not > generate any output. That's because I had spaces in the path. > > I'm still converting the XSL that I have into 19115DIS compliance. > I'm wondering about the element onLineMed. The ESRI DTD has > offLineMed instead. To me, offLineMed makes more sense. If it's > online, the medium type is irrelevant. Offline resources naturally > must be stored on some medium i.e. tape, CD, DVD. Does anyone find > this odd? > > --Paul > Paul Grzeszczak > GIS Analyst / Programmer Dept. of Earth Sciences > University of Windsor > > > > > Jeroen Ticheler <Jer...@fa...> > 2006-01-25 03:46 AM > > > To > pgr...@uw... > cc > Giaccio Roberto <r.g...@it...>, Marco Marsella > <m.m...@it...>, geo...@li... > Subject > Re: ESRI to Geonetwork XML transformation utility > > > > > > > > Dear Paul, > I hope the utility I posted provides a starting point for the > ArcCatalog export function. I have some other VB code that was > developed but not finished to do similar things directly from > ArcCatalog. If you're interested I could provide these to you. > > We currently use ISO19115 DIS that is still using short element > names for instance. It was a considered choice since there was no > DTD for the ISO19115:2003 and ISO19139 should have resolved that > lack long time ago. However, we are still waiting for that final > schema. We now work with the latest 19139 schema we have and will > use that schema in release 2.1 that is currently in the works. We > will than also have a migration procedure to convert 19115DIS into > the 19139 format. > > On the custom metadata format, a manual will be written explaining > what is involved. The main work will be in creating XSL sheets for > display, editing and indexing (by Lucene). I'm not sure what the > status of that document is, copied Marco and Roberto who may > provide more on that. > > My own observations on the handling of multiple metadata standards > is that it works, but is probably not optimal and manageable in the > long run. Especially when moving towards Catalog services for the > Web for instance. All metadata should at least have a common core > set for these searches to work, but anything more advanced quickly > becomes metadata standard specific functionality. E.g. Dublin core > has no good way to store information on something as simple as a > bounding box, it uses a free text field for that.. > > Hope this helps, looking forward to your contributions! :-) > Ciao, > Jeroen > > On 18 Jan 2006, at 22:25, pgr...@uw... wrote: > > > Jeroen, > I'm building a metadata repository for Windsor and area that would > include geographic as well as biological metadata in EML ( http:// > knb.ecoinformatics.org/software/eml/ ). I've started out with > Metacat ( http://knb.ecoinformatics.org/software/metacat/ ) which > is well suited for EML and have customized it for ISO19115, but it > looks like Geonetwork may be better at handling 19115 thanks to the > editor, etc. However, I'm having difficulties importing metadata > into Geonetwork. A while ago I wrote an XSLT to transform the > metadata exported from ArcCatalog's metadata wizard into XML that > conforms to ESRI's own DTD ( http://www.esri.com/metadata/ > esri_iso01dtd.zip ) but Geonetwork's implementation is different > from that. Geonetwork uses string values for attributes such as > CharSetCd, DateTypCd, etc. as opposed to numeric values (ESRI). > There are some other subtle differences such as <Metadata> vs > <metadata>, etc. I'm curious as to why that is. Is Geonetwork > closer to the proposed standard? AFAIK, there is still no > implementation standard (19139). In time, I'd like to write some > plugin for ArcCatalog/ArcMap that facilitates the transfer into > Geonetwork. Would you have some suggestion with respect to how a > batch import could be done programmatically? I'm not yet familiar > with Geonetwork, as I just installed it, but it looks promising. I > just dwonloaded the CVS tree but didn't look yet. Also, it seems > that Geonetwork should be able to accomodate other metadata > standards (such as EML perhaps). How feasible is such a > customization? > In the meantime, perhaps you could send me (or otherwise make > available) that conversion utility that you mentioned in one of > your previous posts that is supposed to convert the XML to/from > ESRI/Geonetwork? > > Thanks > > --Paul > Paul Grzeszczak > GIS Analyst / Programmer Dept. of Earth Sciences > University of Windsor > > > > ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 _______________________________________________ Geonetwork-devel mailing list Geo...@li... https://lists.sourceforge.net/lists/listinfo/geonetwork-devel GeoNetwork OpenSource is maintained at http://sourceforge.net/projects/geonetwork |