Hi Bruce and Jeroen,
It is very interesting so see others looking into the question of supporting
ISO 19115 "profiles". Sorry that my email doesn't pretty up the comments.
You must be using M$ Outlook, HTML format or some other fancy email format.
;--)
Please see my two cents worth of comments below: ;--)
Thanks.
John
-----Original Message-----
From: geonetwork-users-bounces@lists.sourceforge.net
[mailto:geonetwork-users-bounces@lists.sourceforge.net] On
Behalf Of Jeroen Ticheler
Sent: Thursday, 8 June 2006 6:15 AM
To: Bruce Westcott
Cc: geonetwork-users@lists.sourceforge.net
Subject: Re: [GeoNetwork-users] GeoNetwork-users Digest, Vol
1, Issue 92Hi Bruce,
On Jun 5, 2006, at 10:51 PM, Bruce Westcott wrote:
Jeroen et al.:
The difficulties importing ESRI_ISO.XML (or any other XML
which purports ISO compliance) has to do with two larger
issues, I think:1) How "flexible" is the XML import/parsing process in
GeoNetwork, and how can it be made as tolerant as possible
for importing XML files which are "kind of" ISO-19139-ish?
(Sub-issues: how will the parse/import process deal with
different classes of "problems" such as missing elements,
extended elements, schema differences, mismatched data types, etc.???)
The import process is flexible in the sense that you can add
XSL files to convert an incoming XML format into one that
will validate against the schemas as used by GeoNetwork. In
version 2.0 these are FGDC, ISO19115 DIS and Dublin Core. In
version 2.1 we have added the ISO19139 schemas.
The purpose of ISO 19115 profiles is to allow for cut down or expanded
versions of the ISO 19115 metadata standard to suite a country, community of
practice or an organisation's needs. There is supposed to be a way of
registering profiles with ISO but we haven't quite worked that out yet.
Hopefully this will be resolved at the latest meeting.
Here are two reasons for registration:
1. It identifies one location (ISO) for all accepted profiles so that someone
can first check to see if there is an existing profile that may suit their
needs before creating another. That is, it would be best to have only on
Marine Profile rather than hundreds of Marine Profiles and the ISO official
profile web site should reduce redundant profiles.
2. The profile can be checked for compliance to the ISO 19100 standards by
ISO before ISO accepts it. There are many rules for ISO 19115 profiles and
ISO can check that a profile meets these rules before it is accepted.
All profiles should have XSDs that are used to implement that ISO 19115
profile. They must be valid and publicly available. All XML metadata
documents should have references to the full URL of those profiles in the
"schemaLocation" attribute. This will allow any application to access the
XSDs and then validate the XML document when it is being imported. This will
check the validity of the XML metadata record to the related profile. No
document should be imported into GeoNetwork unless it is valid against its
appropriate profile's XSDs.
All profiles should also provide an XSL that allows the translation from the
profile's XML document into the ISO 19139 XML format. Once the translation
of the XML document is complete then the resulting document can be validated
against the ISO 19139 XSDs. This will prove compliance to the ISO 19115
standard. This should always be done when a tool is importing XML metadata
records.
Eventually, I would hope that GeoNetwork will allow the creation of any valid
profile metadata using the XSDs, XSL, XML etc. configuration files. That
means that GeoNetwork will not only allow the creation of XML according to an
identified profile but also translate XML into the ISO 19139 format.
2) What version of ISO-19139 will be incorporated into
GeoNetwork, and how easy will it be for users to
upgrade/modify that version? Will GeoNetwork contain and
rely on the entire schema, which can then be altered, or will
there be complex logic that interprets the schema (and
therefore would have to be modified and tested when the
schema changes)?We have used the latest version as available from the site
http://eden.ign.fr/xsd/isotc211/index_html?set_language=en&cl=en
We will for sure need to update these while updates come out
But release 2.1 is planned for release only by the end of
September, so there is more time to update things in the mean time.
We have created a migration application to migrate from one
version of GeoNetwork to the next. This will convert all
metadata in the catalog through an XSL transformation and
report issues it encounters in that process (you can just do
a test run first). We have had that migration procedure
from version 1 to version 2 and now there's one that will
migrate the ISO19115 DIS to ISO19115:2003 validated against
the 19139 schema.
It's not very difficult to add other versions of schema's to
the application (not very quick either as it involves adding
and updating of a range of XSL stylesheets.
I would hope that the ISO 19139 XSDs use version control to distinguish the
different versions. If the current XSDs do become the official XSDs for ISO
19139 when it is published then the existing GeoNetwork work should be fine.
If another ISO 19139 version is released by ISO then the old XML documents
will still be valid as long as their namespaces refer to the original ISO
19139 version. ISO *should* provide an XSL that translates from the old to
the latest ISO 19139 XSD versions.
When an XML document is translated from any profile or ISO 19139 version into
another then the namespace should be changed to reflect which schema the XML
document is valid against. It should not be necessary to convert from an
official profiles format into the latest version unless the user wishes to do
so. However, all *new* XML documents should use the latest version in its
namespace and for validation purposes.
Again, translation should be by using an XSL and the profile should be
represented in GeoNetwork by XSDs, XML configuration files etc.
There is an unofficial version of the ISO-19139-TS schema
posted on the web; as I understand things, it is the version
forwarded to the 22nd plenary meeting of ISO/TC 211 held in
Orlando, FL (US) in May. ISO-19139 was scheduled for
discussion at the meeting, but there were substantial
comments/objections documented by several countries during
the DTS review period.
Yes, everyone is desperately waiting for a final version, and
so do we
migration path should facilitate this transition phase.
I agree. Also the next version of GeoNetwork, after the official release of
the ISO 19139 or another version of ISO 19139 XSDs, can cater for those new
XSDs.
I have not heard results of the May meeting, but I expect
that there is still some work to be done to gain consensus.
Bottom line -- my best guess here -- a truly "final" TS
version of ISO-19139 is in gestation NOW (following the May
meeting), will emerge from ISO sometime in coming months, and
may take some time from that point to stabilize. (Don't
forget the corrigenda: certain some problems will be noted as
people go to work implementing applications.)
Most of the corrigenda were a result of discrepancies found while trying to
implement ISO 19115 in the ISO 19139 XSDs. It seems to my initial
investigation that the ISO 19139 XSDs do represent ISO 19115 after the
corrigenda are applied.
So if ISO-19139 is truly the Holy Grail of interoperable,
exchangeable metadata, how long will it take all users to
stabilize on a "final" version, and to modify their
applications to include import/export functions that are
somewhat forgiving?
Who knows. I hope the CSW 2 and ISO19115 profile will help a
little in that respect to provide one of those intermediate
solutions while stabilizing.
It doesn't really matter. As long as there are official profile XSDs that
truly represent the ISO 19115 profiles and that there are XSLs that translate
from each profile to the ISO 19139 format then users have already taken up
the ISO 19115 standard.
Greetings,
JeroenBruce Westcott
Geospatial Metadata Consultant
Marshfield, Vermont 05658 -- USA
802.426.3344