T08_InternationalCollaboration: RAL 07032007.mmap58150.mm

File RAL 07032007.mmap58150.mm, 14.4 KB (added by lawrence, 12 years ago)

Freemind version of Jeremy's mindmap, slightly lossy from the original

Line 
1<?xml version="1.0" encoding="UTF-8"?>
2<map xmlns:pri="http://schemas.mindjet.com/MindManager/Primitive/2003" xmlns:ap="http://schemas.mindjet.com/MindManager/Application/2003" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:cor="http://schemas.mindjet.com/MindManager/Core/2003" version="0.8.0">
3<node TEXT="RAL Summit - the 'Feature Workshop'" POSITION="left">
4<hook NAME="accessories/plugins/NodeNote.properties">
5<text>The aim of the workshop is to explore opportunities for convergence in &lt;br&gt;interoperability and data modelling for atmospheric, meteorological, and &lt;br&gt;oceanographic scientific data. We should aim to reach agreement on a modelling &lt;br&gt;approach and a roadmap for building interoperable tooling.&lt;br&gt;&lt;br&gt;+++++&lt;br&gt;&lt;br&gt;Figuring out how to share atmospheric science information with non- &lt;br&gt;atmospheric scientists ... knowledge transfer to the wider community&lt;br&gt;&lt;br&gt;Aim to build a conceptual model that's independent of the persistence format&lt;br&gt;&lt;br&gt;+++++&lt;br&gt;&lt;br&gt;Attributes AND operations</text>
6</hook>
7<node TEXT="Agenda" POSITION="right">
8<hook NAME="accessories/plugins/NodeNote.properties">
9<text>Wed 7 Mar&lt;br&gt;&lt;br&gt;=========&lt;br&gt;&lt;br&gt;* 10:00-13:00 Introductory presentations x 4 ('Unidata', 'NERC DataGrid', &lt;br&gt;Stefano, Jeremy)&lt;br&gt;&lt;br&gt;- Introductory presentations should present background material to get &lt;br&gt;everyone on the same page.&lt;br&gt;&lt;br&gt;- Aim for 1/2hr + 15mins discussion each.&lt;br&gt;&lt;br&gt;(Lunch)&lt;br&gt;&lt;br&gt;* 13:45-14:30 Introductory presentation (Simon)&lt;br&gt;&lt;br&gt;* 14:30-17:30 Agree major information classes (feature types)....</text>
10</hook>
11</node>
12<node TEXT="Day 1" POSITION="left">
13<node TEXT="Introductory presentations" POSITION="left">
14<node TEXT="A Woolf" POSITION="left">
15<hook NAME="accessories/plugins/NodeNote.properties">
16<text>Explicit sematics offer significant potential for advanced processing workflows &lt;br&gt;...&lt;br&gt;&lt;br&gt;Conventions-based approach is fragile ... you need to be 'in the club' to &lt;br&gt;understand ... implied semantics&lt;br&gt;&lt;br&gt;[Caron] All people use conventions; but they don't always declare them&lt;br&gt;&lt;br&gt;[Tandy] Do we have an uphill struggle persuading people to use explicit &lt;br&gt;semantics when there is a lack of concern over the use of existing &lt;br&gt;conventions? We need to provide some carrots to ensure widespread adoption. &lt;br&gt;(or find sticks ... like INSPIRE? Requires a standards-based approach to &lt;br&gt;Atmospheric &amp; Oceanographic themes.) ... part of the roadmap?...</text>
17</hook>
18</node>
19<node TEXT="J Caron" POSITION="left">
20<hook NAME="accessories/plugins/NodeNote.properties">
21<text>THREDDS data server - unidata's common data model; background / summary&lt;br&gt;&lt;br&gt;[TDS has 4 services; WCS, OpenDAP, HTTPServer, NetCDFSunset]&lt;br&gt;&lt;br&gt;THREDDS Catalogue: xml over http, hierachical listing of online resources, &lt;br&gt;container for arbitrary search metadata, metadata can be inherited within the &lt;br&gt;hierarchy, data access via URL, levels in hierarchy linked via hyperlinks &lt;br&gt;(allowing a search-engine to crawl the catalogue)&lt;br&gt;&lt;br&gt;Galeon 2; upgrading to WCS 1.1 &amp; will attempt to return point datasets ... but &lt;br&gt;not clear how to do this currently. [data model will be CDM point data set &lt;br&gt;serialised into netCDF ... CF doesn't cover point data sets ... looking to &lt;br&gt;CF1.1 for standardised profile ... BUT the WCS metadata (describeCoverage)...</text>
22</hook>
23</node>
24<node TEXT="S Nativi" POSITION="left">
25<hook NAME="accessories/plugins/NodeNote.properties">
26<text>WFS cited as possible service implementation ... but this has NOT been done.&lt;br&gt;&lt;br&gt;[Scientific Features are Sampling Features]&lt;br&gt;&lt;br&gt;Coverage feature acts as proximate FeatureOfInterest [O&amp;M] ... there is an &lt;br&gt;implied feature whose shape is the coverage domain extent&lt;br&gt;&lt;br&gt;Scientific Feature Types are (nothing more than) ExtensiveSamplingFeatures; &lt;br&gt;these are implied within ISO19123 ... the coverage sub-types have different &lt;br&gt;'domain' types ... need to relate these domain types to the scientific / &lt;br&gt;sampling features. &lt;br&gt;&lt;br&gt;FeatureOfInterest is application specific ... example: a swath [sampling &lt;br&gt;feature] is used to observe land use [feature of interest]...</text>
27</hook>
28</node>
29<node TEXT="J Tandy" POSITION="left">
30<hook NAME="accessories/plugins/NodeNote.properties">
31<text>"explore opportunities for convergence in interoperability and data modelling &lt;br&gt;for atmospheric, meteorological, and oceanographic scientific data. We should &lt;br&gt;aim to reach agreement on a modelling approach and a roadmap for building &lt;br&gt;interoperable tooling."&lt;br&gt;&lt;br&gt;Why do I care? Because I live in a world where atmospheric science is only &lt;br&gt;part of the picture ... [Nativi "society infrastructures, platforms and &lt;br&gt;systems"]&lt;br&gt;&lt;br&gt;WMO viewpoint not discussed explicitly ... but we should not forget our own &lt;br&gt;history; GRIB, BUFR etc. this is not going away anytime soon.&lt;br&gt;&lt;br&gt;Content vs Container; Information vs Data...</text>
32</hook>
33<node TEXT="Issues" POSITION="left">
34<hook NAME="accessories/plugins/NodeNote.properties">
35<text>Issues from Jeremy's presentation&lt;br&gt;&lt;br&gt;# profiling specifications for specific implementations ... profiling is a key &lt;br&gt;tool to achieve interoperability&lt;br&gt;&lt;br&gt;# constraining the query model ... to limit the type of requests that people &lt;br&gt;can execute ... and tying those query models to roles / access control&lt;br&gt;&lt;br&gt;# we can't forget about the implementation gap ... most current WFS &lt;br&gt;implementations locally map (rdbms) table-space to GML application schema (and &lt;br&gt;users are mostly happy with this; generally speaking these 'simple' WFS &lt;br&gt;implementations can only respond to some 'dumb' queries - they don't support &lt;br&gt;Filter Encoding spec - but even this level of functionality provides far...</text>
36</hook>
37</node>
38<node TEXT="reference" POSITION="left">
39<node TEXT="Coverages" POSITION="left">
40<hook NAME="accessories/plugins/NodeNote.properties">
41<text>"ISO 19123 Geographic information - Schema for coverage geometry and functions" &lt;br&gt;provides information models to describe a variety of coverages.&lt;br&gt;&lt;br&gt;[Noting that domain specifications such as 'irregular girds' are not &lt;br&gt;described. The implication is that the earth-science community will need to &lt;br&gt;provide extensions in order to adequately describe their information; refer to &lt;br&gt;the CSML2 efforts from NERC Data Grid]&lt;br&gt;&lt;br&gt;Previously, I had thought that one needed to derive new feature types; either &lt;br&gt;specialising coverages, or composing a new feature type using a standard &lt;br&gt;coverage.&lt;br&gt;&lt;br&gt;However, on further analysis, it appears that the standard coverage is...</text>
42</hook>
43</node>
44<node TEXT="Observations &amp; Measurements" POSITION="left">
45<hook NAME="accessories/plugins/NodeNote.properties">
46<text>'Observations and Measurements' (O&amp;M) is another model that provides a generic &lt;br&gt;pattern.&lt;br&gt;&lt;br&gt;From ISO TC/211, we know that a feature-type is characterised by a specific &lt;br&gt;set of properties. These properties will either be asserted (name, owner, &lt;br&gt;proce, boundary) or estimated (colour, mass, shape). The significance of an &lt;br&gt;estimated property is that the error in the property value is of interest.&lt;br&gt;&lt;br&gt;O&amp;M provides an information that enables one to describe features with &lt;br&gt;estimated properties, focusing on the data consumer perspective, rather than a &lt;br&gt;data provider perspective (i.e. we care about the information more than the &lt;br&gt;sensor that was used to measure the information)....</text>
47</hook>
48</node>
49<node TEXT="GML Simple Features Profile" POSITION="left">
50<hook NAME="accessories/plugins/NodeNote.properties">
51<text>The GML specification is complex! It can encode dynamic features, spatial and &lt;br&gt;temporal topology, complex geometric property types and coverages. With such a &lt;br&gt;wide scope, interoperability can only be achieved by defining profiles of GML &lt;br&gt;that deal with a restricted subset of GML capabilities.&lt;br&gt;&lt;br&gt;The so-called 'simple features profile' is a mechanism to achieve this.&lt;br&gt;&lt;br&gt;In many cases, the complexity of GML is simply not required - the information &lt;br&gt;encoded within a GML instance 'document' does not draw on the additional &lt;br&gt;elements added to the GML specification for version 3. In fact, one could &lt;br&gt;suggest that the 'simple features profile' is a mechanism to restrict the GML &lt;br&gt;specification back to the scope of GML 2; suiting the needs of most...</text>
52</hook>
53</node>
54<node TEXT="Geoserver Roadmap" POSITION="left">
55<hook NAME="accessories/plugins/NodeNote.properties">
56<text>A roadmap has been defined for Geoserver (open source WMS / WFS / WCS &lt;br&gt;implementation - http://docs.codehaus.org/display/GEOS/Home) that expresses &lt;br&gt;how the current implementation needs to evolve to match the requirements of &lt;br&gt;the information modelling community (see link)&lt;br&gt;&lt;br&gt;As can be seen, there are a number of technical issues that need to be &lt;br&gt;resolved before operational use of the ISO 19123 coverage models can be used.&lt;br&gt;&lt;br&gt;Rationale: a large number of data standards will emerge in the near future &lt;br&gt;based in the ISO19139 metadata patterns. All major jurisdictions are &lt;br&gt;developing data standards based on ISO models, that will introduce this &lt;br&gt;standard as "Feature level metadata"....</text>
57</hook>
58</node>
59</node>
60</node>
61</node>
62<node TEXT="Observations &amp; Measures [Cox]" POSITION="left">
63<hook NAME="accessories/plugins/NodeNote.properties">
64<text>The observed property is associated with something ... the feature of interest&lt;br&gt;&lt;br&gt;The proximate feature of interest may merely sample the more meaningful domain &lt;br&gt;feature ... example: rock sample - ore body, sounding - atmosphere, water &lt;br&gt;quality samples - river etc.&lt;br&gt;&lt;br&gt;When is O&amp;M interesting? if the data-acquisition metadata is of concern&lt;br&gt;&lt;br&gt;SensorML - "aimed high but achieved low" ... got tangled up in syntax.&lt;br&gt;&lt;br&gt;Observation model is general; but it is customised (specialised) for each &lt;br&gt;domain of application.&lt;br&gt;&lt;br&gt;SamplingFeature is characterised by the thing it is intended to sample [the &lt;br&gt;feature of interest]; sampling features are the proximate feature of...</text>
65</hook>
66</node>
67</node>
68<node TEXT="Day 2" POSITION="left">
69<node TEXT="Attempts to harmonise scientific feature types" POSITION="left">
70<hook NAME="accessories/plugins/NodeNote.properties">
71<text>O&amp;M-Samping | CDM | CSML | NcML&lt;br&gt;&lt;br&gt;+++++&lt;br&gt;&lt;br&gt;Note that O&amp;M::SamplingPoint does not have an equivelent of &lt;br&gt;O&amp;M::SamplingCurve.propertyLog because &lt;br&gt;{abstract} O&amp;M::SamplingFeature.property is adequate ... the SamplingPoint &lt;br&gt;does not have a coverage!&lt;br&gt;&lt;br&gt;[suggestion: rename SamplingCurve.propertyLog to &lt;br&gt;SamplingCurve.coverageProperty ... &amp; same for SamplingSurface or SamplingSolid &lt;br&gt;... or could remove them entirely as the &lt;br&gt;{abstract} O&amp;M::SamplingFeature.property can accomodate a coverage ......</text>
72</hook>
73</node>
74</node>
75<node TEXT="Day 3" POSITION="left">
76<node TEXT="Implementation" POSITION="left">
77<hook NAME="accessories/plugins/NodeNote.properties">
78<text>[Woolf] CSML implementation review&lt;br&gt;&lt;br&gt;xlink ... intended to be one of the core xml specifications; it has two &lt;br&gt;profiles:&lt;br&gt;&lt;br&gt;simple xlink ... only 2 objects supported; local resource, remote resource&lt;br&gt;&lt;br&gt;extended xlink ... ability to link an arbitrary number of objects&lt;br&gt;&lt;br&gt;[Tandy] surprised that remoteResource.role is referenceable resource [url] ... &lt;br&gt;urn seems more appropriate&lt;br&gt;&lt;br&gt;CSML 'could' use extended xlinks to link to together multiple physical files; &lt;br&gt;but instead choose to use a 'CSML storage descriptor' to carry the aggregation &lt;br&gt;semantics &amp; use a simple xlink to reference that...</text>
79</hook>
80</node>
81<node TEXT="Roadmap" POSITION="left">
82<hook NAME="accessories/plugins/NodeNote.properties">
83<text>Write stuff up (from this workshop)&lt;br&gt;&lt;br&gt;Expand the operations / services / affordances required ...&lt;br&gt;&lt;br&gt;when the CDM scientific features have been implemented the CSML storage &lt;br&gt;descriptors can disappear ...</text>
84</hook>
85<node TEXT="Data" POSITION="left">
86<node TEXT="NetCDF" POSITION="left">
87<node TEXT="CF" POSITION="left"/>
88<node TEXT="CF2" POSITION="left"/>
89</node>
90</node>
91<node TEXT="Virtualisation" POSITION="left">
92<hook NAME="accessories/plugins/NodeNote.properties">
93<text>aka. aggregation or 'logical view'</text>
94</hook>
95<node TEXT="CSML 'storage descriptors'" POSITION="left"/>
96<node TEXT="NcML" POSITION="left"/>
97</node>
98<node TEXT="root" POSITION="left">
99<node TEXT="Coverage" POSITION="left">
100<node TEXT="NcML-G" POSITION="left"/>
101<node TEXT="NcML-G evo" POSITION="left"/>
102</node>
103<node TEXT="... blurred ..." POSITION="left">
104<node TEXT="SamplingFeature" POSITION="left"/>
105<node TEXT="community SamplingFeature" POSITION="left"/>
106<node TEXT="CDM 'scientific features'" POSITION="left">
107<hook NAME="accessories/plugins/NodeNote.properties">
108<text>6-9 months</text>
109</hook>
110</node>
111<node TEXT="CSML v2" POSITION="left"/>
112<node TEXT="CSML v3" POSITION="left">
113<hook NAME="accessories/plugins/NodeNote.properties">
114<text>3-6 months away</text>
115</hook>
116</node>
117<node TEXT="WCS 1.1" POSITION="left">
118<node TEXT="Galeon2" POSITION="left">
119<hook NAME="accessories/plugins/NodeNote.properties">
120<text>use Galeon2 as a test-bed for 'gridded' SamplingFeature</text>
121</hook>
122</node>
123</node>
124</node>
125<node TEXT="Feature" POSITION="left">
126<node TEXT="WFS 1.1" POSITION="left"/>
127<node TEXT="WFS 1.2" POSITION="left">
128<node TEXT="Data-Access-Query-Model" POSITION="left" COLOR="#3366ff"/>
129</node>
130</node>
131</node>
132<node TEXT="Clients" POSITION="left">
133<node TEXT="NDG" POSITION="left"/>
134<node TEXT="GI-GO" POSITION="left"/>
135</node>
136</node>
137</node>
138</node>
139</map>