NDG PB and AH July 2006 * CSML * not developed in isolation. * conformance to emerging best practice * DECISIONS * all new engineering on csml2.0 should be finished by pre-beta * new developments * processing affordance * GML 3.2 * swath, profileCollection * storage descriptor moving to own namespace * composite domain pattern (remove WHY WHY WHY) * respect standards and autogeneration tools * what is CSML * Core set of base types * trivially derived FTs * schemas * tooling * parser * scanner * cf-compliant netcdf specific * api: basic services * Care in evolution * bodc should be able to do it from sql to python code * nocs may need to transform csml1 to csml2 to avoid rereading entire archive * Need to consider the relationship of delivery service CSML and content at all data providers * How do we plan to deliver qxf files to the consumer of csml files * qxf read methods * MOLES * Decisions * MOLES is an INTERNAL NDG format * Main reason for existence * provide common interface for producing D * content for exporting DIF, ISO19139 * Provides tool for NAVIGATION * Moles should NOT be a profile of ISO19139 * Internal information storage should not be driven by external standards * Needs to be easier to modify, and extend by adding external schema * Needs the "coathanger" functionality. * three types * import content as is * link to content * import content rendered by this service * Major Priorities * Sort out Granules * cross over with CSML * Sort out SpatialTemporal Descriptoin * are bounding boxes sorted * Schema for Stub-B * no explicit catering for big deployments etc * Add Coathanger Functionality * specifiy mechanism * includes what is being hung off * to which nodes * Add ISO19139 export * which profiles * do we have the content? * need code * UML model of existing MOLES * Documentation, Documentation, Documentaiton * xlink pattern issue * transport mechanism * uri * file type * insert or load * role in moles and csml * server side subsetting, how to put in the uri * xlink * roile * arcrole * Services * Discovery * ISO191309 import * browse/discovery view * import of spatio/temporal into backend complex database * Role of OpenSearch * Role of OGC Catalogue Services * Greg Read having an IOC WMO profile meeting in September in Phillidelphia * Z39.50 gateway * Security * evolving to more standards compliance * Java coding * AHM * One free person (registration paid for) * Michael and Helen * Mike * on holiday? * Meetings Needed * Need a bbftp, delivery service, csml discussion, then talk about dx/wXs * Code Review meeting * BInary Image Format Meeting * Issues to trac * Who owns BBFTP * vocab.ndg.nerc.ac.uk * Browse * history should include D * xml rendering/printing clarity needed * WG issue: how to manage role mapping as things change in data providers * can we use the vocab server for role publication * can the aa check whether they have changed * How can we limit discovery searches to specific sites ... * Should Moles creation tools be externally facing? * What advice would one starting from scratch? * database configuratoins * dataset/datagranule decisions, reasons, and limitations * clear guidance needed for use of csml docs (e.g. can cope with 10,000 files) * Automatic population of OAI repository from updated MOLES ... * PML do that ... * Still need to resolve OAI respository quality issues. * CSML API read methods need coordiation (ie. dom isn't doing the work, but he needs to implement them); relationship with CF-API work. * testrig via equivalents to ncdump and ncview? Exploiting CDAT * Phenomenon needs to be fully described (cell methods problem) * Expecation: where a standard name exists, exploit parameter mapping in vocab server to provide discovery vocabularly. Where a standardd name doesn't exist, we need to face the fact that moles records don't have parameter discovery ... records ... * csml api needs to be able to point at a file or an exist db csml file * what do we do about duplicates in our oai harvested repository * namespace problems in xqueries in exist and csml etc * we have one oai repository * should have two exist databases * one live one development * both populated directly * harvest now button * link checker for harvested material * could send an email if some exists to do so * flag broken links in display * parse access constraints a bit more carefully * Better branding for data providers (icons in browse etc) * How to show html in Trac ... [[SimpleFreeMind(NDG-AllHands200607.mm)]]