Random Post: get_results("SELECT ID,post_title,guid FROM $wpdb->posts WHERE post_status= \"publish\" ORDER BY RAND() LIMIT 1"); $p=$post[0]; echo ('' . $p->post_title . ''); ?>
RSS .92| RSS 2.0| ATOM 0.3
  • Home
  • About

    At risk over the weekend of 19th-20th Feb

    February 11th, 2011

    There may be some disruptions to our service over the weekend of 19th-20th Feb 2011. We’re replicating our servers across two datacentres at the University of Edinburgh, and one of them is being switched off over that weekend.

    Hopefully the running of Unlock won’t be affected – we survived a recent “at risk” weekend without downtime – but just so you know.

    Unlock service status and plans

    July 20th, 2010

    Note, this post was originally written in May 2010

    This is an attempt to set out the status of the Unlock services and describe the future roadmap, written in response to much missed EDINA colleague David Medyckyj-Scott who’s now shaking up the New Zealand scene at Landcare Research.

    Right now Unlock provides a gazetteer search service and a placename text-mining service. Unlock Places searches across different sources of data for references to placenames, returning “footprints” where possible and points (e.g. latitude,longitude coordinates) where not.

    Unlock Text uses the gazetteer to locate places extracted from text documents and XML metadata using natural-language processing tools provided by our colleagues at the Language Technology Group, School of Informatics, University of Edinburgh.

    Now we have good open data sources of footprints from the Ordnance Survey. This helps to justify implementing those parts of the Unlock Places API that deal with searching using shapes. For example searching for a name or a type of place within a detailed shape, or drawing a buffer around a shape and searching within that (e.g. “Within five miles of the boundary of the City of Edinburgh”).

    Before too long we plan to add the ability to search through OpenStreetmap data and run an equivalent to OSM’s Nominatim service so researchers can do large volume batch geocoding.

    We plan to create an open source release of the gazetteer and geoparser, beginning summer 2010 if possible.

    Longer term plans:

    • Add more historic placename and footprint data. Leading bid to extract placename records from authoritative source for England and Wales.
    • Separate the *geotagger* from the *georesolver* and re-implementing the latter using shapes rather than points where possible.
    • Add a temporal reference parsing service which works similarly to the geoparser but for temporal event references – “Unlock Time”.
    • Separate out the personal name parsing part of the geoparser into a distinct service

    Use cases:

    • Cross-search between different coordinate reference systems – for example, you know a postcode and want to search by latitude, longitude. The Archaeology Data Service uses Unlock Geocodes for this.
    • Use KML output from the gazetteer to project local area statistics onto Google Earth.
    • “GeoPrints” plugin to EPrints extracts locations from documents uploaded to institutional repository
    • Geotagging large reference collections of documents, such as the proceedings of the parliament of Northern Ireland

    "At risk" 7-8am March 23rd, 8-9am March 30th

    March 9th, 2010

    The Unlock services will have a couple of one-hour periods at risk of downtime for network maintenance at the end of this month.

    7-8am Tue Mar 23rd a router upgrade planned on the University of Edinburgh network may disrupt access to the Unlock service.

    8-9am Tue 30th March is an “at risk” period for our connection to JANET while it undergoes resilience testing.

    That’s all.