- From: Dan Brickley <danbri@w3.org>
- Date: Sat, 5 Apr 2003 17:18:24 -0500
- To: www-annotation@w3.org, www-rdf-calendar@w3.org
- Cc: djweitzner@w3.org
(crossposting the rdf-calendar and annotation lists) I've done a bit more towards writing up the opening hours meets restaurant review SemWeb use case. What writeup I have is in the ESW Wiki, http://esw.w3.org/topic/RestaurantRecommendation http://esw.w3.org/topic/OpeningHoursUseCase and nearby, http://esw.w3.org/topic/GeoInfo Each of these Wiki pages points to (perhaps sketchy) RDF/XML sample markup. The OpeningHoursUseCase example was discussed on www-rdf-calendar previously. For GeoInfo we show the use of a lat/long/alt markup of Web pages, and for the RestaurantRecommendation scenario I've excerpted some RDF/XML examples from the Open Directory's "Chef Moz" project. The Chef Moz example is interesting. There are various ways in which their data and markup could be improved, but it does show a fair sized body of restaurant review content available in (slightly buggy) RDF/XML. In particular, RDF cal folks might be interested to see that opening hours are marked up as follows: <Hours>daily 6.30pm-11pm</Hours> <ParsedHours>18.5-23|18.5-23|18.5-23|18.5-23|18.5-23|18.5-23|18.5-23</ParsedHours> ...and that timezone seems to be left implicit, perhaps discoverable via <Country>United Kingdom</Country> <Location>United_Kingdom/England/Bristol</Location> or <Zip>BS6 7AH</Zip> I encourage anyone from www-rdf-calendar or www-annotation with an interest in such apps to take a look at the sketchy docs in the Wiki and add pointers or suggestions directly to the pages. I'm particularly interested in finding more freely accessible data or lookup services that could be used for a realistic demo. You can edit the pages using the "EditText" link at the foot of the page. RDF cal folk, any suggestions on what we might do to convert the opening hours information into RDF? Particularly the timezone issue? Help/suggestions welcome. On the Web annotations front, I find this interesting to contrast with Annotea in that it is a scenario where the thing one is really annotating (ie. describing) isn't a Web page (eg. restaurant homepage) but the thing that the Web page is about, ie. the restaurant. We could, however, use chefmoz-like data to auto-generate Annotea annotations attached to the restaurant homepages, for those restaurants which chefmoz knows a homepage URL for. There are about 13,000 URLs for restaurants in their dataset currently, which should be taken in context (stats from chefmoz.org homepage): 185116 restaurants; 12459 reviews; 47955 links to reviews; 595 editors. Big but not vast. Perhaps someone would find it interesting to generate an Annotea service from parsing the Chef Moz restaurant reviews and turning them into Web annotations on the homepages of those restaurants that have them? The next thing I want to do is get the calendar aspects translated into RDF cal, and associate some additional metadata (lat/long/alt), photos etc., with some sample records. Thanks for any suggestions, links, help with the writeups etc., cheers, Dan
Received on Saturday, 5 April 2003 17:18:26 UTC