Difference between revisions of "Property talk:Occurred on"
Jump to navigation
Jump to search
m (reply) |
m (reply) |
||
| Line 11: | Line 11: | ||
:I'm still very much at the experimental stage too but, as presently defined, the property [[property:is assassination date|is assassination date]] refers only to the date. IOW 22 November 1963 is the date on which an assassination occurred. It does NOT include the event or page as such. The property has the type 'date' and the JFK Assassination page RDF feed only says that it is used on that page. Other properties can be defined for assassination events or specific assassination pages. I'm still feeling my way about the best way to delimit the use of categories and properties but can see considerable advantages for the extended use of semantic properties. You've probably noticed that I've switched on the display of 'semantic facts' at the foot of each page. It is currently set to display only on pages that contain at least one semantic property. Initial potential issues are that a refresh is often need to display them because of caching and an unfinished work queue (SMW batches a lot of its indexing into the MW jobs queue). I am also experimenting with assigning greater resources to working the queue. It is specified as x jobs per page load and uses a lot of memory when there are a lot of pending jobs and the figure is set too high. Worth a lot more attention me thinks. I'm away in Ireland for a wedding for the next few days and won't get much more done 'till next week now. --[[User:Peter|Peter P]] ([[User talk:Peter|talk]]) 16:50, 23 October 2013 (IST) | :I'm still very much at the experimental stage too but, as presently defined, the property [[property:is assassination date|is assassination date]] refers only to the date. IOW 22 November 1963 is the date on which an assassination occurred. It does NOT include the event or page as such. The property has the type 'date' and the JFK Assassination page RDF feed only says that it is used on that page. Other properties can be defined for assassination events or specific assassination pages. I'm still feeling my way about the best way to delimit the use of categories and properties but can see considerable advantages for the extended use of semantic properties. You've probably noticed that I've switched on the display of 'semantic facts' at the foot of each page. It is currently set to display only on pages that contain at least one semantic property. Initial potential issues are that a refresh is often need to display them because of caching and an unfinished work queue (SMW batches a lot of its indexing into the MW jobs queue). I am also experimenting with assigning greater resources to working the queue. It is specified as x jobs per page load and uses a lot of memory when there are a lot of pending jobs and the figure is set too high. Worth a lot more attention me thinks. I'm away in Ireland for a wedding for the next few days and won't get much more done 'till next week now. --[[User:Peter|Peter P]] ([[User talk:Peter|talk]]) 16:50, 23 October 2013 (IST) | ||
| + | |||
| + | :: Ref previous reply. After all that spiel, I see what you mean. I'll change 'is assassination date' to 'occurred on'. Much better. --[[User:Peter|Peter P]] ([[User talk:Peter|talk]]) 16:55, 23 October 2013 (IST) | ||
Revision as of 15:55, 23 October 2013
Semantic web
I am jumping in in the middle of this, have read no documentation, but from my general semantic web experience, it seems overspecialised to state:
(JFK-Assassination has-assassination-date 1963-11-22)
I suggest instead using two predicates:
(JFK-Assassination occurred-on 1963-11-22)
(JFK-Assassination has-event-type assassination)
I am using RDF style semantic triple notation. Basically two simple predicates is likely to be more use and more extensible than one more complex one. This is a general opinion devoid of context on how SMW works, so if this makes no sense, feel free to go ahead and ignore it. Robin (talk) 14:36, 23 October 2013 (IST)
- I'm still very much at the experimental stage too but, as presently defined, the property is assassination date refers only to the date. IOW 22 November 1963 is the date on which an assassination occurred. It does NOT include the event or page as such. The property has the type 'date' and the JFK Assassination page RDF feed only says that it is used on that page. Other properties can be defined for assassination events or specific assassination pages. I'm still feeling my way about the best way to delimit the use of categories and properties but can see considerable advantages for the extended use of semantic properties. You've probably noticed that I've switched on the display of 'semantic facts' at the foot of each page. It is currently set to display only on pages that contain at least one semantic property. Initial potential issues are that a refresh is often need to display them because of caching and an unfinished work queue (SMW batches a lot of its indexing into the MW jobs queue). I am also experimenting with assigning greater resources to working the queue. It is specified as x jobs per page load and uses a lot of memory when there are a lot of pending jobs and the figure is set too high. Worth a lot more attention me thinks. I'm away in Ireland for a wedding for the next few days and won't get much more done 'till next week now. --Peter P (talk) 16:50, 23 October 2013 (IST)
