From: Joel N. <joe...@gm...> - 2013-08-07 17:31:19
|
Hi Yaron, Thanks for the hint. And revisiting "caching" Semantic Drilldown, the cached drilldown file perhaps, can even be used with MigaDV, as an offline browser of an SMW site. Best, Joel ======================================================= Think Different! (http://en.wikipedia.org/wiki/Think_different#Text) Imagine Different! (http://www.youtube.com/watch?v=H5tOgRD4EqY) On Wed, Aug 7, 2013 at 12:36 PM, Yaron Koren <ya...@wi...> wrote: > Hi, > > Thanks for the feedback. > > Joel - you can already exclude categories from the drilldown, using > __HIDEFROMDRILLDOWN__. > > -Yaron > > > On Wed, Aug 7, 2013 at 12:13 PM, Joel Natividad <joe...@gm...>wrote: > >> +1 for smart filtering, but still give admins the ability to optionally >> tweak the default behavior on a category by category basis and perhaps, >> exclude the category from semantic drilldown. >> >> While you're thinking about SemanticDrilldown, you may also want to think >> about how to make it more "performant" for large sites, maybe even explore >> the concept of doing "cached" drilldowns that are precomputed on a schedule >> (think generateSitemap<http://www.mediawiki.org/wiki/Manual:GenerateSitemap.php> >> ). >> >> Best, >> Joel >> >> ======================================================= >> Think Different! (http://en.wikipedia.org/wiki/Think_different#Text) >> Imagine Different! (http://www.youtube.com/watch?v=H5tOgRD4EqY) >> >> >> On Wed, Aug 7, 2013 at 10:37 AM, Yaron Koren <ya...@wi...> wrote: >> >>> Hi everyone, >>> >>> You may have seen the project I released last month, the Miga Data >>> Viewer: >>> >>> http://migadv.com >>> >>> In brief, it's a Javascript-based application that provides a browsing >>> interface that's somewhat similar to Semantic Drilldown, but more >>> generic, >>> and it can handle any structured data, not just data from Semantic >>> MediaWiki. (There are a few other differences, like that the interface >>> can >>> work offline.) >>> >>> Miga is based around the "Data is data" theory, described here: >>> >>> http://wikiworks.com/blog/2013/02/06/data-is-data/ >>> >>> In brief, it's the idea that the best way to display, and browse, a set >>> of >>> data can be determined solely by the size, structure and layout of the >>> data, not its subject matter. >>> >>> I bring all this up because creating Miga has made me reconsider some of >>> the design decisions behind Semantic Drilldown. Miga takes a simpler, >>> more >>> automated approach to browsing: when a user filters on a number or date >>> property/field, Miga generates the set of ranges for that property/field >>> automatically, based on the current layout of the data. With Semantic >>> Drilldown, conversely, the admin specifies the ranges ahead of time: for >>> Number properties, the admin has to specify ranges like "0-9", "10-99", >>> etc., while for Date properties they have to specify whether the values >>> should be subdivided by month or year (those are the only two options), >>> or >>> whether the input should be a date range, with a "start" and "end" input >>> for users. Besides being more work for admins, this approach results in a >>> more inflexible system. >>> >>> To take an example: a wiki has a category holding buildings, including >>> properties for the date they were completed, and their height. Both of >>> these have filters in SD, and the admin picks a reasonable setting for >>> the >>> completion date (filtering with a date range, since the time span is in >>> hundreds of years) and a reasonable set of number ranges for the height. >>> But then a user goes to Special:BrowseData and filters down by city, and >>> selects a city that has only relatively short buildings, that were all >>> built in the last 20 years. For the "Completion date" property, a simple >>> subdivision by year would have been the better choice at this point, and >>> for the "Height" property, all the values will wind up in the lowest >>> number >>> range, which makes that filter rather useless. >>> >>> And actually, perhaps having a "Date range" input is never an ideal >>> option >>> anyway - perhaps date properties should simply subdivide by whatever the >>> best unit is for the current set of values (years, decades, centuries; or >>> months or days), and then let the user keep subdividing after they >>> select a >>> value. That's how Miga does it, and it seems to work well. >>> >>> A similar change would be, for Text/String properties, to have the input >>> type depend on the number of values: a "combo box" input type if the >>> number >>> of values is greater than some large number, like 50, vs. simply listing >>> the set of values if it's smaller than that number. This would differ >>> from >>> the current situation, where again the admin has to specify ahead of time >>> whether they want a combo box or not. >>> >>> My strong inclination is to switch to this approach, which would result >>> in >>> the removal of the special properties "Has value", "Uses time period" and >>> "Has input type". But this would take power away from admins, so I wanted >>> to ask before looking into this any further. Would anyone object to this >>> sort of major change to Semantic Drilldown? Are there filter-value setups >>> that it makes sense to keep, regardless of the current layout of the >>> data? >>> >>> -Yaron >>> >>> -- >>> WikiWorks · MediaWiki Consulting · http://wikiworks.com >>> >>> ------------------------------------------------------------------------------ >>> Get 100% visibility into Java/.NET code with AppDynamics Lite! >>> It's a free troubleshooting tool designed for production. >>> Get down to code-level detail for bottlenecks, with <2% overhead. >>> Download for free and get started troubleshooting in minutes. >>> >>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Semediawiki-user mailing list >>> Sem...@li... >>> https://lists.sourceforge.net/lists/listinfo/semediawiki-user >>> >> >> > > > -- > WikiWorks · MediaWiki Consulting · http://wikiworks.com > |