From: Sergey C. <ser...@gm...> - 2009-06-03 22:36:06
|
Guys, As patch for imported vocabularies from Li Ding worked fine and RDF export now functions properly, I finally published semantic sitemaps for my projects - you can and example here: http://www.techpresentations.org/semantic-sitemap.xml What it does - it allow Semantic Crawlers like Sindice to use static dump instead of crawling your wiki's Special:ExportRDF pages. See more info on Semantic Sitemaps here: http://sw.deri.org/2007/07/sitemapextension/ Thank you, Sergey -- Sergey Chernyshev http://www.sergeychernyshev.com/ |
From: Sergey C. <ser...@gm...> - 2009-06-09 15:07:46
|
I think Sindice is using it and Richard Cyganiak is the guy who advertised it ;) I saw Sindice bot coming for it and all other regular bots too (with no result probably ;)). BTW, there are some issues with command line tools (e.g. SMW_dumpRDF.php) and enableSemantics() which I think is worth converting to just be called at the end of SMW_Settings.php - the problem is with $wgServer and $wgServerName not being set at that point while default value for $smwgNamespace should be based on $wgServerName, but should be bound later in the game when it's actually used. I hope to really get rid of this call with parameter - there is no need for the custom value unless there is something magical happens on the server. BTW, maybe it's worth describing mod_rewrite rules neede for resolving custom namespace values. Sergey On Tue, Jun 9, 2009 at 10:54 AM, Markus Krötzsch < ma...@se...> wrote: > On Donnerstag, 4. Juni 2009, Sergey Chernyshev wrote: > > Guys, > > > > As patch for imported vocabularies from Li Ding worked fine and RDF > export > > now functions properly, I finally published semantic sitemaps for my > > projects - you can and example here: > > http://www.techpresentations.org/semantic-sitemap.xml > > > > What it does - it allow Semantic Crawlers like Sindice to use static dump > > instead of crawling your wiki's Special:ExportRDF pages. > > > > See more info on Semantic Sitemaps here: > > http://sw.deri.org/2007/07/sitemapextension/ > > Great, thanks. I had been discussing this with the guys from DERI some time > ago, but nobody was found to actually implement it. > > -- Markus > > > > > Thank you, > > > > Sergey > > > > > > -- > > Sergey Chernyshev > > http://www.sergeychernyshev.com/ > > > -- > Markus Krötzsch > Semantic MediaWiki http://semantic-mediawiki.org > http://korrekt.org ma...@se... > > |
From: Markus K. <ma...@se...> - 2009-06-09 17:17:52
|
On Dienstag, 9. Juni 2009, Sergey Chernyshev wrote: > I think Sindice is using it and Richard Cyganiak is the guy who advertised > it ;) > I saw Sindice bot coming for it and all other regular bots too (with no > result probably ;)). > > BTW, there are some issues with command line tools (e.g. SMW_dumpRDF.php) > and enableSemantics() which I think is worth converting to just be called > at the end of SMW_Settings.php - the problem is with $wgServer and > $wgServerName not being set at that point while default value for > $smwgNamespace should be based on $wgServerName, but should be bound later > in the game when it's actually used. > > I hope to really get rid of this call with parameter - there is no need for > the custom value unless there is something magical happens on the server. > BTW, maybe it's worth describing mod_rewrite rules neede for resolving > custom namespace values. There are various reasons why the base URI was required to be set when installing SMW. I have seen that you have changed this in SVN recently, but I wonder if this is a good idea. There are various problems: (1) The value of $wgServer depends on the caller that accesses the site. It can be some URL with domain name or just the IP address. For many sites, these already allow for a number of values. It also can be localhost if accessing the wiki locally. In effect, the semantic data exported by the wiki uses different URIs depending on how you read it. I think this is potentially confusing and certainly not useful from a Semantic Web point of view where you look for as much data integration as possible. (2) Local script calls, including bulk export, would not have any reasonable value for the URIs, or would they? Maybe some PHP configuration allows you to supply some URL that is used for $wgServer in this case but this is certainly not the usual setup. The way in which $wgServer(Name) is calculated is rather unpredictable, also across MW versions, see [1]. (3) Any RDF store based storage backend, as e.g. our RAP adaptor, will get into serious problems if URIs change all the time or are defined based on localhost. Also, using SPARQL-queries based on such a store requires you to provide an explicit URI -- but what do you choose if the URI changes depending on your way of accessing the site? These were the main reasons why SMW has always forced admins to supply some concrete string that will be used for making stable URIs. The effort seems small for the stability you get in return. Btw.: please try out all code before committing it to SVN. This can be hard for complex features (so many things to try; see the unstubbing bug I just closed for the import ... ;-), and I know that you are careful with the code you write, but some minimal checks should always be done. As far as I can see, the current code cannot actually work as intended (missing global declaration). Markus [1] http://www.mediawiki.org/wiki/Manual:$wgServer > > Sergey > > > On Tue, Jun 9, 2009 at 10:54 AM, Markus Krötzsch < > > ma...@se...> wrote: > > On Donnerstag, 4. Juni 2009, Sergey Chernyshev wrote: > > > Guys, > > > > > > As patch for imported vocabularies from Li Ding worked fine and RDF > > > > export > > > > > now functions properly, I finally published semantic sitemaps for my > > > projects - you can and example here: > > > http://www.techpresentations.org/semantic-sitemap.xml > > > > > > What it does - it allow Semantic Crawlers like Sindice to use static > > > dump instead of crawling your wiki's Special:ExportRDF pages. > > > > > > See more info on Semantic Sitemaps here: > > > http://sw.deri.org/2007/07/sitemapextension/ > > > > Great, thanks. I had been discussing this with the guys from DERI some > > time ago, but nobody was found to actually implement it. > > > > -- Markus > > > > > Thank you, > > > > > > Sergey > > > > > > > > > -- > > > Sergey Chernyshev > > > http://www.sergeychernyshev.com/ > > > > -- > > Markus Krötzsch > > Semantic MediaWiki http://semantic-mediawiki.org > > http://korrekt.org ma...@se... -- Markus Krötzsch Semantic MediaWiki http://semantic-mediawiki.org http://korrekt.org ma...@se... |
From: Markus K. <ma...@se...> - 2009-06-09 15:31:28
|
On Donnerstag, 4. Juni 2009, Sergey Chernyshev wrote: > Guys, > > As patch for imported vocabularies from Li Ding worked fine and RDF export > now functions properly, I finally published semantic sitemaps for my > projects - you can and example here: > http://www.techpresentations.org/semantic-sitemap.xml > > What it does - it allow Semantic Crawlers like Sindice to use static dump > instead of crawling your wiki's Special:ExportRDF pages. > > See more info on Semantic Sitemaps here: > http://sw.deri.org/2007/07/sitemapextension/ Great, thanks. I had been discussing this with the guys from DERI some time ago, but nobody was found to actually implement it. -- Markus > > Thank you, > > Sergey > > > -- > Sergey Chernyshev > http://www.sergeychernyshev.com/ -- Markus Krötzsch Semantic MediaWiki http://semantic-mediawiki.org http://korrekt.org ma...@se... |