Help to remove imported libraries and cleanup automated rdf mappings

This will revert all your mappings to defaults and remove additional namespaces:
- Uninstall evoc
- empty/TRUNCATE the tables rdf_mapping and rdfx_namespaces
- clear your caches (admin/config/development/performance or run drush cc all)

if all you want is to support schema.org terms, you don't need to enable evoc or RDFx/RDF UI.

re the 'Warning: RDFa tree has unknown type' displayed on the rich snippet testing tool, it's not a bug in your markup, it's just Google warning you that some terms were not recognized (because Google only support a small subset of the terms defined on the Web). As long as you can see your type and the right properties in the output of the testing tool, you can ignore this warning.

I never imagined that the modules would not give option to remove those vocabularies and they automated mappings.

yes, it's a missing feature, let's work on this at #1296820: Allow for imported vocabularies and namespaces removal.

All of the things that you list are because of the core RDF module. The rdf_mapping table is installed by core. The mappings you list in #2 are because of the bug that I pointed you to in one of the other threads. Most of the namespace mappings at the top of the document are defined in core. The only one that isn't is the schema.org namespace at the bottom of the list. This is defined in the Schema.org module.

If you do not want #2 and #3, your only option ATM is to turn off the core RDF module. The fix for #2 probably won't come for a while and I doubt #3 will change during the 7.x development cycle. Of course, this means that you wouldn't be publishing Schema.org terms in RDFa.

Thanks for your support.

1.Yes I use a custom theme.

2. I though that they would create problem to google.
For sites to appear with rich snippets, webmaster must email google (they have a specific form for that) and then google will decide if it will or not include rich snippets in search results for the particular site.
I thought that the cleaner the code the more possible it is to get included by google.

3. "HTML+RDFa 1.1": should it be replaced with something better?

4. @linclark : are you referring to http://drupal.org/node/1229974 ?
"The fix for #2 probably won't come for a while ".
Do you mean it's a matter of time because of your limited time, or it depends on something else?

As I understood it is only linclark and scor who are developing semantic-web modules.
How we can attract more developers, or persuade acquia to deal with this important matter? Semantic web is the most important evolution that happens in web after web itself and if Drupal can be pioneer then it can attract bigger usage share.

5. If it is core rdf module behavior, then I have to compromise. I thought that all the headers as described in #3, where results of the imported vocabularies.
My header markup is huge because my theme repeats the same for 5 times just to add a class depending on the IE version. This is because of some scripts that try to emulate html5 and css for older IE versions.

6. So the only other way to implement the results that I want, is to disable rdf modules and schema.org and create custom tpl.php for that content type, I suppose, right?

3. The RDFa output by Drupal 7 core is actually RDFa 1.0. Drupal 8 will use RDFa 1.1. You will want to change this to 1.0
4. It is a matter of limited time, and thank you for your recognition of those limitations. Unfortunately, it's hard to find committed volunteers for many of the different parts of Drupal development and documentation, which is why there is a recent focus among core developers on strategies for "on-boarding" new contributors.

   I agree that this new phase of the Web is really just emerging and that Drupal is in a really good leadership position. Even folks in the W3C and WHATWG have commented on how future-foward Drupal is with these technologies.

   I think that the best way to bring more volunteers on is to find very immediate and very simple uses for the technology and then to share them, either through screencasts and other documentation or through downloadable configurations. By spurring use of the technology to solve simple problems, it will increase the number of users playing with the technology, which will increase the understanding and the capacity for testing and development that we need.

5. As a non-optimal solution, you could remove the namespaces that you don't use in RDF module. This would be 'hacking core', which isn't the best thing to do, but if you are aware of the ramifications then it isn't so bad. You would go into rdf.module and remove unwanted namespaces from rdf_rdf_namespaces. You should save this as a patch for yourself so that when you update Drupal, you can just apply the patch again.
6. Depending on the fields you use, you could also consider microdata module. It is still in development and I tell people it isn't ready for production sites because it is still changing a bit and most fields do not support it yet, but it works for certain fields. Be aware that until the module gets a stable release, you may need to update your mappings when you update the module.

Re: comment 7

Scor, the HTML element inside IE conditional comments sounds like the approach taken by the HTML5 Boilerplate. Perhaps dropbydrop's theme uses a similar approach. See Conditional stylesheets vs CSS hacks? Answer: Neither! for an explanation of how this works. It means that you get 5 html opening tags. Not pretty IMHO but a reasonably effective way to provide a browser class.

To use this approach in Drupal however, you need to put $rdf_namespaces> (D7) or $html_attributes (D8) inside all of the html tags (i.e. inside each conditional comment), so all the namespace declarations are duplicated as well. Around 80 lines of HTML source code before you even reach the head tag!