incorrect merge of title prefix in search engine

[solrize]

http://openlibrary.org/search?q=thebest finds a lot of wrongly indexed titles, e.g. "thebest of leigh brackett", because the title_prefix "the" has gotten improperly concatenated with "best". There are similar errors for other prefixes. I thought we were not supposed to insert a space because of prefixes like L-apostrophe. Is there a precise rule we are supposed to use?

[solrize (solrize)]

It looks like we have an explicit trailing blank in title prefixes like "The ". That is easy to get wrong. I think inserting a space should be the default and we should have a way to suppress it, perhaps an alternate prefix field. Are there any examples other than L-apostrophe (like "L'Hotel du Nord") that shouldn't get a space? Maybe I should insert the space for every prefix except that one.

[Edward Betts (edwardbetts)]

Please read https://bugs.launchpad.net/openlibrary/+bug/262265 it contains other examples.

Maybe we should merge these bugs.

[solrize (solrize)]

Hmm, Karen writes:

a list of title prefixes (with language, but unfortunately not using the
ISO language code!)

http://www.loc.gov/marc/bibliographic/bdapndxf.html

Should we just strip out those prefixes automatically?