Stemmed searches by default off in ML9

[grtjn]

ML9 seems to have changed the default for stemmed-searches to off. Does it make sense to enable this for content-db by default in slush, like we enable uri lexicon, collection lexicon?

[dmcassel]

Based on internal conversations, we've confirmed that the new default in MarkLogic 9 is stem=off, word=on. (Motivation relates to more customers using MarkLogic for data rather than text.) In light of this, the following changes make sense:

• let the change in defaults flow through by not setting the explicitly in ml-config.sample.xml
• update default-docs-config.sample.xml with the MarkLogic 9 defaults; include a comment that explains the change
• add commented-out settings to ml-config.sample.xml so that it will be easy to change if desired; I suggest these be the ML9 values, but add comments noting the change and listing the old values