Question

The Apache Solr wiki states that mlt.minwl =
minimum word length below which words will be ignored.

A concrete example of this is appreciated.

Example query (decoded)

(string:822) qt=mlt&fl=nid%2Ctitle%2Cpath%2Curl%2Css_simple_geo_position%2Cis_cck_field_sponsored_content_yn%2Cis_cck_field_compound_review_yn%2Cis_workflow_state%2Cds_cck_field_publish_date%2Cds_cck_field_publish_expiration_date%2Csm_timeout_search_event_type%2Csm_timeout_search_event_genre%2Csm_timeout_search_venue_type%2Csm_timeout_search_venue_feature%2Csm_timeout_search_venue_genre&mlt.fl=body%2Cname%2Ctaxonomy_names%2Ctitle&mlt.mintf=1&mlt.mindf=1&mlt.minwl=3&mlt.maxwl=15&mlt.maxqt=20&fq%5B0%5D=-is_cck_field_exclude_from_search%3A1&fq%5B1%5D=-type%3Aimage&fq%5B2%5D=-type%3Aoccurrence&fq%5B3%5D=%28nodeaccess_f98a254002c9_all%3A0+OR+nodeaccess_f98a254002c9_workflow_access%3A1+OR+nodeaccess_f98a254002c9_workflow_access_owner%3A0+OR+nodeaccess_all%3A0%29&facet.limit=21&version=1.2&wt=json&json.nl=map&q=milk&start=0&rows=4

Answer 1

If an “interesting term” (alike), found by MLT, has less than mlt.minwl number of characters it will be ignored (excluded from MLT results).
Default value of this param is 0, which means that the param has no effect.

The essences of the internal workings
of MLT operate like this:

1. Gather all of the terms with frequency information from the input
   document:
2. If the input document is a reference to a document within the
   index, then loop over the fields
   listed in mlt.fl, and then the term
   information needed is readily there
   for the taking, if the field has the
   termVectors enabled. Otherwise get the
   stored text, and re-analyze it to
   derive the terms. If the input
   document is posted as text to the
   handler, then analyze it to derive the
   terms. The analysis used is that
   configured for the first field listed
   in mlt.fl.
3. Filter the “interesting terms” based on configured thresholds, one of which is your mlt.minwl param.
4. Construct a query with these interesting terms across all of the
   fields listed in mlt.fl.