I recent posted this and received an appropriate response from Bradley:
- Open LUT 84
- Select word in the German that is also a word in English (I picked "sage")
- Select search all resources (green arrow)
- Note that the search is built simply for the sequence of letters without the language indicator from LUT 84, I would suspect this also causes the wrong stemming results further corrupting the results. (red arrow)
- Note that the results include both German and English occurrences (red circle 1) i.e. a meaningless mic.
- (blue arrow) the search that I expected - limited to German because the word I selected was in German.
(blue arrow) the search that I expected - limited to German because the word I selected was in German.
As you noted, you can override the default behaviour to only find results in one particular language.
I would suspect this also causes the wrong stemming results further corrupting the results.
It doesn't.
This got me thinking about 4 separate Search issues:
- the building of a datatype entry when most users (erroneously) expect Milestones
- the building of a generic search when a user could reasonably expect it to be limited by language; a similar problem may exist re: the treatment of diacritical marks.
- the number of forum questions that are the result of the 3 Search options "hidden" in the panel menu not having the expected values.
- the expectation that the user knows that the default behavior is not what they need and know the syntax to change it.
I would like additional options and better visibility of the options, these include:
- Match case (exists)
- Match all word forms (exists)
- Match language (new)
- Match diacritics (new)
- Match equivalent references (exists)
- Match occurrences (datatype --> milestones)/Match references (milestone --> references)
I would suggest that these options be a dropdown menu on the main search specification section of the Search screen.
