The priority can be overridden if the accuracy of the match from a resource (be it Glossary or TM) is higher than in another resource. For example, Exact matches override Stem or Prefix matches. The priorities work if the matches are on the same accuracy level. This rule is good and allowing any other options would result in a maze of numerous options.
I can see that. But in the past there was an option in the preferences to let matches from the glossary override identical matches from the TM. As far as I know, this setting has been removed.
There's no way to have it back?
That option has been replaced. Why don't you just make sure that the conflicting Glossary matches are exact? Then, the priorities should work just fine. Otherwise, there would be no logic whatsoever in allowing any fuzziness to overwrite exactness.
Okay, I now understand what you mean. When you allow stemmig for glossaries, a full match from a glossary will have the attribute 'Stemming' and get a lower priority than an exact match from a TM:
The confusing part (at least for me) is that the glossary matches actually aren't derived from stemming, but are EMs too.
Sorry I don't understand this. I have the same issue. I don't think I have "conflicting" Glossary matches.
Please can you provide a clear description how to make sure that my Glossary entries prevail over fuzzy matches?
Try increasing the priority for the given glossary - right-click at the glossary pane and choose "High priority". Also, note that longer fragments have a higher priority in auto-assembling than the shorter ones if the longer matched fragment contains the short term.
ok I've found it ...
I go to the Glossary menu on top of the screen, then to Edit glossary info.
Will see if that helps :)
> But I'm on a Mac.
On Mac, you can also open a context menu for the glossary or TM clicking at the TM/glossary pane and holding CTRL key.
I'm having this priority issue with the combination of a TMX (priority Low) and two glossaries (priorities High and Medium).
I've set the TMX file to Segments memory only.
The issue is that when a word like 'Is' is present both in the TMX (priority Low) and in a glossary (priority Medium), the translation from the TMX is used, instead that of the glossary.
Can this be fixed/can I change settings to prevent this effect?