There are many ways to deal with terminology, here's my personal approach:
In CafeTran I have an Active Glossary (AG.txt), a Background Glossary (BG.txt) and sometimes a Client Glossary (CG.txt).
Client glossaries are fed with Excel (CafeTran opens them directly) and (exported) Multiterm glossaries from my clients. But I seldom receive them (which is fine by my, since I can then use my own preferred terms). These glossaries are also used for CafeTran's terminology consistency check (QA). CG.txt has the highest priority and is case insensitive.
New terms are added on the fly to my Active Glossary (case-sensitive). I regularly dump the content of this glossary into my:
Background Glossary (case-sensitive, read-only), where it is merged (newest target terms at first position) with existing term pairs.
So I'm constantly optimising my BG for my own purposes. Both via periodical merging (e.g. once every month) and on the fly during translation projects (CafeTran stores glossaries in plain text files that I can very quickly edit in a text editor, e.g. to make global modifications when I've learned a better translation or when the spelling has changed).
When I'm translating files that have been translated by colleagues previously, I use CafeTran's <b>term stack feature</b> to have the target term that the colleague used, auto-assembled in the current project (context menu, left-click on the preferred target term to give it the highest priority).
If I don't agree with the target term that my colleague used (e.g. because it has a typo or old spelling), CafeTran's Find and Replace dialogue box will allow me to correct the target term in the TM, the project and the glossary. Modifications to the TM and translation project are made in one run (case is automatically adapted) and the modification of the glossary requires an extra run. All this during the translating process.
Glossary searches can be fuzzy (via Hunspell), but when I really need fuzziness, I use Fragment Memories in TMX format instead of glossaries. CafeTran supports them too, as another (and equivalent) terminology storage format. Here I can use attributes/properties (e.g. client, subject field) too.
Sometimes it's not possible to always have the correct target term auto-assembled, no matter how specific you define the subject field etc. This applies to generic words with multiple (equivalent) translations. In these cases I connect the target terms to related words (e.g. verbs) that can occur in the source sentences. So if verb B appears, translation C will be chosen for source term A. (If B' appears, target term C' will be used instead.)
If the term has a few translations (in your TMs or glossaries). Right-click at its suggested translation in the target editor and choose the one you prefer from the pop-up menu. CafeTran remembers your choice and will use it subsequently in auto-assembling.
Hi HL. This is really cool, thanks. Especially this paragraph:
"When I'm translating files that have been translated by colleagues previously, I use CafeTran's <b>term stack feature</b> to have the target term that the colleague used, auto-assembled in the current project (context menu, left-click on the preferred target term to give it the highest priority)."
One question: what is the "term stack feature" (or the context menu)? I guess these were names of features of an earlier CafeTran version that have been replaced by something else? Thanks.
wery intereeesting Mister Hans. Will ponder this and report back to you.