Start a new topic

Tags give specs

How about this idea: use tags to auto-identify term pairs in pre-translated



I think much more realistic solution + the specially designed interface is via the current Task > Frequent words > Extract frequent words from the current segment.

By default  "Extract frequent words from the current segment" interface adds a phrase to the memory. You can switch for glossaries by choosing  Task > Frequent words > Add frequent words to glossary option.

Yes, but that's a different thing. I was referring to pre-translated files where both source and target terms are surrounded by the same tags. Which identifies them, which would offer an opportunity to easily harvest them as term pairs.

The other day, I was musing over it but abandoned the idea. Simply, the mere change of the tagged word order in the target segment would mismatch the terms.

Yes. I don't say that it's a fool proof approach. But it can be time saving. The argument that human intelligence is needed after the harvesting, is valid for the Extract fragments feature too.

Other candidates for term pair marking are (from likely to very unlikely):

  • All kinds of quotes
  • All kinds of brackets
  • Different types of character formatting (with different likeliness!)
  • Number of words per segment (the less words, the higher the likeliness)
  • etc. (to be added by Michael)

Identifying term pairs based on tags

Login to post a comment