Start a new topic

Use glossaries or fragment memories?

Your interpretation of the difference between glossaries and fragment memories is correct. There's only one way to find out which concept suits your needs and preferences best: try them out. Both have pros and cons. CafeTran doesn't force you to choose either glossaries or fragment memories. You can even use both at the same time. You can create a new glossary in three ways: as a project glossary (via the Dashboard), via Glossary > New (to start from scratch) or via Glossary > Add to add an existing tab-delimited UTF-8 text file. In the latter case, you no longer have to create a glossary resource file (kind of the URL to your glossary on your disk). CafeTran 2016 takes care of the creation. Make sure that you place the existing glossary file in a smart and permanent location on your disk, e.g. in your Dropbox folder, subfolder Glossaries. Changing the path to the glossary (which is stored inside the automatically created resource definition, is a little complex, so if you can avoid it by thinking about a suited permanent location for the glossary: please do). Note that fragment memories can do fuzzy matching and stemming automatically, whereas glossaries offer exact matches, unless you code the individual term pairs manually (by inserting pipes at appropriate positions). Then you will have "controlled fuzziness". Please experiment with both concepts.

[ For people who have no idea what is refers to:]

Alyissa: I don't understand the difference between adding a term pair to the "fragment base" vs. the "glossary". What is the difference in their functionality?

My answer:

It depends on what you want. The glossary is a simple text file that does not allow for much "fine-tuning" nor for fuzzy matches. However, you can use regular expressions in it, or add synonyms to an entry. The memory for fragments is a TMX file for words and subsegments. It allows for far better fine-tuning and subsegment matches, important when you use Auto-Assembling.

Unless you use regexes, I suggest you use the memory for fragments for both words and fragments. And if you use regexes, create a glossary that you use exclusively for regular expressions. Synonyms can be added to the memory for fragments as well, albeit not in the same entry.


I think that you're very honest and moderate in your answer this time. That's good! No hysteric shouting about glossaries in your posting. Perhaps you've learned something? The caveat is of course in the phrase does not allow for much "fine-tuning". As if she would know what you mean (and if it would be relevant ;)). And by the way, what do you mean with these quotation marks? Don't you mean real fine-tuning?

Thanks for your effort! On behalf of me personally. Mine. As in moi.

Login to post a comment