Good question, Ivan.

As with many bodies of collected information, the trick is to start early and grow the collection, rather than trying to build it all at once when people are busy with many other tasks.

The first step is to get people to agree that a glossary is a useful thing to have. The second step is to set up a structure and mechanism for collecting, storing, and accessing the information. The next step is to get people to think about whether any work they did today revealed terms that could profitably be added to the glossary, and to make that addition as part of completing that piece of work.

Building a glossary incrementally like this takes very little work or time on anyone's part. Someone does need to be responsible for editing and maintaining the collection, like removing duplicates or reconciling conflicting definitions.

Does this advice help? If so, I'll add it to the article! Thanks for bringing up this point.