Field Linguist's Toolbox
software.sil.org
software.sil.org
1. Gloss alignment in the file format is governed by magic (is it driven by the unicode display width? Is it UTF8 encoding size in bytes? Is it both?)
2. The file header has this magical number in it that just has to be the right one, otherwise you get weird glitches in the UI (like the file can't be edited). We get these regularly once per week when working with version control — the magic number seems to change randomly with no explanation
3. Does not work well for large teams due to shared dictionary. We have a decent git workflow in the meantime, but it's not easy.
4. Complicated project setup, especially if you want to share things. Not version control friendly — too many config files, too much system-specific stuff in them.
5. No native macOS version. This is a big one. We got it running under Wine on the new ARM Macs but I feel like it's a ticking bomb. It's fine for field people who will use a cheap disposable Windows laptop anyway, but not ok at all for people doing research on these corpora.
Overall, Toolbox was a great tool when it was first released, but a lot of time has passed and Toolbox did not keep moving. Unfortunately, newer tools (such as FieldWorks) made things more complicated without really improving anything. Toolbox's selling point is still its simplicity — at least we can parse the files with our analysis tools.
Another reason for complexity in FLEx is that it tries to prevent you from shooting yourself in the foot--e.g. by by inconsistent spelling or inconsistent use of fields.
FLEx does not however try to cover all possible dictionaries, e.g. its ability to work with dialectal variation is not its strong point.