GNOME Bugzilla – Bug 308498
Change the rule used to determine what is written to training file, and when
Last modified: 2021-05-26 09:58:32 UTC
At present the learning of Dasher's language model happens on the fly as the root node leaves the canvas. That is grand. What I would like us to change is what gets written to the training file. I suggest that _what_is_in_the_text_box_ should be written to the training file, when (a) the user hits "new" or (b) the user selects a new alphabet. This will mean that modifications put into the edit box by hand will be included in the training file. At present, I believe that manual modifications to the text box never make it into the user's training file.
How this interacts with various components needs to be considered
Indeed, I would suggest, if possible to put the new learned text into a temporary memory and only save it the training text if the user whishes it. - Why am I talking about a temporary memory? I don't know how the prediction works, but I would like to have new words immediately available; not only after having saved them to the permanent training text. - The user should have control about whether a training text should be modified or not. The reason is not only to avoid the text of becoming to huge, but also to avoid spoiling up a training text that satisfies the user.
GNOME is going to shut down bugzilla.gnome.org in favor of gitlab.gnome.org. As part of that, we are mass-closing older open tickets in bugzilla.gnome.org which have not seen updates for a longer time (resources are unfortunately quite limited so not every ticket can get handled). If you can still reproduce the situation described in this ticket in a recent and supported software version, then please follow https://wiki.gnome.org/GettingInTouch/BugReportingGuidelines and create a new enhancement request ticket at https://gitlab.gnome.org/GNOME/dasher/-/issues/ Thank you for your understanding and your help.