You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As we have continued to test on real site, I have found myself continually relaxing the restraints on what is considered for an element to be interactable. We had pretty hard criteria (still mostly do), but I am now starting to automatically allow links to have mouseover events because of running into https://coag.gov/ where we need that to explore the dropdowns.
So the question is, at what point do we just start trying actions on every element on the page that we can see. In the example above, there is no reason that those couldn't be div elements or something else and our current solution would fail. The two big downsides I see are:
Runtime - Trying all of these actions, especially if they create a lot more edges, could increase runtime significantly.
Global Listeners - Similar to 1, but we have seen carousels that move no matter where you are on the page, with trying everything we would significantly multiply the number of times we interact with the carousel (or whatever global object).
The positive would be that we can be more sure of actually creating a full graph and not missing any content.
The text was updated successfully, but these errors were encountered:
As we have continued to test on real site, I have found myself continually relaxing the restraints on what is considered for an element to be interactable. We had pretty hard criteria (still mostly do), but I am now starting to automatically allow links to have mouseover events because of running into https://coag.gov/ where we need that to explore the dropdowns.
So the question is, at what point do we just start trying actions on every element on the page that we can see. In the example above, there is no reason that those couldn't be div elements or something else and our current solution would fail. The two big downsides I see are:
The positive would be that we can be more sure of actually creating a full graph and not missing any content.
The text was updated successfully, but these errors were encountered: