search

phetsims / molecule-shapes

"Molecule Shapes" is an educational simulation in HTML5, by PhET Interactive Simulations.
http://phet.colorado.edu/en/simulation/molecule-shapes
GNU General Public License v3.0
5 stars 6 forks source link

Metacog wrapper recorded wrong behavior #208

Closed KatieWoe closed 2 years ago

KatieWoe commented 2 years ago

For https://github.com/phetsims/qa/issues/745 When recording the sim on iPadOS 15.1 with metacog, I went into options and turned on some options such as lone pairs. When watching the playback, a screenshot was downloaded instead.

KatieWoe commented 2 years ago

Watching the playback, I think I know what happened. On iPad, there are three things in the PhET Menu: Options, Screenshot, and About. On computer (where I was watching back, there is a forth (Full Screen). So when I clicked Options on iPad, it corresponded to where Screenshot is on other platforms.

jonathanolson commented 2 years ago

Ahhh, presumably we'd need to include this metadata for playback to work. I also recall that we'd want to include platform details, font size metrics, etc.

@samreid thoughts?

samreid commented 2 years ago

There is generally a constraint that you are supposed to try to watch the playback on a similar platform to where it was recorded. @zepumph do we have anything beyond this?

KatieWoe commented 2 years ago

I thought that the reason for the metadata testing on iPad was because the files can't be stored and used by the playback wrapper on the iPad, so it has to be watched on a computer?

zepumph commented 2 years ago

Hmm, In general @samreid communicated our philosophy thus far (playback should occur on the device that it was recorded on). It is infeasible to expect that we can generally cover every single possible discrepancy between two different runtimes of a sim. That said, this specific bug seems like something we should fix since it is for a common recording device that doesn't support the playback wrapper to my knowledge. And THAT said, it doesn't seem trivial, and so even if we don't just write it off as a won't-fix, it still may not be high enough priority to fix it. To me this really comes back to how important the playback WRAPPER is. Perhaps we should ask designers.

samreid commented 2 years ago

We dropped support for playing back input events and are moving toward state-based playbacks. So I don't think work is warranted for this issue. Closing.