I created this task because the developers of Modelviewer pointed it out to me in this Twitter thread as a vehicle for allowing us to implement T246901: https://twitter.com/modelviewer/status/1319785320246775808?s=20

Personally I don't know what would be for/against considerations for us to consider using such existing software, hence this ticket, which allows for focused discussion.

@Aklapper are you aware of any other examples where preexisting software has been used in Wikimedia projects we could use as a model?

No as I have nothing to do with this field. :) Please ask on a public mailing list or such - thanks!

I am not a developer, but I have a hunch that it definitely needs very strict code review, and thorough investigation whether embracing external software will benefit us in the long run. Is it indeed open source, and in that regard, is it totally compatible with our own codebase? Is it hack and tamper-proof and will it not be a magnet for various problems (important, considering how visible Wikimedia projects are and how much vandalism and other malicious attacks we are already subject to)? What kind of community is behind it, and is it reliable and in line with Wikimedia's values? Will it be around, supported and further developed for a good amount of time (I'd say 10 years or more) so that we don't end up with something we need to maintain ourselves with our own very limited resources?

These are just a few of the considerations I'd come up with that seem common sense to me. It would be great if others can chime in.

Since we seem to have tried this before for the Kaltura viewer, it's worth attempting. The linked ticket about 3D model viewing suggested it might be faster than rewriting. And doing it w/ support from the Modelviewer devs would bring new coders w/ experience we wwant into the ecosystem; that may be the deciding factor.