So, a radical concept, possibly, but let me know what you think about this.
In the Pyblish QML application, you’ll be able to “login”. The credentials will come from this forum and in so doing your currently running instance of Pyblish QML will gain access to the forum on your behalf.
Having logged in, we can do all sorts of cool things.
Store settings in the GUI on a per-login basis, as opposed to per-computer or user in a company
Introduce comments on plug-ins and instances, that are mirrored between forum and GUI.
Link conversations to a particular raised exception, like a validation failure. They could appear on the forum, an live in the GUI.
And finally, drumroll please, to automatically create a new topic for each publish, with data relevant to the particular publish, like metadata, images, video, 3d preview. Basically everything Pyblish Event is meant to handle.
To avoid cluttering up the forum with hundreds of publishes, they could all reside under a particular, unlisted category.
Clash of the Titans
This would of course overlap with what Shotgun and FTrack is doing, as is Pyblish Event, but for new users and for those who don’t want or care about that, this could be a gentle introduction to it.
It sounds like a great concept but I fear that it would remove the line between what belongs to a company and what is public domain. Even if you made the publishes unlisted, information would be stored online which is not ideal.
The location of data is always going to remain flexible.
Using the public (this) forum would be the default behaviour, to get people going and to demonstrate benefits. But once you’re looking to sandbox your data (which you should) then you could re-target any queries made to your own database, either backed by forum or otherwise, or set-up a forum similar to this one but running locally within the company.
It’s a bit bloated as well; we would end up with a database of conversations and pipeline data baked into one. Both tools and humans would be accessing it, which means, for example, during times of heavy conversations, tools would see a performance drop and vice versa. But perhaps more importantly, loss of data in pipeline data could result in loss of topics on the forum and vice versa.
To be honest. I don’t see many benefits in this. At this stage it’s roaming very far into ftrack/shotgun territory but only scratches the surface which makes it stuck somewhere inbetween. If one needs to send information about a give publish somewhere outside pyblish, then making a conforming step should be enough to achieve that.
On top of that, Adding a feature to the UI that is so specific to a certain platform like this forum (or company private forum) seems a bit counterproductive and a waste of time. It might look cool as a promo feature, but I can’t quite imagine managing even small production using a forum.
