Kraynic said:
Ravenknight said:
Why isn't UDL properly tested on the Dev-server?
I expect because not enough of us were actually testing it by running games on the Dev server.
It is the job of a software provider, not the users, to ensure software is fit for purpose before releasing to production. The Pro users were given just a few weeks to test. The message was made loud and clear by Pro users on the Dev server that Dynamic Lighting, and later the Conversion Tool, were not ready for live and why.
That feedback was ignored. Instead, the software was pushed on the general populace on the live server as if it was ready for prime time. While a user still can avoid UDL, most users would not have known that was a good idea based on the following messages, variants of which came up in social media, blog announcements, and site notifications.
https://twitter.com/roll20app/status/1247947647153037314
https://app.roll20.net/forum/permalink/9053854/
Every major Roll20 feature release for the last two years that I can think of (bar one, mentioned below) has had major, game-breaking bugs that are revealed after 5 minutes of use and are not fixed for months. Unlike downloadable software, you cannot defer updates, so Roll20 updates need to be bulletproof before they are moved to live, but this same scenario has happened
again (Text Editor Update, I still have an outstanding bug that I ask to be fixed every 6 months or so)
and again, part 2 (AFoW update, the previous time the Dynamic Lighting system was rewritten from the ground up, dropped features, and made AFoW unusable with Global Illumination)
and again, part 2 (Animation, still has amazing upload times a year and a half later)
and again, part 2 (My Audio / Bring Your Own Beat, great that Roll20 prioritized this with Fanburst shutting down, but it was still released well before it was ready)
and again (WebRTC Update, made video chat unusable for me and my group for a year and then was fixed without fanfare by an update about 5 months ago. Yay!)
and again, part 2. (A New Light, our current situation)
The Charactermancer is one feature release that I can think of where the feature was tested on the Dev server first, and most issues that made it to live were edge cases that were handled relatively quickly. However, my players and I had no need for it until you could level up, so that could be clouding my judgement. Even then, it seemed like most of the Charactermancer bugs were edge cases of special combinations of features.
The other time Roll20 did well in this time period was the Token Bar update. This was necessitated from fallout over changes made (within the AFoW rewrite if I recall correctly) and is the one time in my recollection that Roll20 actually rolled back a poorly implemented feature, took it to the dev server, and used the Pros feedback to actually get it right before releasing it again to live.