I'm a little bit late to the party here, but I thought I'd better chirp up and correct this snippet.
Gordias' immense difficulty spike was predominantly down to SE's QA team trying to save time by testing the encounter purely phase by phase. It's critical to be aware that not only was this with fresh cooldowns and full HP per phase, but for much of the testing period they also had various invuln buffs. It's been openly stated by Yoshida that the QA team never cleared A3S or A4S end to end.
The proof is in the pudding really. Whilst Yoshida claimed that the difficulty would be tuned back down somewhat for Midas, A6S ended up being significantly tougher than A2S. It's also not much of a stretch to suggest that A8S was mechanically a much more complex encounter than anything Gordias had, it just wasn't riddled with latency dependant mechanics and the DPS requirements were tuned to much more realistic levels.
TLDR: - The community's feedback was not to blame for Gordias' difficulty. Rather, the fault lay primarily at the hands of Yoshida's QA department.
https://www.4gamer.net/games/278/G027835/20151224179/ - One of the interviews where Yoshida talks about this.
https://www.bluegartr.com/threads/12...tions-requests - And here's a translation.
To get back ontopic. The talk about the overall weakness of the design of healer gameplay is bang on point. Honestly, in my opinion it goes much deeper than just the nature of each healers toolkit. There have been some encounters over the years with enjoyable quirks and gimmicks that did well to alleviate the expectation of a healer to just sit there throwing rocks. A2S and A5S are probably the best examples of this.