The editors of Topology resigned over the high price (apparently it was $1558 in 2004: https://pages.uoregon.edu/dps/journals.php), and started Journal of Topology which at the time of announcement (https://www.lms.ac.uk/sites/lms.ac.uk/files/About_Us/news/20...) was said would be published by LMS (in partnership with OUP), and priced at $570. Looking at the page now, it is still published by LMS (in partnership with Wiley now), and I can't figure out the pricing. It is fully possible that the price has increased, but there has been no change from free to non-free. Back in 2006, "free" was not one of the conventional options for a journal. Now it is.
The editors of Topology resigned over the high price (apparently it was $1558 in 2004: https://pages.uoregon.edu/dps/journals.php), and started Journal of Topology which at the time of announcement (https://www.lms.ac.uk/sites/lms.ac.uk/files/About_Us/news/20...) was said would be published by LMS (in partnership with OUP), and priced at $570. Looking at the page now, it is still published by LMS (in partnership with Wiley now), and I can't figure out the pricing. It is fully possible that the price has increased, but there has been no change from free to non-free. Back in 2006, "free" was not one of the conventional options for a journal. Now it is.
As for "a journal is not an easy task to manage", my favourite article is about the Journal of Machine Learning Research, the leading journal in the field of machine learning: https://blogs.harvard.edu/pamphlet/2012/03/06/an-efficient-j...