The main problems are there have been no released versions since 2017, combined with an annoying "open core" development model where some nice to have features have been held back from the free version (although not ones we care about too much). For our use case (virt-v2v) speed is not an issue, but data integrity when reading, writing and modifying many dozens of files absolutely is, and ntfs-3g has been very reliable.
(Disclaimer: I've made occasional contributions to ntfs-3g upstream)
To make backups of data on a Windows partition. Playing games residing on a Windows partition. The main purpose I can think of though is forensics (images might contain NTFS).
HN gets all bent out of shape about minor differences between btrfs and zfs. Like how zfs supports raid5/6 but btrfs raid5/6 support is immature. Meanwhile NTFS doesn't even support checksumming, multiple block devices, copy on write, the performance on SSDs is garbage, etc etc.
The whole of the NT kernel was essentially a futurized implementation of VAX-VMS.
Both VMS and NT had the same manager, in the person of Dave Cutler. Microsoft hired a large team away from DEC to build NT (which became the subject of litigation from DEC).
I have read that NTFS shares common architecture with Files-11/ODS-2&5 (native VMS filesystems). Aspects were also lifted from OS/2.
ZFS was developed a decade after NTFS. Most UNIX was still on FFS when NTFS was born.
Stability hasn't been an issue for us. Virt-v2v uses ntfs-3g intensively and we have migrated a large fraction of a million VMs, many (the Windows ones) all using ntfs-3g. We've not had problems with ntfs-3g at all.
I'm still interested to see if an in-kernel driver can be better.
(Disclaimer: I have contributed a few patches to ntfs-3g upstream)
