Last Week on My Mac: 10.16 is all about tools and maturity

In just over two weeks, Apple will host its first online Worldwide Developers Conference, at which we expect it to release the first developer beta of macOS 10.16. I can’t promise you any sneak preview of what it will announce there, but here are my personal thoughts as to what it must get right this time.

The last major releases of macOS have brought huge structural change, more than any other series of operating system releases in the history of the Mac: High Sierra brought a brand new file system; Mojave made that obligatory for all boot disks, added stringent and intrusive privacy protection and introduced notarization and substantial changes to security; and Catalina dropped all support for 32-bit code, made the greatest change to boot disks since the introduction of Mac OS X, enforced notarization, extended Gatekeeper checks and privacy protection, and more.

We now need a period of consolidation, in which the tools provided with macOS catch up with these changes.

Starting with fundamental tools, both Disk Utility and its command line equivalent diskutil have yet to catch up with the most recent changes in file systems and their features. Disk Utility still lacks essential features. For example it’s unable to compact sparse bundles, and preparing a disk for a clean re-install of macOS 10.15 is elaborate and usually obstructed by a succession of errors.

Although one of the former headline features of APFS, Fast Directory Sizing is currently unsupported by either Disk Utility or diskutil. That command tool is now a labyrinthine mess, and needs to be fundamentally redesigned, either into two separate tools, or APFS commands need to be properly integrated into a coherent verb and option structure. Although macOS provides rudimentary support for working with APFS snapshots, third-party tools like Carbon Copy Cloner provide far superior features, and Disk Utility is unable to inspect or manage snapshots at all.

A year ago, at WWDC 2019, session 710 spent some time delving into the new features in Apple Software Restore for working with APFS, extolling these to developers as a means of restoring volumes from snapshots, for instance. Since then, developers of third party cloning and backup tools, such as Mike Bombich of Carbon Copy Cloner, have looked closely at asr and found it too unreliable for production use. Most recently, because of a bug in macOS 10.15.5, they’ve been forced to return to use asr to work around that bug. However, as Mike Bombich writes, that uncovered another bug in asr which makes replicating Catalina system files to a disk image file unreliable.

I think we all, with the apparent exception of Apple’s marketing staff, recognise the problems of trying to start a Mac up from an internal hard disk on which Catalina is installed. Yet two of the three base configurations of 21.5 inch iMacs currently offered by Apple come with 1 TB internal hard disks. If Apple is going to continue to saddle users of brand new Macs costing £/$/€ 1200 and more with APFS and its known performance problems, then they and others with earlier Macs need tools to maintain them. As Apple has still failed to provide sufficient documentation about APFS to enable third-party developers to attempt to defragment APFS volumes, there seems nothing those users can do other than pay out more to have their brand new hard disk replaced with an SSD.

Once you look more widely, there’s no shortage of essential bug fixes outstanding in macOS 10.15.5. There’s still a great many users – including plenty with the MacBook Pro 16-inch 2019, for example – whose expensive Macs suffer frequent kernel panics, particularly when trying to wake from sleep. There’s the mysterious checks performed to see whether unsigned shell scripts have somehow (how?) been notarized, when they’re first run, a feature which was never mentioned in any of the presentations of WWDC 2019. There’s the longstanding issue that Console still doesn’t provide sensible access to anything in the log which happened the moment before you open it, without creating a logarchive and opening that – a feature which Apple has deliberately omitted for four years now.

What I’d really like to see before anyone launches into Keynote addresses later this month is a review of all the fixes and improvements which are coming in 10.16 to address these shortcomings in tools and the fundamental maturity of the last three years of macOS. Without them, 10.16 has weak and flawed foundations, and the more it changes, the greater the risk it will fall flat.