Paul Hammant's Blog: Apple Dissatisfaction
I have become dissatisfied with Apple Mac’s OS X and iOS over a number of years. I’m really ready for something else. That would be Linux, I think, but only after the last nail has been put in X11’s coffin.
HFS+ Bit Rot bubbled up a few days ago. It struck a chord with me because I have images in my iPhoto database that are now corrupt (like the author). That would be black bars half way down the image. I’ve no early warning system like the author has, so I can’t systematically say how bad my corruption is. I also can’t determine which ones are already corrupt in order to go back in Time Machine to retrieve the uncorrupt version. I mean I could, but it’s going to require a lot of work. Reddit discussed the authors blog entry, as did HackerNews. Those conversations meander though.
It’d be nice to have a file system that could spot the corruptions for me. Maybe it’s something DropBox could add to their sync daemon: “Hi, we notice that IMG_00882.jpg became corrupt, so we’ve restored it from history.”
Over the years I’ve had hard drive failures, and I’ve come to be very thorough with backups. Time Machine that is. The reason is a OS X wisdom I’ve come to agree with: your Mac hard drive will fail at some point, and not be recoverable. It’s going to be easier to recreate a Mac image from backup than attempt to repair. At least if Time Machine does not let me down.
I hate it and its database. My wife has iPhoto databases that, on load, will lose whole sets of pictures. This happened after a Software Update in January. We still have them in Time Machine, and can do something to recover them, but iPhoto the app isn’t part of that recovery while it won’t process the other iPhoto databases on import. She was happier with Windows Explorer’s unobtrusive presentation of photos, and therefor happy with a directory structure she managed. That was before I pulled her into the land of Mac. I’m inclined to agree that a directory structure is better. Death to proprietary database formats!
I don’t have the features I want in iPhoto either. There’s easy plugin architecture for things I might want to do, like “reduce resolution”, or “merge original and modified”.
I feel I want to use LynApp for photos. I’d be throwing a lot out with that. Integration with other applications and general and sync with iDevices, would be the key losses.
iTunes is more of a directory structure under the hood, than iPhoto is. It allows the archiving of shows/movies by hand to give a short-term boost to available space. Thats a relief at least.
It’s just a bad app though, both on the Mac and the iPhone. That I have to use it to manage iOS applications on my devices too seems inappropriate. That experience is definitely sub-standard. There is not enough fidelity. Say I want “App X” on my iPhone, but not my iPad. It is really not apparent in a tabular form what Where App X is installed and where it is not. I’m also allowed “on the device” and the Mac’s iTunes database. I’d like iTunes to manage a ‘removable storage’ place too.
Brian Windheim says it all in Time Machine sucks, use rsync instead, although I want better than Rsync, but I leave that for another day. In short: Time Machine is not robust enough for emergency usage.
Half life of accessories
We’ve had an original Apple TV that died when it was 18 months old. Ditto an Airport Extreme. After that I vowed not to but an apple accessory over $99. I also had a trusty Airport Express that I used to travel with a lot that worked well in Hotels where my Wifi was more reliable than theirs. It works perfectly, but I can’t administer it from Mountain Lion - they’ve just dropped support for it in the latest Airport Utility. It’s as good as it’s last configuration for me, or I’ll have to find someone with an old Mac if I want to tweak it again. I’ll not buy a WIFI hub that doesn’t have a web interface again. Apple can’t be trusted to not do that again.
Airport (Admin) Utility fiasco
Corey J. Mahler tells the story Features in Airport Utility 5.6.1 were taken out for 6.x onwards. Really Useful features that you would want to have going forward. Cory is writing about Mountain Lion (10.8), and did a Mavericks (10.9) update. He’s got some tricks for you so that you can continue to run Airport Utility 5.6.1 as well as 6.x as Apple would wish you to run. Great work Apple, you’re love of rewrite over refactor isn’t best practice, you know.
It has failed me before, and failed me again on the recent migration of my user account from my old Mac-Book-Pro (end of lease) to a new Mac Book Air. Same OS version on both, no disk permissions to repair (said to be essential), and a Thunderbolt 2 cable strung between the two. Yet Migration Assistant couldn’t find the other mac. It’s a very common problem it seems. I wasted a huge amount of time trying to get it working and it would not.
No matter, over the same cable I was able to use ‘ditto’ on the command line to speedily migrate my user account. I was willing to take the hit that I have to reinstall applications again, but sometimes that’s not a bad thing.
Migration Assistant works best when one of the disks is in Target Mode, and ditto is the same. After the copy completed, I had to do some ‘chown’ fixes to make the permission right.
I’ll never use Migration Assistant again as it is remarkably short sighted and unhelpful.
OS update uncertainty
You shouldn’t just reflexly update iOS or Mac any more without waiting a couple of weeks to hear what all the bugs are. Not just bugs, but what’s deliberately changed that you may not like.
I’m really not impressed. Whilst I get that there were problems with the synchronization in .Mac/MobileMe, I still really liked the fact that I had (essentially) a WebDAV mount. Everything has been serviceified, and the sync happens at moments of Apple’s choosing. I can’t trust that.
Also in terms of features, it does not have what I want. There’s me in iCloud, and my wife, and we want separate stuff, but we also want some joint stuff. Why can’t I fiddle with user permissions for arbitrary directories/files like GoogleDocs and DropBox has? I’d also like versioning built into it, so I can undo mistakes.
To some degree, WebDAV feels like an abandoned spec. It was always a little controversial, but there is a ubiquity to it some ten years later. At least all OSs can mount WebDAV urls as file-systems, and handle the reads/writes over the with transparently. WebDAV-Sync, and its commercial variant DAVbox provide more DropBox style capability.
Meanwhile Content Management Interoperability Services (CMIS) has come into being, and technologies like cmissync could be the enablers of it’s adoption. CMIS is “over HTTP” so would likely work with corporate firewalls.
Greg Stein (who pioneered lots of web technology stuff, and helped define the WebDAV standard/spec) tole me “CMIS is the New Hotness”, and suggests that a question on StackOverflow describes the changes well: CMIS vs. WebDAV.
I still like WebDAV though.
Some of these are more evolved than others…
Colleague Scott Robinson says “Dropbox has an infinite restore old versions feature, that’s good (but proprietary). I wonder why we can’t have this as a default thing on filesystems? Collapse and expire old snapshots. “Click here to have them auto put in Tha Cloud!”, he suggests.
I’m inclined to agree.
There’s also webhooks which was recently announced. It is very alluring, and I think it should be another thing that OSs should generally support (./index.html heuristics on a file system).
DropBox isn’t perfect though. It does not currently allow files and folders inside the “Apps” folder to be shared with other users.
Ex colleague Michael Robinson suggests Bittorrent Sync. He says: “It was designed from the ground up for exactly this use case (idiot-proof, secure, “cloudless”, multi-device, cross-platform folder synchronization), and does the job brilliantly. There is even a per-folder option to archive files on delete.”
I like SparkleShare a lot. That it has “git under the hood” is very alluring, especially as that’s only a server concern, and clients are receiving sync’d directories/files in the same optimized way as DropBox.
In lieu of better technologies, I would deploy my own SparkleShare instance and configure everything to rely on it for important files.
I’m just going to list them, and not dig deeper: