Brammy wrote:The point I was trying to illustrate is that Dropbox is way more than just a sync service now with the Slack et al integration.
If that is your problem with Dropbox, have you looked at the huge list of bloated things that iCloud handles?
It sprawls as badly as iTunes.
iCloud is a complicated and over-engineered monster. I alluded to it above, but just because Apple hides this thing inside the operating system does not make it a slick, lightweight and cool. What that really means is that you have no choice in whether it is installed or not, and they can sprawl it deep into stuff that sync should have no business messing with in my opinion—and I state that as an ethical concern as well as a technical one. Just look at it this way: you have to give Dropbox permission to look at your photos. Did you give iCloud permission to do anything
? Apple doesn’t ask permission, they just upload your address book to the Internet like that’s cool to do, whereas in most cases we’d call that spyware.
But Dropbox integrates with a development toolkit, so it is clunky and bloated—abandon ship. Sure, sure.
Now the 3-device account limit on their freebie service, yeah I get the grumbles there (despite the whole Gift Horse Mouth bit), and like I say I get the preference
aspect, that is fine. But technical arguments for Dropbox being worse than iCloud just make my eyebrows rise.
Clearly, Apple is really good at selling stuff, while Dropbox does not care so much about that kind of thing. Now to a person like myself, that is actually not a point in Apple’s case; Bose is good at selling things too.
Rayz wrote:Mmm. I’ve found iCloud sync to be rock solid on the four devices connected to it. The apps sync seamlessly, even when they’re open on different machines.
The problem to my mind, and the thing to pay attention to with this kind of technology, is not the affirmations of when it works correctly, but what happens when it does not
. Because what happens when it fails is what we have to prepare for even when it seems to work fine for years. We post advisories about not using Google Drive even though the failure rates are a fraction of the people who use it (many of whom have used it for years). But because those fail states happen often enough, and because their condition is so catastrophic, we warn people away from the service entirely.
Now for iCloud, I can confirm some of the behaviour I have seen lunk describing (I have never used it long enough to see anything like extended outages). I have created isolated test accounts using iCloud throwaway accounts, and with a few simple steps, made in under two hours (mostly because iCloud was so slow
that every test took ages) I was able to observe the following alarming or detrimental behaviours:
- A tiny change to a Scapple file (editing one note), resulted in a 45 minute wait for the file to sync. In the meantime if I had not realised that and assumed iCloud were as snappy as other services, I would have conflicted the file.
- Changes were made to .txt files on an offline system, while an online system made changes to those same files. When the offline system was brought online, iCloud not only did not alert me to the fact that there were conflicts, it obliterated one set of the modified files. They were not in Trash, they were not archived anywhere (as far as I can tell, iCloud has no safety net like Dropbox’s deleted files for file versioning systems, but I am no expert), and since iCloud is so complicated and makes your actual file storage locations opaque, getting to where the files are at the system level to try and find backups was extremely difficult. The modifications were just gone.
So yeah, that is kind of more seamless in the fact that you are not presented with the fruits of your own errors, but err…
- The above was an extreme condition, iCloud does also handle conflicts (when it manages to detect them). If you edit a Scrivener project in two locations, or fumble the timing in waiting for it to fully sync (and good luck knowing when it is done if you do not use Finder much). Its method for handling them means duplicating the entire project and having to somehow figure out what is different between the two copies (or worse, you reflexively mess up and click one of the buttons to choose A or B instead of keeping both, and again irrevocably lose one of the conflict copies).
For myself, this kind of catastrophic failure over a rather common misuse scenario is something I have a “One Strike You’re Out” policy on. When I saw iCloud destroy data, that was it. I had no interest in that system and I never will. I have never seen any other sync technology do anything like that, and every other sync tool I have tested has extensive safety nets built in, like archival folders, versioning and trash-buffer systems. Maybe iCloud is better now than it was when I tested it—but like I say, at this point I do not care and I never will
, because any system published with the flaws I saw has said enough about itself.
…My experience (especially with very large projects) is that the syncing is not exactly seamless (I have spent a good few hours sifting through files to sort clashes because I’d forgotten to sync on one device, or because Scrivener had been unable to shut itself down for some reason).
This is the same sort of thing you would have to do with iCloud if you messed up—but as I pointed out above, the experience is a whole lot more opaque and difficult to recover from.
- You open Scrivener and it alerts you to the fact that Dropbox has created duplicate conflict files.
- They are imported into the binder into one convenient location and cross-referenced to the items they are associated with if possible, making it easy to compare the two copies side by side.
- You end up with two whole copies of the project.
- And then what? Well, I guess you can search by “mdate:<2d” or something to find everything modified in the past two days then methodically trawl through the list one by one.
In neither case would I say these outcomes are evidence of a lack of reliability in either system. They are symptoms of improper usage of them, and their own individual mechanisms for coping with said usage. Neither case is going to be fun to recover from, but at least Dropbox lets you know precisely what went wrong instead of “hey, somewhere in this 80,000 file trove you have here, a sentence within one file triggered a conflict, have fun finding it”.
…and if the cost of iCloud syncing is a complete rewrite of Scrivener from the ground up (and I am certainly not saying that this is Apples fault) then I’m not sure if such a move would be worth just for iCloud syncing, especially if I still have to keep one eye on the syncing process while I’m trying to write.
One thing worth noting is that rewriting Scrivener from the ground up is a euphemism for gutting pretty much all of the things that make it a unique tool for writers, with regards to large-scale research and text storage. We’re talking turning Scrivener into Yet Another Single File Text Editor, with all of the limitations that come with storing 200,000 words in a single file (and nothing much beyond those words, no research).
The other thing worth noting, and I see this misconception all over the place, is that unless Scrivener were gutted, then if we’re speaking of a hypothetical universe wherin Apple finally adds safe and simple support for bulk folder+file sync, using the iOS protocols—nothing
would likely change about the procedure.
What people refer to as “clunky” with Dropbox is more the technical limitations of the platform they are choosing to write with: a scalable system that isn’t fettered by single-document limitations. They are in essence blaming the road for being bumpy while choosing to use a vehicle that uses tracks instead of wheels and suspension. If Scrivener supported iCloud as the tool that it is, then you would be driving on a different highway with a different Brand Name, but it likely wouldn’t feel any different, because in this vehicle driving on a road feels about about the same as driving over a field of boulders.
The point is: try doing the latter in a lightweight electric with 4cm of clearance. You can’t, and that’s why Scrivener is what it is and not a gutted version of itself solely so that it can use the A-50 as well as the AC-14 highway. Scrivener can hoist a 250gb repository of research that encompasses a 10-year project to write a 300,000 word box set. Try that
in NovelWritr Xtremist, or whatever.
Scrivener came about before Apple really knew what Cloud was.
True, but it is even further back then that—you’ve been around since the beginning so you know. Scrivener came out before anyone on the planet started using that jargon to refer to a centralised synchronisation network that keeps local content updated. Scrivener beta (post Gold) begin in the summer of 2006, and the inventors of this genre, Dropbox, published in the summer of 2007, around half a year after Scrivener hit 1.0. At that point of course, only geeks were messing with this thing, and criticising it for basically being a simple rsync clone for dummies.
It wasn’t really until Snow Leopard hit shelves (literally, with people lined up around stores made of blocks of stone and other physical matter, to buy it) that this was something typical people were using. That’s about when we posted an advisory on how to use “cloud” technology safely.
Around this period of time, if you preferred Apple technology you were (a) utterly mad and (b) using a MobileMe service called iDisk which is just a WebDAV folder in a GUI wrapper that was notoriously, and in this case actually, unreliable. At this point Dropbox was already two years old, and the clones were starting to appear.
Apple seemingly still did not understand what Dropbox was when iCloud was announced. The first iteration, launched four years after Dropbox hit the scene, was best described as a complete misunderstanding of what made Dropbox popular. It was also unethically used as a tool to bludgeon developers into adopting the Mac App Store, and thus siphoning revenue meant for developers into their own pockets (but I digress).
I go a bit into the history here not only to point out the differences in the companies that bring these two services to the market, but to raise another point: since around 2008 or 9 or so, when people start using this technology with Scrivener (and Dropbox was pretty much the only game in town), I have not seen a single technical
failure on their part. I have seen an awful lot of people causing conflicted files, and some of that may be down to Scrivener faults (like the shutdown bug you saw), etc. But Dropbox itself actually deleting files or damaging data? Nope—it is probably one of the most solid pieces of technology I’ve seen mass usage of.
In conclusion, and importantly, I have no love
for Dropbox, lest anyone think I’m a champion of it. They have issues with how they store data without encryption for instance, and I don’t like how you have to use one monolithic sync folder (at least without geeky workarounds). I have to use Dropbox a little bit, and for that I have a command-line tool that is about as complicated as wget. It doesn’t run in the background, I tell it to list X folder as one command, download Y as another and upload Z as a third command. I don’t have Dropbox on any of my iOS devices, either as the little viewer they make or enabled in any my software that supports it.
So I don’t like Dropbox! But I trust it, as technology, a lot
more than anything iCloud. The way it works is simple and elegant. iCloud Drive is a Rube Goldberg machine, like much of Apple’s technology. It does 100 things where a simple switch would suffice. Apple’s really good at putting all of that into a box and making you think it is simple—but here’s the funny thing: even the box
is over-engineered (again, in typical Apple fashion). I have an automatic distrust of anything that has so many working parts within it to do something that should be pretty basic. With Dropbox and most other sync systems, you give it a file and it handles the file from that point on. With iCloud, your system is hijacked; files deleted locally; those that remain are moved into mysterious hidden folders with obfuscated naming schemes; Finder is subverted to make it look like they are where they were—but other tools won’t see them; no permissions asked to start uploading personal data, etc.
So yup—I don’t get it when I hear people going on about how great iCloud is and how clunky Dropbox is. All I can say is what I said before: Apple is really good at selling things
(and with a good dose of people conflating industrial strength workflows with the highway being used to transmit them, too, and wishfully thinking that if they used a differently branded highway, their tank would magically become a bicycle, and not realising what all they would lose by asking us to stop making tanks).