Given that extended attributes are available in most of the filesystems used in Linux today, Dropbox still hasn't answered why they decided to go "ext4-only" for a while.
I feel like in open source this tends to work itself out with community patches, but with closed source it can become a real pain.
My point being that the 0.1% of sales might be true per device, but "supporting Linux" might actually bring more revenue than just the devices themselves would indicate because the whole team's revenue would be lost if they didn't support Linux.
Typically Linux support in products like this is organic: It happens when the engineers themselves use Linux and push upwards to support it.
Anyway, with the amount of testing and verification we do on our product, desktop Linux is impractical. We try to verify on as many possible user configurations as possible. (Multiple Windows versions, 32-bit and 64-bit, Multiple Mac versions, remote desktop, ect). Linux is so heterogeneous that we could only really support it if it was a large customer where we targeted their typical configurations.
How would you know if you were?
“The sales team says that's a requirement for the companies they're talking to”
“Competitor X has Linux support and they're growing faster with no other obvious explanation”
I’ve been using Linux for around 25 years, some of that selling commercial software and most buying. There’s a very noisy contingent of the community who make a lot of noise but never actually buy anything. You’ll go broke chasing their feature requests because there’s always one more thing before they can switch.
In which context it might actually be perfectly correct to say that you aren't losing any sales over it, even if there are thousands of users who would use Linux support, were it there.
It's frustrating to me because academia is a major target for Box, and there are many of us linux users here. I wish the university admins that negotiate these deals would consider us.
It also feels like Box has a dismissive attitude towards us. Linux support has been a much requested feature for a long time, for example see this years-long thread: https://community.box.com/t5/Desktop-and-Mobile-Forum/Status.... IIRC mods encouraged people to vote for this feature request in BoxPulse, where it subsequently became one of the highest-voted features, but then labeled WONTFIX. Also annoying that Box supports linux-based Android, and also supported the less-used Windows phone. Yes there are a lot of linux distros out there, but Box could put out a snap, or just target one distro and let the other distros patch for themselves (like Steam does).
Thus, what happens is that so-and-so runs such-and-such which happens to be incompatible because of some whacky configuration alignment that no one considered.
The test matrix then becomes much more complicated than a traditional test matrix targeting common Windows and Mac configurations.
That's fine for expensive software where the end user might have a very close relationship with the vendor, but for "cheap" software the cost of supporting every possible way someone configured their computer is significantly higher than targeting Windows and Mac.
(Otherwise, the Linux version of a program might cost 2-10x what a Windows or Mac version will cost.)
I like to think of desktop Linux as a DIY hobby; but not something that you can expect commercial software vendors to support.
If you are targeting businesses, virtually nobody will run Mandrake (which does not exist anymore), Mandriva, Arch, or whatever. When you target Ubuntu LTS and Red Hat Enterprise Linux, you have probably covered most of the enterprise user base.
(Otherwise, the Linux version of a program might cost 2-10x what a Windows or Mac version will cost.)
This surprises me, because Apple changes and breaks a lot of stuff every macOS release and Apple only supports one or two versions back for security updates. In the meanwhile, RHEL only releases every half-decade or so and supports every release for much longer. Ubuntu LTSes are only released every two years and are then more or less frozen as well.
I like to think of desktop Linux as a DIY hobby; but not something that you can expect commercial software vendors to support.
Except if you care about servers. Or developers for that matter.
Sure, for everyday end users, supporting Linux can be a pain, but if we’re still talking Enterprise, there are only a few distros that would need to be covered.
I don't use desktop Linux myself and I'm not usually a desktop Linux advocate at all, but I know that a lot of developers use desktop Linux professionally for good reason. Public administration in Europe is also very gradually moving towards using more Linux on the desktop (in spite of well publicised setbacks). There is a lot of public pressure in that direction.
Desktop Linux may not be important now, but it probably has more growth potential than anything else. And if it grows even just a little bit, the impact on Windows/Mac-only cloud solutions will be disproportionate. It doesn't take a huge percentage of Linux desktops to make Windows/Mac-only cloud solutions very cumbersome.
I think it's a smart move by Dropbox to occupy that niche now, even if it's loss making. There could even be a bit of a moat, because Microsoft and Google both have reasons (Windows/Chrome OS) to drag their feet.
But using 'any' as an universal quantification is false, because some users like me are paying for products like pCloud precisely because they support Linux.
As a Dropbox competitor I know that you've lost at least a couple of sales because with all the companies I worked thus far the only sync solutions used at the company level are Dropbox and Google Drive. And the later is only used because it's attached to Google Suite, but nobody pays for extra storage.
And all the companies I worked with have some Linux users. And I know, correlation is not causation, but I know that at least one of those companies haven't gone with Box because of missing Linux support.
If you haven't noticed lost sales it might be because companies with Linux users aren't even bothering to check you out.
And yet, there are many open source projects have the scale of a file synchronization tool or larger, that manage to work fine (and run a large battery of tests) across different Linux distributions, BSDs, and macOS.
I don't think it is impractical, it is probably just not a good business case.
(I don't necessarily agree, I think Linux users will typically send more fine-grained and useful bug reports and might help you shake out bugs more quickly.)
Even if they haven't and it's literally 0.1% Linux users, I think that it's unlikely a team will pivot on a sale because of a few users who insist on not using ext4. Instead, those Linux users will probably just end up creating ext4 partitions somewhere so everyone can get on with their jobs.
1. Dropbox isn't just storage. If it was just storage, it would never have taken off. The reason Dropbox makes tons of money is because it's actually selling a process. Still commoditized, but it's not as lightweight as toilet paper.
2. If you feel like creating an ext4 partition is analogous to remodeling your bathroom, I would question where you choose to hold strong opinions.
I think you have your analogy backwards. I'd happily switch out the toilet paper brand (the file system on one partition) in my bathroom (how my team and I manage shared documents) if that's what it takes to make the bathroom fit its purpose.
Besides, Dropbox didn't kill Linux support, they scaled it down to ext4, which is a perfectly inoffensive choice of file system for an end user.
Replacing this functionality on ext4 requires an entire stack of alternatives. Worse if you opt to use zfs and ext 4 you now need 2 complete stacks or to use an inferior set of tools to be compatible with a commodity service.
I work in a bioinformatics lab, and I access Dropbox through the command prompt and do a lot of file cleanup through scripts for dozens of researchers, which is impossible on Windows or Apple.
This BTRFS reversal last year was a nightmare for everyone's research even though I'm the only one who uses Linux.
Out of curiosity, why is that? macOS does provide a Unix environment and Bash & Zsh shells. (and Homebrew or MacPorts are package managers that can provide updated versions of whatever)
Or just use Python across all three platforms?
My point was more generally about: surely tools exist on the different operating system to achieve similar outcomes?
I use Linux because the scripts are easily moved between the cluster computers and my computers, but I suppose YMMV
You’d definitely want to use Hombrew: you can get the GNU utilities instead of the BSD ones Apple tends to ship, at the latest versions, and it’s a single command to update everything.
May be another explanation.
Also the number of bugs that get marked as Won't Fix that are real problems by upstream doesn't help so a lot of people like myself just don't bother after a while and just end up using something else or living with it being broken.
In the end this makes the code more robust for all users, not only Linux users.
Why would it make the code more robust for users on other platforms if they aren't experiencing those bugs?
Is exposing these bugs worth the extra development time? Surely just compiling with a different compiler is as likely to find bug in the compiler.
I am not saying you are wrong but you are making a lot of assumptions.
Because different compilers and different stdlibs do things in subtlety different ways, that may just work on one platform by chance. Memory alignment/layout is often slightly different for example.
Then IO behaviour is often quite different on Linux and OS X despite both being POSIX... (and once you need high performance or Async IO, you're using native APIs anyway).
Memory behaviour is also quite different in terms of buffering, when memory gets freed, etc - e.g. on Linux you often need some malloc_trim(0) calls to actually release recently-free'd memory), on other platforms that's not as necessary, so there are subtleties there.
Different platforms often have different includes as well (on Linux, you generally use GCC's STL/stdlib, on Windows it's a totally different set).
A new compiler / OS update / graphics card driver in the future might trip the bug on the platforms it seems to be working fine on currently.
If Pear OS 8.0 is release a month later and breaks my application because of an API change their end. I would just release 2.0.1 of Music Box that fixes the defects. If I was proactive get hold of a beta or release candidate of Pear OS 8.0 and QA my application on there to see if there was any obvious defects.
The same with a newer compiler or a newer hardware driver. I would just get hold of the beta / release candidate and fix it then. The one thing I wouldn't do is try compiling it on a completely different OS that none of my users use.
Tickets showing actual bugs not tech support requests from people that can't read directions are valuable resources.
Might be fair to note that these numbers are for a video game, not a file syncing app. Linux is relatively new territory there, let's say when Steam started supporting it.