Future of FUSE for OS X

[Benjamin Fleischer]

Hello all,

As some of you might know FUSE for OS X will turn five years old in a couple of days. I would like to use this occasion to say thanks to everyone who has contributed to the project and helped to reach this milestone. It has been an interesting time and I believe FUSE on OS X is in a better place than it was five years ago.

In the past 12 months FUSE for OS X has been downloaded over one million times from the project’s website, not counting third party redistributions. It has become the basis for many successful products.

Nonetheless, we have not been able to reach all of the goals set out five years ago:

• The project is still very much dependent on one person. The list of contributors is very short. Contrary to popular believe the core team consists of only one developer and there hasn’t been any code contribution to the FUSE kernel extension by anyone but me in years.
  
• Another goal was to modernize the project and port new features from Linux FUSE to OS X. This is still a work in progress. I’ve been releasing previews of version 3.x for one year now but there are still some issues that need to be worked out. Free time is a limiting factor.

I spend most of my spare time keeping the software alive by fixing bugs, modernizing components, and implementing new features. But in my opinion simply keeping the project alive is not enough. The world keeps on turning and the todo list seems to be growing faster than I can scratch items off:

• Finish work on version 3.x
• Improve locking code in FUSE kernel extension (non-exclusive locking)
• Implement new FUSE features e.g. notifications
• Modernize the Objective-C framework, look into Swift support and libfuse 3
• Write a test framework to catch regressions
• Rewrite Preference Pane
• Improve documentation
• Modernize demo file systems
• ...

FUSE for OS X is being released under a very liberal open source license and can be used freely by everyone. In this context the word free expresses that everyone is granted the right to deploy the software for any purpose, to study its source code, to modify it as desired, and to pass it to others. But it does not mean that the development of open source software is free of cost. 

In order to take FUSE for OS X to the next level I would like to spend more time working on it, but this won't be possible without being compensated for my work. I’m not the one who invented FUSE or ported it to OS X in the first place but I’ve been maintaining it for the past five years. Certain tasks, like improving the FUSE kernel code, are complex, time consuming and generally require more work than a few hours every now and then. I believe this to be one of the reasons there haven’t been more third-party contributions.

My day job does not revolve around FUSE and I don’t sell any FUSE-based end-user products. To be able to spend more time on FUSE and cover the development costs I started taking small FUSE related consulting jobs and developed a rebranded pro version of FUSE for OS X, that can be distributed as part of an app and does not need to be installed separately.

Turning open source into a business without changing the licensing model can be difficult. Why should anyone pay for something that is seemingly available at no cost? And of course, FUSE is not an end-user product, it is a software development kit primarily used by developers to create end-user products. Therefore, I would like to ask companies selling FUSE-based products to consider supporting its continued development financially on a regular basis.

FUSE for OS X is valuable to developers and users alike that rely on it for their daily workflows. That’s why I would like to start an open discussion about the project’s future. I invite everyone to share his or her thoughts on the matter. 

Best regards,

Benjamin

[Peter West]

No responses I see.  It's been too many years since I wrote any system level code for me to be able to contribute, and I don't have the time necessary.  I would be happy to pay (depending on the cost, of course) for a commercial product.  Have you spoken to any venture capitalists about what would be needed to commercialise the product?

Peter

[Alex Aranda]

I agree, osxfuse is a great software, and if paying for support or some kind of access account will keep it alive and growing, i will gladly pay for it, of course keeping a open source license is nice for the source, but i see no problem in paying

Sent from my iPhone

--
You received this message because you are subscribed to the Google Groups "OSXFUSE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to osxfuse-grou...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Michael McGlothlin]

More general response than just for this project..

I'd be willing to contribute but I have limited time and energy to get myself up to speed. Good developer documentation is vital to making it easy for people to help. It can't be expected that contributors are already an expert in your code.

I have decades of experience with open source projects and sadly a lot of this experience has been negative so I tend to avoid trying to contribute these days. I just assume that projects are set in their ways and not interested in help. Unfortunately overcoming the bad behavior of others is necessary if you'd like contributors.

Something somewhat unique to this project is that it's confusing what the relationship is to other FUSE projects. I guess this goes along with better documentation. From the Linux world I'm familiar with FUSE and if I'm looking to porting to OSX I will probably find a couple options, including this one, and it's confusing to determine which I want and how to get started.

Mostly I want to be able to write cross-platform file systems in my choice of languages (usually C#) with as little wasted effort as possible. If I can work in from my own projects and expand into the core and know I won't get crapped on for trying to contribute I may well do so. Knowing you want help makes me interested.

Thanks,

Michael McGlothlin

Sent from my iPhone

--

[Benjamin Fleischer]

Hi Peter,

Thanks for your response. I’m kind of disappointed to see that only so few people seem to care enough about FUSE for OS X to write a short reply. In all fairness, in addition to the public replies here, I’ve received three off-list replies, but that’s it.

To answer your question, I have not spoken to venture capitalists. I think the most effective way to commercialize the project would probably be to drop the open source label and change the license to a non-free for commercial use model. But I’m not a big fan of this option. I was hoping more companies would see the value in FUSE and offer to support the project.

Regards,
Benjamin

[Benjamin Fleischer]

> Am 05.07.2016 um 02:44 schrieb Michael McGlothlin <mic...@kavlon.com>:
>
> I'd be willing to contribute but I have limited time and energy to get myself up to speed. Good developer documentation is vital to making it easy for people to help. It can't be expected that contributors are already an expert in your code.

I agree. The problem is that good documentation does not write itself and it needs to be updated. This can be quite time consuming. The FUSE API is documented and there are a few demo file systems. loopback and LoopbackFS are pretty good starting points. The other file systems are somewhat outdated and need work.

Improving the documentation and updating the demo file systems is on my todo list but so is a lot of other stuff.

> Something somewhat unique to this project is that it's confusing what the relationship is to other FUSE projects. I guess this goes along with better documentation. From the Linux world I'm familiar with FUSE and if I'm looking to porting to OSX I will probably find a couple options, including this one, and it's confusing to determine which I want and how to get started.

There is only one FUSE project for OS X, that is alive. MacFUSE was abandoned in 2009 and Fuse4X has been unsupported since 2012. So, if you want to use FUSE on OS X you don’t have too many options.

Regards,
Benjamin

[Alexander Skobelev]

Hi Benjamin,

It is two years later now from your original post, but have you ever thought about starting a project on kickstarter.com? I see that Magit for Emacs has been able to gather more that $77000 recently. 

[Heinz Zerbes]

Good idea, Alex!

I would also support Benjamin or anyone else who helps to keep osxfuse alive.

[Heinz Zerbes]

Thanks a lot for your continued involvement in OSfuse, Benjamin!

Being able to encrypt a shared Dropbox volume helps a lot in my daily collaboration. There are so many good and important tools out there that it's hard to thank every single developer for his efforts, of course. As mentioned, a kickstarter campaign might get in some money but that also requires a fair amount of effort to set up and then run. Do you have a simple paypal link that people can contribute, especially while it still seems to be only you updating the software?

If so, it might make sense to repeat the Paypal link here and in other posts every ow and then, or put it in your .sig ?

[Rich Schmitt]

Ben,

If you still could use help, feel free to get in touch.  Not sure what I can help with but as I've said in a thread today, I am moderately familiar with fuse on Linux (at least from an application perspective).  I have a custom SMB stack that I add to linux as a fuse module (for cases where samba licensing is problematic).  I am in the process of porting that fuse module to MacOS.  I would be happy to help in any way I can.

Rich

[Chris Jones]

I'd start a https://www.patreon.com/ for the project and add a link to it on the github readme as most of the active projects I run across on github tend to be using that service to aid in funds.  And I totally agree, there is a high cost to open source software that people often overlook, even more so if there isn't a corporate backer that financially supports the product.

One project that seems to be doing well in collecting funds would be webpack

https://github.com/webpack/webpack

The webpack guys n gals use a service known as opencollective, and another project that I've been active with is Inkscape, and those guys n gals promote people and companies who have donated to the project.

cheers

github.com/ipatch

[mixtly]

Hi Benjamin,

you mentioned that you started taking small FUSE related consulting jobs and developed a rebranded pro version of FUSE for OS X, that can be distributed as part of an app and does not need to be installed separately?

Do you still offer these services and is there some place where we can learn more about the 'rebranded pro version of FUSE for OS X'?

Thank you!

[Urban Kronenberg]

Hello Benjamin

The product manager of the company I am working with would like to get in touch with you regarding a financial contribution to the project. Is there a procedure in place for that? Thanks in advance for a quick reply! – Keep the good work up. Best, Urban

[Yeon Seunghoon]

Hello Benjamin,

I sent you an email about using FUSE for macOS in commercial product in May and didn't received a reply.

Thanks for your work.

Best,

Seunghoon

2016년 5월 29일 일요일 오전 12시 16분 22초 UTC+9, Benjamin Fleischer 님의 말: