Several months after "It's time to talk about Free Software Again" was published, I was contacted by a reporter from a business news magazine as a result of some Linux-related volunteer work I had been doing. The reporter started asking questions about the intersection of Linux/free software and business, and I directed him to Bruce Perens.

I don't remember the exact rationale at the time, but having led Debian during a period it was making regular releases and at the same time writing the Debian Free Software Guidelines and Debian Social Contract, I can't think of any better person I could have directed a member of the press to talk to about these issues.

Regarding codes of conduct, it's not really an area I have much basis to speak on, but the Linux Foundation's Professional Code of Conduct seems about right to me: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fdocs.linuxfoundation.o...

While the corporations are doing their thing with app stores and privacy-related issues, the PinePhone and PineBook series look like they're shaping up to be a nice option that keep the values that drew me to Linux intact: https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.pine64.org%2F

From my point of view, the OSI uncovered that the Linux community was primarily made up of technical geeks who were interested in economical, reliable systems and decent jobs more so than ideologues interested in debating the moral and ethical implications of various licensing schemes.

This was accomplished both by shifting the focus to the practical from the ideological tone struck by the FSF and also when ESR returned from a meeting with Apple to pitch a license that was incompatible with the GPL as being OSD-compliant and "the community" was having none of it.

The Coherent Open Source effort by Bruce Perens reflects a continuation of what I have seen him stating over the years - that a few good licenses allowing software to be linked is the operative need when building a system.

Regarding Slackware and Debian in the 1990s, I had installed Slackware from a stack of floppies on a few systems before trying Debian, and often needed to download source code from ftp sites to get fairly basic software. Most of this software was available through dselect on Debian where it was then kept updated through releases.

Then KDE was released and Debian release cycles started stretching out, resulting in it becoming more of a barrier to getting the software I wanted. I would consider dselect/apt and the DFSG to be some of the most important contributions to Linux/FOSS of the era, though.

I wasn't trying to write a timeline, rather to discuss the significance of the OSI back then. I was already using Linux at work in 1996 as well, but it was a smaller backwater shop running on a dual booting desktop converted into a server running at my desk. The main server even at that small shop was running Solaris.

At bigger shops, operating systems like Solaris and AIX were used for anything official until around 2004. That was the year when when a shop where I stood up the first Linux webservers, migrating from AIX made the Linux news for doing so. The OSI framing "Open Source" as being the important part about Linux was hugely helpful back then. In the late 1990s, "free software" was understood to be shareware with [UNREGISTERED] prominently displayed in the UI, nag screens and full functionality in a paid-for registered version or freeware. The quality of the freeware internet applications for Windows 3.1 was actually one of the key points that pushed me to use Linux at home back then.

By comparison, the actual work of the OSI doing things like determining whether a license met the open source definition always seemed relatively insignificant.

This was echoed by sentiments attributed to Bruce Perens last year upon his resignation from the OSI:

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fnews.slashdot.org%2Fstor...

"We've gone the wrong way with licensing," he said, citing the proliferation of software licenses. He believes just three are necessary, AGPLv3, the LGPLv3, and Apache v2.

Whether or not one agrees with the choice of the licenses, the original impetus behind the DFSG which became the Open Source Definition was to be able to ship a CD full of free software that could be dynamically linked in a Linux distribution. Debian was the first distribution to ship a wide range of software instead of just a core UNIX-like system and fortunately, Bruce and the others who helped write the DFSG rose to the occasion.

The way I remember it, the Open Source Initiative was created as a way for us to be able to us Linux and other DFSG-free software at work. Microsoft had been telling people that Linux is a cancer ( https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.theregister.com%2F20... ), implying that running anything on Linux would render it subject to the GPL.

The Open Source name also helped frame the market for companies like Red Hat with Enterprise Linux offerings. UNIX also had source code available and was licensed with support contracts available for business use, so it was a known concept for the businesses that would be using it. By contrast, the FSF website would probably read similarly to the timecube website in a corporate boardroom.

I have been using Linux at work as a top tier operating system running business-critical systems for a couple of decades now, so it looks like the Open Source campaign worked.

That's before you add in support revenues for the render farms for these shops whose server OS vendor would become obvious. :-)

I have a number of systems that are still running on btrfs and can't be easily migrated, and it contributes to a feeling of dread. The reason is that btrfs-balance, which is included the btrfs-maintenance service and enabled by default by the distro, has a tendency to destroy the filesystem.

Disabling that job has stopped the majority of the btrfs filesystem loss on my systems, but then I have to wonder if there is something btrfs-balance would have been doing that now isn't being done because running it corrupts the filesystem.

The only special feature I have intentionally used in btrfs is the snapshots. Fortunately, I have only ever used it for root filesystems, so none of the other features have really come into play. Filesystem checksumming seems to generate a lot of buzz, but in 10+ years of running hundreds of systems on ZFS and BTRFS, I have never seen a file checksum error reported.

The idea of btrfs watching file checksums like a hawk on an enterprise all-flash array or modern hyperconverged system brings this to mind: "Why do you look at the speck that is in your brother’s eye, but do not notice the log that is in your own eye? Or how can you say to your brother, ‘Let me take the speck out of your eye,’ and behold, the log is in your own eye?"

Personally, I would rather have a reliable filesystem like XFS with the extended features being snapshotting with previous views accessible under /.snapshot and online resize, including shrink. Ideally, it wouldn't require a host of routine maintenance tasks to keep it going, such as how btrfs is supposed to have scrub, trim and balance jobs run regularly. Also, it wouldn't block IO to the filesystem for 30 seconds or so while deleting a number of snapshots. I would rather keep volume management of it - LVM does a nice enough job of being a volume manager for those cases where a filesystem needs to be striped across devices or mirrored off from one storage array to another for migrations.

My most recent yearly payment went through in April and was refunded in its entirety, so I got four months free.

Not really, but I did just notice that Accuweather has a text widget. I saw the widget showing "Rain may form in the next 47 minutes", but opening the app, it reported that rain may form in the next 11 minutes. Not sure why they were that far apart.

Darksky has a meter that catches the eye when looking down at the phone and audible alerts when it's going to rain. That's what really makes the app for me - alerts and indications of rain I didn't know to expect. I am in the market for an app with that functionality. I have only had the other apps for a matter of hours, but the only one I noticed sending an audible alert for rain was Shadow Weather.

The special thing about Darksky for my use was that it would play a chime when rain was imminent in an exact location and had a widget with a bar chart meter showing how heavy the rain would be over the upcoming hour or so in fine-grained increments. It was great for getting an idea at a glance of whether it would be worth it to wait a few minutes for the rain to let up before going outside and when to check if the car windows were rolled up.

So far, I have found Accuweather with a similar 2 hour rain meter, but no widget for it, and Shadow Weather sent a notification about rain and has a "Next hour" rain forecast with a meter that can be opened by clicking on the Next hour text, but doesn't seem to have a widget for the meter, either.

Hopefully it's only a matter of time before one of the other apps adds a home screen widget to fill the void.

> Your reply was all about this from the point of view of Linux, making it the centre of everything.

It was more from the point of view of a spurned Solaris admin who lost some of his favorite tools after Sun made licensing and process decisions that precluded widespread adoption on the platforms now used by business. The point I was intending make is that had CDDL not been an effective poison pill to prevent Linux cherry picking from OpenSolaris, I expect that Sun would have found other means.

I consider the module situation a trade off - on one hand, vendors supplying binary modules that have to be relinked each time a kernel is updated is a headache. On the other hand, it encourages vendors like Microsoft and VMware to finally contribute hv_netvsc and vmxnet3 source to the kernel rather than only releasing binary modules. Vendors never seem to keep those updated on a timeline that matches when I want to deploy a new release of an OS.

Maybe one day Oracle will be able find it in its heart to license ZFS as GPL like how they did with ocfs2 :-)

Linux was already GPL both in name and in spirit when ZFS was released under CDDL. If Sun had wanted it to be used in Linux, they could have dual licensed it like perl with Artistic/GPL. One of the reasons given for CDDL instead of GPL was reportedly to play keep-away from GNU/Linux so OpenSolaris source couldn't be meaningfully reused there.

https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fen.wikipedia.org%2Fwiki%2F...

"According to Danese Cooper [ the one who actually wrote the CDDL ] one of the reasons for basing the CDDL on the Mozilla license was that the Mozilla license is GPL-incompatible. Cooper stated, at the 6th annual Debian conference, that the engineers who had written the Solaris kernel requested that the license of OpenSolaris be GPL-incompatible.

"Mozilla was selected partially because it is GPL incompatible. That was part of the design when they released OpenSolaris. ... the engineers who wrote Solaris ... had some biases about how it should be released, and you have to respect that. "

Fortunately for those who do want to use ZFS in an integrated system, FreeBSD has done the work and accepted the licensing ramifications.

As Linus pointed out, Oracle Corporation has the option of GPL licensing the ZFS code. I haven't seen any reports of Linus threatening to go after e.g., Ubuntu for GPL-related issues. Meanwhile, an $8.8 billion case involving Oracle and open source copyright issues is waiting to be heard by the Supreme Court...

On the other hand, after Linus' clear comments of "don't use it" alongside dismissing its overall value, it would appear difficult to make the case that he was contributing in any way to any type of infringement. I much prefer for Linus to err on the side of not overvaluing the worth of something with a dodgy license. Rather than criticizing Linus, perhaps that energy would be better spent lobbying for a usable license for ZFS.

For the typical enterprise use case, ZFS was nice on Solaris to avoid paying Veritas licensing for software RAID and online resizing. On Linux, LVM + XFS has been good enough for that. As much as I liked LiveUpgrade + ZFS on Solaris, I have made peace with what Oracle assuming ownership means for those technologies.

In my experience, btrfs is has been more of an attractive nuisance than anything. While SUSE will field support cases about it in SLES, the level of support is along the lines of "try btrfs.check --repair", and when that destroys the filesystem entirely, to recommend restoring from backup.

ZFS on Solaris was a solid production-grade filesystem. btrfs sounds like it offers similar functionality on paper, but has not been production quality on SLES 15. File checksumming is nice, but it's hard to really get excited about when btrfs balance regularly corrupts filesystems.

The irony of ironies would be if they brought it back and by virtue of not having systemd, it shot to the lead in the Enterprise Linux market.

No matter how well coal plants operate their scrubbers, they leave huge amounts of toxic coal ash that have a habit of ending up in our rivers and groundwater.

From https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fwww.cbsnews.com%2Fnews%2Fd...

"Environmental groups have hailed the various charges against the company as vindication for their years of efforts to get regulators to hold Duke accountable for the pollution leaking from 32 coal ash dumps at 14 power plants scattered across the state. The ash, which is the waste left behind when coal is burned to generate electricity, contains toxic heavy metals."