PC-BSD Evolves into TrueOS | Hacker News Reader

PC-BSD Evolves into TrueOS

trueos.org

164 pointsjlgaddis87 comments

The same excellent team at http://ixsystems.com is behind this. I've not only had a breeze integrating TrueNAS gear into a VMWare environment but I have personally had a FreeNAS mini for years at home and recommend it for many SOHO users who want local storage.

Hire a designer to make Lumina less ugly. Sorry I know that sounds harsh, but it looks very 2005.

Make this install on modern laptops. I'm tired of being recommended five year-old Thinkpads.

Fix those and I would consider overwriting my Ubuntu installation

> it looks very 2005

I have literally no knowledge or opinion about Limina, but kinda with desktop UIs would attempt to be timeless rather than trying to follow the latest style fads.

> timeless

What does that even mean in this context.

IMO the Xerox Star UI and NeXTSTEP/GNUstep:

https://www.youtube.com/watch?v=Cn4vC80Pv6Q https://www.youtube.com/watch?v=uXHFfc3btCI

You can run a variant of GNUStep today, yet no one does.

Before switching to Lubuntu a few years ago, Window Maker was all I ran. LXDE is unobtrusive enough that I don't mind it.

As a registered NS developer, I had a NS Turbo Color. What about NS do you find superior to OS X?

Not OP but also a NS developer (custom built Pentium), and I must say like the menu system, digital librarian, workspace manager, and eof much better. A separate interface builder was much better, but that's more a developer tool thing.

[edit: left side scroll bar also]

For sure I liked the separate IB/PB. Good call! It was faster to develop in NS than in anything else for me.

At the time, I loved the digital librarian and menu system. After leaving and going back a few years later the WS seemed a little too dated for me. I use the context menus in BlackBox on my linux box. OS X seems to be my comfort zone now, which tends to color memory. Though I also found the IB/PB split much faster to work with. That's for sure.

You've got me fired up. I'm installing 4.2 ent in virtualbox now, which will be the newest version I've used :) Thank you!

I still have my 3.3 discs and the 4.x discs with the wall of documentation. I'm tempted to install it again, if for nothing else to get at some very old e-mail I sent.

I think some of the "dated" is that it wasn't very high resolution, or put another way, it really didn't get to evolve into the high resolution / anti-aliased world we live in today. I've seen multiple attempts at redoing the icons in a modern setting and many of them look great. I particularly like the old folders. The menus suffer from the same problem, but more from a modern font choice scenario.

For me, OS X stripped a lot of the NeXT goodness to appease Mac lovers. I'm just not a fan of a UI designed for a 9" screen that really should have a better menu system. Recently, I've seen some 3rd party software that simulates part of the old NeXTSTEP menu system. Sadly, it doesn't seem to do the tear-offs.

I guess OS X is loads better than Windows, but it sure isn't my favorite. My two big "not working" projects are an agent oriented programing language and some form of modernized NeXTSTEP UI on OpenBSD. Sadly, more likely to do the former rather than the later.

I installed it last night, making sure to get the patches in order to use VESA not just SVGA. That gave me color. NS sure is beautiful. My color slab wasn't this good looking, since the slab had 16 bits of color.

The dated part for me is having to setup the OS for some workflows that have come about since NS days. Some things for scroll wheels not getting recognized out of the box and other polish. WS is great, though I'd change the Drag and Drop just a little. My memory of WS must have tarnished or faded some over the years, because it was and is really easy to work with.

I'll likely install the developer cd this weekend.

Thanks, again, for getting me to reinstall.

On a side note, there are commercial projects that use GNUstep. I use Hopper, a disassembler, that uses GNUstep.

From the developer training I still remember one statement in particular, "Even if NS isn't around in a few years, there will be a company that deeply understands what we do and they will do it." I guess that prophetically was OS X (at least in part).

What guide did you use to do the install?

This one looks pretty good. http://openstep.bfx.re/

I haven't been able to figure out the networking. I have the ne2k driver, but it doesn't seem to install. I suppose the copy I have is somehow corrupted.

Less chrome, no shadows, no rounded corners. Scroll bars on the left side. Two button mouse (not three, but good enough).

That's not timeless, that's very very timed.

Motif and CDE!

> What does that even mean in this context.

IMHO: come up with a set of initial will-thought-out UI design choices that follow an internal logic, then evolve thoughtfully (and conservatively) in a way that respects those choices.

That has the added bonus of not constantly invalidating people's investment in learning the desktop environment.

I would expect a environment that follows that would "look" similar in 2005 and 2015, and that wouldn't be a bad thing.

> it looks very 2005

I want things to look very 2005 (or even 2001), which is why I use XFCE4 as a desktop environment on xorg / debian, and not a modern "fancy" thing like the latest gnome or KDE.

Cheers! I liked gnome 2, I can't stand unity, gnome 3 or the latest KDE. LXDE was too minimalist. XFCE was just right.

I liked GNOME2 also, and I found MATE to be the perfect recreation of it.

Xfce was just perfect. Sadly, xfce 4.6 unintroduced manual context-menu in favor of autogenerated one (from cryptic sources). It also introduced many good, but menu changes broke my workflow. I was so disappointed that I left entire Linux/FreeBSD after years of using it. Now I have mac, and it is just great for years.

I guess the question is whether this thing that follows 2005's style is timeless... I suppose from the reaction of people that it doesn't?

Though the only interface that I can qualify as timeless is maybe the PS2/3/4 interface XMB (10+ years and still basically the same).

I think as far as desktops environments go, I imagine the first goal is to make it functional, and only then make it pretty.

Hopefully now after recently hitting 1.0, someone spends some time making it a bit nicer looking (nicer being relative of course -- I tend to like simpler desktop designs).

> Hire a designer to make Lumina less ugly. Sorry I know that sounds harsh, but it looks very 2005.

I'm not trying to sound harsh and just have to respectfully disagree.

I don't use Lumina because I don't use DEs, but the looks of it are just right. It's not a regression like most "styles" these days where it's impossible to know half the time what's an interactive widget, or why suddenly the designers think large empty white spaces and large widgets and titlebars are a feature. If I'm at a desktop, I need a desktop environment, not a tablet experience.

I don't want it to look like modern stuff, but it could use some tweaking and usability consistency improvements here and there. It's the whole 'papercuts' thing, nothing terrible, just a bunch of small issues (font sizes, click target issues, etc.)

Polish is probably what you're missing then.

My first language is English, and from the screenshots in the TrueOS handbook I'd also agree that it looks a little dated (although not particularly more so than most other open source desktop environments).

>> Polish is probably what you're missing then.

> My first language is English,

Did you just mix up polish, as in "polished ui", with Polish, the language of Poland?

Yes.

Hopefully he doesn't also think that a totem pole is a guy from Poland who looks after a totem. :-)

Also, we all need to be using polish notation https://en.wikipedia.org/wiki/Polish_notation for better good

> I don't use Lumina because I don't use DEs [...]

In fairness, that probably means your preferences will be very different from those people who actually do use desktop environments.

I also think a lot of people (myself included) have a certain nostalgia for older, leaner desktop environments. That doesn't mean they can't be designed better, with more polish, using those same older design cues.

I failed to see how control panel was any better in each 95/xp/7/10, except that my enemies can't find anything in it just like me. I also fail to see how w7 explorer became more 'convenient' without up button (and changing backspace key to back) but with annoying Libraries that are not even folders. Or removal of quick launch, so that icons on our taskbars now float and you search for a button to start that app.

What you call polishing is just breaking familiar old-simple and building new-super-innovated-final-solution each year. Leave us power users alone, you hipsters!

> What you call polishing is just breaking familiar old-simple and building new-super-innovated-final-solution each year. Leave us power users alone, you hipsters!

I wish it was only confusing to power users. If it was actually intuitive to end users, it would have a justification. But no. I've yet to see a single user who can wrap their head around Windows 7's virtual libraries and why they're sometimes not acting like a real folder but most of the time do.

Well, I do use a DE (MATE), and I think the appearance of Lumina DE is just fine. What kind of "polish" do you find missing?

The linux drm compat shims should go a long way toward modern laptop support.

> Make this install on modern laptops. I'm tired of being recommended five year-old Thinkpads.

This branch does carry the (work-in-progress) i915 driver updates that bring support for modern Intel graphics (Haswell, Broadwell, Skylake).

If anyone else was concerned that they are claiming this without actually being affiliated with PC-BSD, check out http://pcbsd.org/development/moretrueos/.

Yeah, there's a post on the PC-BSD blog [0] also, but it just links to the submitted URL.

[0]: https://blog.pcbsd.org/2016/09/pc-bsd-evolves-into-trueos/

> TrueOS® soars above the competition with advanced security features, such as PersonaCrypt and GELI disk level encryption, to keep your important data safe and secure.

Has it got kernel/userland ASLR yet?

Does their ZFS implementation supports dataset level encryption?

Both ZFS and ASLR are more interesting for servers than for desktops. I would like ASLR on my desktop OS, but it's a rather high hanging mediocre fruit. WRT to ZFS, I couldn't care less if it existed on my desktop OS.

> Both ZFS and ASLR are more interesting for servers than for desktops.

On what basis are these claims made? I'm tired of hearing this excuse trotted out every time someone posts about ZFS or ASLR.

ZFS (or a filesystem possessing many of the same characteristics) is absolutely still valuable even on desktops. With consumer storage devices reaching ever loftier maximum capacities, data integrity becomes even more important as the loss of that data can have a profound impact on a person's life.

Given the questionable quality of hardware in many consumer devices, having a filesystem that detects corruption as it is happening, and works to ensure that it is corrected automatically when possible, seems like an obvious win.

While a typical consumer is unlikely to have a ZFS raid device even if they had ZFS, they would still benefit from all of the other protections it offers. It should be fairly telling that Apple is adopting many of the same technologies ZFS offers in their own consumer filesystem (except, regrettably, checksumming) because they too understand that it does matter.

As for ASLR, that's important too because it's yet another mitigation tool to make it less likely that an individual system is compromised. Haven't the past few years of various malware/ransomware announcements been enough? We should consider every practical tool available at our disposal to mitigate such attacks as valuable.

> ... I couldn't care less if it existed on my desktop OS.

Then you either don't know about snapshots, don't care about data integrity, or have a backup system in place and don't want to replace it with:

  zfs send zroot@snap63 | gzip > /mnt/remotebackup/snap63.gz

I remember just before Oracle took over Sun, OpenSolaris was a pretty viable desktop machine. They had a feature called "Time Slider" that was built into Nautilus. It was like Apple's Time Machine, only based on ZFS snapshots instead of backups to a separate disk. You could slide a widget back to get to a previous snapshot of the disk. OS upgrades also worked together with ZFS snapshots so that you could atomically roll back to a previous version just before the upgrade took place. It was all really awesome.

Yeah, out of the box FreeBSD just barely scratches the surface when it comes to ZFS - there is a lot of untapped potential. Same story with jails and vnet.

PC-BSD ... sorry ... TrueOS Desktop (and Server) has had that idea of ZFS snapshots for "boot environments" for some time.

* https://blog.pcbsd.org/2012/07/9-1-feature-multiple-boot-env...

* https://blog.pcbsd.org/2013/06/pc-bsd-status-update/

* https://www.ixsystems.com/blog/a-closer-look-at-the-changes-...

* https://www.ixsystems.com/blog/the-revamped-life-preserver/

Beadm first appeared in Solaris, long before it hit FreeBSD (and then PC-BSD) - so the idea had been around for a while. The issue is fit and finish. Solaris had it tightly integrated and made full use of it. I haven't used PC-BSD for a very long time, but I don't remember it coming anywhere close to Solaris in that regard.

Is there a way to download the ISO hashes over valid HTTPS? I'm not sure how to verify the ISOs I've downloaded

FWIW, this reasonable request can be accommodated without necessarily setting up HTTPS on the download site. The devs could post the checksums on a distinct, 2FA-protected HTTPS site like Twitter, for example [1]

[1] https://twitter.com/coreinsiderprog/status/69213854723989504...

Wouldn't it be easier to just setup an automated Let's Encrypt set up and avoid depending on a third party website?

In any case, they could also just include the keys somewhere in trueos.org, which does support HTTPS.

Reading behind the link made me learn about Lumina desktop. I'm curious to why it is not making use of the LxQt effort? They seem to be pretty close in goals...

If I'm not mistaken one of the big goals of Lumina is to produce a desktop with its roots in BSD, since most (if not all) DEs currently in development got their start on Linux. I imagine the idea is to develop Lumina as part of a single-piece OS, BSD-style, as opposed to the Linux model where everything is developed independently and then glued together to form a distro.

It still feels like it'd make more sense to at least fork something existing (or better yet, find a desktop that doesn't mind BSD-specific conditional code inclusion). I mean, the one-base-system mentality doesn't preclude the BSDs from using gcc and/or Clang - even though they weren't developed for BSD specifically, they are clearly the best available choice.

I wonder why not Xfce, specifically. Of all the DEs out there, it seems most complete in a sense of being an actual DE, with the usual configurability and stock apps that one would expect (a great file manager, for one). But at the same time, it's not the mammoth that Gnome and KDE are, gets out of the way easily enough, and doesn't demand that applications are aware of its existence beyond following XDG specs.

Not sure, but i think XFCE relies on a fair bit of Gnome plumbing, and that pluming in turn is becoming increasingly systemd-centric. Not even sure how well GTK3 plays with BSDs, and XFCE will move to using GTK3 "soon".

I don't think that's the case, at least not anymore.

I'm running Xfce on two systems, one of which is Gentoo (in the default OpenRC configuration, no systemd), and the other is FreeBSD. Neither has Gnome installed - just recently I had to actually install gconf to do something, as a matter of fact, and was surprised that it wasn't there - but they handle Xfce just fine.

I believe it used Gnome settings daemon at some point, but it has its own these days.

>I imagine the idea is to develop Lumina as part of a single-piece OS

That makes no sense to me, given that it runs on Linux as well as BSD.

From what I've read it relies on Qt, Fluxbox, XScreensaver etc, which are hardly BSD based components.

>as opposed to the Linux model where everything is developed independently and then glued together to form a distro.

What's the difference here ?

Lumina is developed independently, the components Lumina builds upon are developed independently of Lumina, it is then glued together with PC-BSD into TrueOS.

No, Nothing about Lumina ties it to BSD. On the contrary; it's fundamentally easily portable to any OS. It runs just fine on Arch linux. The "idea" was to generate a basic DE that is not mired into any lame dependencies such as systemd.

>I'm curious to why it is not making use of the LxQt effort?

I would guess the primary reason is licensing. The other is a motivation to integrate with FreeBD/TrueOS specific things like ZFS and their administration tools.

Development on LXQt's been rough lately. I really wish the Lumina guys would have contacted us at some point, reached out to us.

There is a very big lack of manpower, lack of designers, lack of general availability from sysadmins or developers competent at UX.

It saddens me greatly. I used to take care of releases, outreach and UX but focusing on my company means I don't have the time to do that anymore. Despite taking great care to reduce the bus factor as much as possible, right now, LXQt is stuck without a release because there's nobody that both cares about the project and has the time and skills to do this kind of stuff.

This is how the Linux desktop dies. A couple of "big" players doing their own thing and a plethora of small players never trying to partner up with similar projects. LXQt was successful for a long time because it was two small projects that merged into a larger one, but people move on and the momentum gets lost.

As far as I know Lumina is a one man project. I remember listening to a podcast with him. It sounded very much like a guy started messing around and trying stuff out and ended up with Lumina. Kinda like Hammer2 is also mostly a one man project. That's why it's pretty slow moving.

Sorry about LxQt. Wish I had some Qt experience to help out

> A couple of "big" players doing their own thing

So much for Freedesktop standardization...

FDO is a failed project. It got us a few good things, but now it's a joke. Nobody cares about or reads the mailing list, nobody takes the various requests in it seriously, blah... I'm pretty upset about it.

The problem with GNU/Linux is that they just replaced the UNIX wars with Linux wars.

For a long time, starting with Slackware 2.0 and for around 10 years, I wanted to use fully use GNU/Linux and used most of the distributions until eventually settling on Ubuntu. Which I only use on a travel netbook anyway.

As someone that is focused on UX and the whole desktop stack, never found a UNIX that cared about the user desktop experience.

Compared to Atari, Amiga, BeOS or even Windows.

NextSTEP was probably the closest to it, but even its UNIX underpinnings were more a side effect to speed up its development and make it easy to port software than anything else.

So my daily OSes have become Windows, OS X and Android. The fact that Apple and Google have a UNIX kernel is totally irrelevant for the programming languages I make use of.

Just like LXDE and Razor-Qt merged, wouldn't it be a good idea for LXQt to merge with Lumina?

It would not only ensure that LXQt's torch is carried on but also set a great example to all other open source projects that merging provides as great a solution as forking does, although for opposite problems.

Sure would be, unfortunately nobody's talking or available to actually do it.

The Razor-Qt merge wasn't an easy project - it took us almost 6 months to complete.

Licensing issues make this unlikely, a big reason Lumina was developed is to be able to ship a BSD licensed OS with a GPL free desktop environment.

Previous discussion: https://news.ycombinator.com/item?id=12366140 (6 days ago, 30 comments)

Maybe I'm getting old and my eyes are on the way out, but the readability of this web page can be greatly improved by using the following custom css:

article { font-size: 1.4em; line-height: 1.4em; }

It's not just you: http://imgur.com/a/CfFaS. Fails WCAG AA accessibility for contrast. (even at your increased font size, I think)

God save us from the insane low-contrast mafia. No, that's not intended as sarcasm.

I really despise this trend. People who make such webpages completely disregard the existence of users with bad eyesight and those who still use low contrast TFT displays mostly because they can't afford to buy a fancy new IPS panel or just don't know any better. We all should complain about it, otherwise soon it might get to the point where you only will be able to read some webpages on the newest 40 inch OLED displays.

Having not followed the development of PC-BSD for the last 5-6 years, can somebody please explain to my why they felt the need to fork FreeBSD? FreeBSD by itself already has a really tiny market share, and to a certain extent many things are much more easily deployed with better package management on a mainstream Linux platform (Debian stable or CentOS 7).

OpenBSD has its own unique niche, but I don't see what great need PCBSD fills that FreeBSD wasn't already doing.

Effectively, TrueOS _is_ FreeBSD, plus the work in progress graphics stack that will eventually land in .org base or ports, plus an installer, plus some opinionated settings (kernel build, /etc, /boot/loader.conf), plus some original software (control panels, lumina DE).

This is a valid way to contribute to *NIX -- downstreams that push and pull code to evolve the entire ecosystem.

It's less a fork and more of an add-on, PC-BSD (now True-OS) is FreeBSD with a fancy installer and a preconfigured desktop. TrueOS server is FreeBSD with their admin tools installed. Kris Moore, the creator, is a member of the FreeBSD core team, the projects are close.

So, is this a fork, or...?

I know they're affiliated with PC-BSD, and I know that PC-BSD is a separate system, but beyond that, I'm lost.

It looks like it may be an RHEL/Fedora thing, with PC-BSD being stable and rock-solid, and TrueOS being bleeding edge. But I could be wrong.

Also, while Lumina looks kind of nice, if I installed this, my next step would be installing either XFCE (the best full-on DE out there) or i3 (the best tiled WM out there).

It looks like it's the successor to PC-BSD. It's from iXsystems, the developers of PC-BSD.

PC-BSD is no longer a separate system. PC-BSD was merged into TrueOS, probably to save some maintenance powers and save on trademarks. PC-bsd blogs tells you to go to TrueOS page:

https://web.pcbsd.org/development/moretrueos/

> PC-BSD being stable and rock-solid, and TrueOS being bleeding edge. But I could be wrong.

TrueOS was basically the same as PC-BSD but for servers.

PC-BSD to TrueOS is like BackTrack Linux to Kali Linux.

An evolved product from the same developers.

PC-BSD

It's a rebranding, plain and simple.

How does this handle turning off Intel ME or the AMD equivalent ? Having a always on PC on my motherboard not under my control does bother me ...

That has nothing to do with the OS. The Intel ME is nothing more than some tech which can be leveraged/controlled by the BIOS/EFI - or not. There should always be a setting to just turn it off.

The hysteria over Intel ME amuses me. The clamor should be for a simple on/off switch to always be present in the BIOS/EFI.

But A) can you trust the switch really turns it off

B) there is not yet a switch ... So what do we do now while we wait?

Great decision unifying PC-BSD under TrueOS.

The real innovation lies in SysAdm:

A new way to manage your Server, Desktop or Cloud-based system. By exposing an API via encrypted REST or WebSockets, it is now possible to remotely control all aspects of your machine, including management of software, updates, boot environments, users, backups, and more.

Truly a flagship FreeBSD "distro".

I'd love to learn more! <Clicks "Learn More" button on main page> "Localhost not found". No, not on an iPhone it won't be. EDIT: ah, good, looks like it got fixed.

Reply on news.ycombinator.com