Reg FOSS desk test drive: First beta of Fedora 38 drops The Fedora project has released the beta version of Fedora 38. The Reg took it for a spin, and it handles well. The beta version of Fedora 38 is here, hitting the "early target" in its schedule. As usual, there's an extensive changeset, which is worth scrutinizing for the full details of all the various subcomponents. We tried …
To be honest, I don't care a fig about Redhat nor anything they produce, and here's why:
In '96 I found myself in a far-off land and, trying to set up a workable PC for myself, all I could get hold off was a Rehat 4.8 CD (you needed an intenet connection before you can download anything from sunsite). I took me three effing days to sort out what the heck had been installed (when you're used to Slackware, which leaves you perfectly in control, you have every reason to be peeved.)
Having returned to the Sceptered Isle, I had no such problems, and managed to create (WABI had by then become quite affordable) setups for others, knowing that I'd never receive a support call. Linux was by then stable, libc5 had been thoroughly debugged, XFree86 no longer spontaneously froze, SANE was pretty good, Hylafax worked flawlessly, and Ghostscript had vastly improved. And when it comes to stability, Linux ran rings around NT4. Life was good.
What happened next? Redhat 5.0: five years of hard-earned reputation blown to dust. Rickety as hell.
Than came Drepper. Torvalds futzing around with virtual memory management. The fileystem wars. And, if that's not bad enough, Poettershite - did anyone of us ask for systemd? Did anyone of us ever need this?
There once was a time when Linux provided a degree of freedom - but with the preponderence of Shithead and Ucuntu - Debian being the eternal retards (openssl anyone?) I've called it quits. (My private and professional obligations are such that I no longer have the time to fix others' code.)
Yet you took the time to come here and post...?
If you are an IT professional, you should know that you should always tread carefully when moving to a .0 release. I had the same with VMS 4.7/5.0. 5.1 resolved most of the issues in 5.0.
These days, and IMHO, Linux is a lot more stable than the POS that comes out of Redmond/Bangalore. I use AlmaLinux which is built from the RHEL Sources. Apart from the disaster that Gnome has become (easily rectified) it does what it says on the tin. For most people, what goes on with systemd is irrelevant as it should be.
As for fixing other people's code? I don't do that but I do chip in a few £££ every so often so that others can do it for me.
Yes, I can code. I spent most of my 45 years in IT coding. These days, I find that I am writing shell scripts more than coding although some would contend that shell scripts are code.
Thank you for your considerate reply.
Hope springs eternal, and maybe we can get Linux back on track. Gnome, as you rightly state, is a disaster (again, did anyone ask for this? We've had better.).
What has changed since them there days is that, starting with Yggdrasil, Linux had become user-friendly and also encouraged users to learn something and turn it into their own.
This has completely changed by now (ever had the displeasure of having to work with AWS?)
When it comes to Linux proper (i.e. the kernel) certain things have definitely improved (e.g. SMP which was less of an issue in the 90s), but in most other respects I can no longer recommend it, because what Linux once gave you was the freedom not to be chained to a particular supplier. Since the packagers have begun to run the market, it no longer makes a bleedin' bit of a diffenrence wether you go for Microsnot or what not (there's commoditizatin for you). At least with AIX I've still got something resembling a roadmap, Snoracle being a price-gauging nuisance, and HPox no longer worth talking about. (BTW, I don't work for IBM.)
So what's left? On the Linux side I'd say Slackware - but the guy needs all the help he can get - apart from that there's FreeBSD and OpenBSD, both suffering from an extreme lack of device drivers, plus some objecting to their license models (isn't life strange?).
Quote:-
Hope springs eternal, and maybe we can get Linux back on track
Which given your first post and the name calling, seems unlikely. You can't get it back on track from the outside.
I started with Unix... ok, BSD Unix and Ultrix. Then one day, there was Slackware 1.1 given away with a computer mag. Not looked back since.
As for 'packagers' the people behind SNAP need to be lined up on the top of a cliff and pushed over the edge one by one. IMHO, Canonical have totally lost the plot in their chase to be bought out by Microsoft.
"Which given your first post and the name calling, seems unlikely. You can't get it back on track from the outside."
From a (current) Linux outsider (by choice, my personal Linux experience mirrors Golgafrinch's almost exactly), as I read both reviews and comments from users, I am finding it interesting how changes in Linux (apart from anything introduced by Poettering) is taken with far less grief than any changes made to Windows. MS changes a [power] button: the end of the world as we know it. Linux changes a windowing paradigm: well, it's here, oh well. It's an interesting dichotomy to watch, as Linux users seem oblivious to their own conflicted standards regarding this but often suggest changing distros to solve this issue, which only serves to continue to problem of Linux's fractured distro landscape which impedes greater desktop adoption.
@Snake:
(Fascinating animals, by the way)
Anyone who quotes both R.E.M and the original Fleetwood Mac (as in "I ain't pretty and my legs are thin", which applies to me - oh well!) must be slightly bemused, though not amused, by all the developments (or frankly put, regressions) that have occured in the last 25 years.)
"I am finding it interesting how changes in Linux [...] is taken with far less grief than any changes made to Windows."
A couple of possible reasons
People who decide install Linux tend to be more confident with this IT malarkey. My understanding is that most Windows users basically use what the computer came with
Linux distributions provide a much wider range of choice. Don't like Gnome Shell? Use xfce instead. Prefer old-school init and admin? Slackware is still going strong. You get the drift.
Back on topic: had a play with the rawhide live iso last evening on my test bed Thinkpad L440
Anyone using smbus based trackpad who notices a frozen pointer after waking up from sleep with recent Fedora versions (and Debian - it is kernel related), have a look at...
https://discussion.fedoraproject.org/t/touchpad-dont-work-after-suspended/73733/22
and
https://askubuntu.com/questions/1097080/ubuntu-18-04-mouse-on-lenovo-thinkpad-x240-not-working-after-suspend-hibernate
Linux distributions provide a much wider range of choice. Don't like Gnome Shell? Use xfce instead. Prefer old-school init and admin? Slackware is still going strong. You get the drift.
There's some truth to that, but it's less true than it was historically. To my eyes Linux peaked around or shortly before the millennium, say the time of the 2.0 and 2.2 kernels. You had a decent traditional Unix system, slap a copy of Motif on there (proprietary at the time, LessTif existed but was buggy) and you had a decent enough system, a few quirks and limitations but nothing's perfect.
Then came the desktops, and other large scale projects counter to the Unix traditions such as CUPS - SystemD is just the latest in a long line. At first KDE was a library and few applications, Gnome had ambitions but wasn't ready for primetime. Slowly they became integrated wholes and you lost the ability to select components on a piece by piece basis that had always been the strength of Unix. For example, what if KDE has that really nice applet to configure your WiFi, but you ditch KDE in favour of something else. Sorry can't use that now...
The ultimate evidence of this is these very reviews of each latest distro. Rarely do they discuss anything of substance, mainly it's the interface and eye candy. Sure you can change it, that doesn't mean it's easy. If it was why are there so many advocates for one distro over another?
There's some truth to that, but it's less true than it was historically. To my eyes Linux peaked around or shortly before the millennium, say the time of the 2.0 and 2.2 kernels. You had a decent traditional Unix system, slap a copy of Motif on there (proprietary at the time, LessTif existed but was buggy) and you had a decent enough system, a few quirks and limitations but nothing's perfect.
Yup. Those were the days that you mainly installed X because you could get more terminal sessions on the same screen, and a copy of a web browser (spyglass, I think) to poke at this newfangled URL/web thing. And you could have fresh PFYs manually editing sendmail.cf for days before telling them about m4 :).
TBH, I have found that all desktops are quite inconsistent in how they're configured, it's never one system like textfiles or one config GUI that adapts as applications are added and God help you if you try to figure out where the configs are and how they work (instructions tend to be written for people who code for a living, and not all of us do), and there has been as yet no workable replacement for the one thing that MS has managed to integrate reasonably, Outlook (which, BTW, happens to be the major hook by which people are kept using Windows). I had my high hopes of Vivaldi dashed by their frankly baffling and inexplicable omission of carddav capability which rendered the email client basically unusable. But I digress.
I know that some people swear by GNOME or KDE, but I found in both cases that I more swear AT them, so the next time I spin up a Linux box (mine are packed up as I'm moving) I think I'll stay with xfce. The problem is also that all these desktops are designed by people who are technical, not artistic. No, let me correct that, they are DEFINITELY not artistic. So you can't sell either desktop to non-Linux users either. Enlightenment tried to do better, but has little support so it's still sort of bumbling along in the margins of distros alongside other experiments and is not going really anywhere (although I think that KDE and GNOME picked up ideas from there).
And then we have the other aspects: server functionality versus desktop. The main reason why I use Linux is because I basically have an Internet in a box (well, multiple boxes), so not only do I have a desktop but I can also spin up some services if I want to experiment with something like VLAN numbering or cachcing in Apache - anything. Sometimes just because I can or want to see how something works. For that I find the UIs also, umm, wanting, but maybe that's me expecting to much. I find myself still preferring to into a terminal for that, even with a text GUI as you can find in, for instance, SuSE's yast so I don't have to set up every bloody file myself but it's just faster.
Anyway, this is getting too long (and late) for basically saying that I agree :)
> Linux changes a windowing paradigm: well, it's here, oh well.
I suspect it's more like "well the new paradigm is bollocks, thankfully I can [do|install|configure|etc] something else".
That freedom of selection found in Linux (and BSD) is not typically found in Windows.
Conversely, when some paradigm is shoehorned into Linux which does NOT have choice, at least not simple/straightforward ones ("go fork it yourself" doesn't count), perhaps even actively seeks to create lock-in and prevent other options, the hue and cry is pretty noticeable.
Hence the backlash to things like systemD, Snap, etc.
Agreed on almost all, except "what goes on with systemd is irrelevant" -- IME it is definitely *not* (would that it were), since it has actively broken some things we relied on, requiring extensive re-work and crafting new unit files and similar hacks e.g. to workaround systemD's inability to figure out "is the network up", among other things.
systemD continuing to snaffle more of the operating system (creeping featureism) seemingly at each new release is an ongoing concern. We should be so lucky that systemD were truly "irrelevant".
To be fair, working out whether the network is "up" requires a precise definition of what "up" means. Given the sheer number of networking topologies that exist, and the fact that the 7-layer model is no more than a figment of the imagination, there is no precise definition of "up". macOS and Windows fare no better in this respect.
Regardless, it wasn't a problem pre- systemD.
Probably even worse, systemD's stance on addressing the situation apparently boils down to writing a not-that-helpful wiki page about the topic, and blaming other developers, system services, users, etc. for "doing it wrong".
Redhat 5.0 was a bit mmmmrrrrr. RH5.2 was very good, stable, reliable. The same with RH6.0 vs RH6.2.
The thing is... Linux (I use the term in blanket fashion, NOT just the kernel but Xorg, desktop environments, printing from desktops, server applications, etc) is in the hands of the developers. It so happens that Redhat (IBM), Ubuntu, etc, have the resources to do the coding. They have the ability to do this day after day, which is unlike a lot of unpaid open source contributors. If Bob is ill or is fed up doing $project or leaves the company, they can put Bill on the job. In short, it is about delivery.
I have no liking at all for systemd, or its whole ethos of tentacles everywhere. It's all very making the complaint, but unless enough users are able to get off their arses and continually put in the time their words aren't worth much. Otherwise it's pointing a finger and saying "I don't like what you are doing but can't/won't do something different myself".
Re Linux v NT4. Linux was more stable, but in networking performance tests NT4 crushed Linux.
Quoth VoiceofTruth:
"Re Linux v NT4. Linux was more stable, but in networking performance tests NT4 crushed Linux."
Thanks to the efforts of a certain Alain Cox, this was no longer the case by the late 90s. In the end, Multiple Sclerosis had to swallow the bitter pill and adopt the BSD stack (hence. NT5). All their old stuff was geared towards X.400, (they'd never understood anything beyond LANs and WindowsforWorkgroups - reading the first edition of "The Road Ahead" might get you into your paces) and apart from that, NT4 was, just about, useful as a muliprotocol print server. The reason why it became somewhat useable for office work was Citrix.
I can tell you how much fun I had with this crap: remember 'winnuke'? NT 3.1 was a bit of a pretentious joke, and NT4 was absolute shite.
The same goes for their officeware (Microserfs is a nice book to read): because they think it works for them, they think the rest of the world needs it - or rather must have it shoved down their throats; I can't think of a bigger productivity killer. We have similar problems with SAP, Atlassian, and (I think I better stop here)
-> Thanks to the efforts of a certain Alain Cox, this was no longer the case by the late 90s.
That's not correct. Let me refer you to this: http://www.mindcraft.com/whitepapers/first-nts4rhlinux.html (article year is 1999), which indicates the problem clearly.
I remember this era very well. I supported and still support Linux. But we saw this sort of difference in performance too. It's no good saying "it ain't so", when it clearly was the case. I remember the furore, penguins were saying Mindcraft should have done this and that. It was all just noise.
As I wrote, we found Linux more stable but NT performance was better.
Thanks for the link - to be perfectly honest, at that time I wouldn't even have dreamt of using Linux as a fileserver on an industrial scale, for a number of reasons (no journalling for starters) ... most of that stuff was still in it's infancy at the time. And SMP support was rather patchy. But when it came to supporting small offices who needed internet access and were on a tight budget, users storing their files on their own PCs/Macs anyway, it was unbeatable. Plus Samba and Netatalk (on SunOs CAP was preferred) 'n all that.
'At that Time' are the key words.
Nowadays, things are very different. The world has moved on from those bad old days. Whilst I am nominally retired, I look after a number of companies E-Commerce sites. These all run on Linux. Some are CentOS 7.3 but they are slowly moving to AlmaLinux 8.7. A few of these were originally built for Windows/MS Commerce but the continual f**king around forced the companies to think again.
Other than Linux, they all use Wordpress.
Crashes? unheard of.
Uptimes? I will be patching one Centos7 server this weekend. It's current uptime is 325 days.
Back in the old days, yes, those of us early adopters had to work hard to get a workable system Those days are gone IMO.
At least with Linux, we don't get the OS provider forcing changes on us with every update.
If you still can't be arsed to try out a distro that isn't bleeding edge than that's on you. Fedora is the testbed for what goes into RHEL and its forks. I'll test it on an old Macbook when Beta 3 drops and yes, I will feed back any issues. I will put my test Wordpress domain on it, not just test the install.
I sure as hell hope that you ain't gonna plan on retiring any time soon. There's a whole cohort we'll have to outlive. (Mouse to the left of me, ashtray to the right, here I am stuck with my keyboard and you ;-)
Unfortunately, once you work for a company which has "adopted" Linux, you're stuck with something unbearable determined by the packager they've chosen, and then have to try and make the best of it. And the various Linux packagers - which is why I gave AIX a favourable mention - have never given us even the slightest indication of a roadmap, let alone target dates. It's just the same me-too-ism that MS practised and is still practising. I can assure you that life was easier in the 70s - and, BTW, rock music was much more interesting to listen to. At that time.
>> Uptimes? I will be patching one Centos7 server this weekend. It's current uptime is 325 days.
Peanuts. :) I have an Ubuntu server I installed when 20.04 came out. I'm going to have to upgrade it soon because that's going out of support. I have never rebooted it since it was installed. That's 3 years of uptime.
.....have mostly been superior to the (expensive) alternatives. Namely dropping $100 to M$ when I buy a new machine.....or dropping $1000 when I (don't) buy something from Apple.
Pick your poison, guys.....................................................but there's NOTHING wrong with Fedora 37!!!
I love everything about Fedora except the software Management. I used fedora 14 as a daily driver for years and loved it. The changes in gnome drove me to Mint, and LMDE5 has been a dream compared to everything else. Mind you I'm not running a server, a touchscreen, or a programming workstation. For a standard one-size-fits-all desktop I see no reason to leave LMDE.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
