I found Solaris to be a surprisingly reliable system, especially since 'Orrible had bought it, so I would consider Solaris to be a very fine alternative for most people who want "a Unix". This reliability even descended into illumos, at least to some extent.

The one big "advantage" of Linux is the low overall cost of a more-or-less complete operating system with some of the software ecosystem required for "usual tasks", minus multimedia design. In Germany, even big newspapers (from non-IT realms) suggest that all administrations move on to Linux "because Windows is expensive for taxpayers". I, for one, would prefer an administration which has reliable software to one which has free software, but that could be just me. Even here on LQ, a lot of people seem to assume that there only is a choice between Windows and "the only free operating system". I mean, I find it perfectly acceptable to use something which does the job well and remains unknown by mere mortals, but still...

After two decades of playing with C++, I tried Rust recently. I gave up rather quickly. Hooray, it has a package manager which is actually quite usable in build scripts. And then what? - Anyway, Kotlin is quite good, to be honest. It is a good mixture between Smalltalk and Lisp and it even is "a standard" now in the Android world. (Not really looking forward to Fuchsia. It will come with Dart. Writing apps in an almost-JavaScript language is nothing I would do for money. Luckily, I moved on to iOS for my personal needs.)

I would argue that Go is an acceptable "C for the web" too. That could be related to who designed it though.

Again, we live in a time where PL/I features are proclaimed to be the latest hype in programming languages. (I wonder how many youngsters who consider to become a developer even read our little exchange here? Hopefully, there will be quite a few. If so: You're welcome, youth!)

Technically, both were written by high school students. The main difference seems to be that the BSD guys actually had money.

You should chime in with a monologue like this ...
Making the future better by teaching the past.

It never could!

This is true for Ubuntu (and probably for RedHat too) but I don't think it can be generalised the way you do. If you think it can I'd like to see examples/proof.

Nevertheless, I think your subsequent criticism & analysis applies.
I just don't see it as darkly as you seem to.
As you rightly point out, most things technical have grown "organically", which is usually less than ideal for something technical.

I like Solaris. Hey, I worked at Sun Microsystems for 14 years. I didn't learn to be an engineer at MIT. Honestly, I learned at Sun... and at this startup where I worked before moving up to Silicon Valley. Both places had great engineers. We all learn by standing on the shoulders of the giants who came before us... except, those who think they know better. But to those I say "you don't know what you don't know."

But to be fair, Sun systems (hardware and software) were not as robust or capable or as fit for purpose as mainframes. That's not a criticism. First of all, the design goals were different. So my previous comment about the IBM mainframes is consistent with this latest statement. I was trying to say that what many people (most notably the Linux community) thinks is some new-fangled invention is really not. IBM had virtual machines running on System 360 in the '60s.

I'm not saying Linux has no attractive qualities. What you say in your last post @YesItsMe is true. And I freely admit that I'm cynical... as a result of many, many years of experience. I think it's amusing that the Linux fan boys think they're so hot but turn out such crap. I mean, relatively speaking, it's a fair statement in my opinion.

I won't go on a lengthy diatribe, but in other posts here on LQ I detailed how I tried to sincerely contribute to improving Linux. I was ostracized and unceremoniously pushed out. The problem was ego, protection of domain, power, influence and money.

I truly believe that the majority (not all, there were some excellent folks I met) of the people starting or running distros did not want standardization. I won't repeat what I said in this thread earlier. You know it already. ;-)

Yes, I'm critical. But that's because I've seen what good engineering can produce. And, if these fan boys were reasonable, or if they were mentored by those who do have the experience and good engineering, they would be willing to stand on the shoulder of giants instead of trying to just stand on their own.

My first job at this start up was fantastic. I was working with some incredibly talented, experienced engineers... some of them the original creators of the CDC Cyber-70. You mentioned Burroughs in an earlier post. @YesItsMe. I worked with some of these guys who came from Unisys (previously Burroughs) from south Irvine in Southern California. I learned so much about computer architecture from them by just shutting my mouth and listening after asking a question. The knowledge I gained is what enabled me to get into the OS group at Sun, first working on SunOS device drivers, then contributing to the DDK.

A couple of the guys at this startup wrote some of the first Fortran compilers. That low-level foundation, and their explanations and mentoring, gave me an incredibly solid foundation... not something you learn in school....

So, yes, I'm critical. And I'm impatient. But mostly, the bad taste in my mouth comes from the arrogance, and defensiveness of these immature types. @ondoho, you saw my post on the ridiculous comments I got from the manjaro community. I mean, how ridiculous when their X-Window system desktop manager is less configurable than the MIT distribution I used in 1987.

And why would anyone make a new distro but make it more difficult to use, less documented (inaccurately documented) and change one of the best pieces of software ever made available to the public. I don't call that good engineering. In fact, it's not even engineering. It's hacking. That's not a compliment.

@ondoho, I had volunteered with the San Francisco and San Jose LUG back in the '90s. I was shut down. As soon as I said anything about architecture, I was silenced. Two of the best engineers I've ever met (and I can't mention their names here without their permission), kernel developers at Sun, quit in disgust for the same reasons. No one would listen to their vast experience. To me that's shocking. You might disagree. But this is how I feel; it's my experience.

I'm not saying everyone is like this. But it does seem to be a community that has lost its way. And I believe it's a fair statement to say that one of the primary reasons is that people without experience or perspective don't know what they don't know.

@ondoho, you're right about RedHat. Their "product manager" is one of the guys who had me removed from technical forums.... because what I was proposing was a unified design, loose coupling, high cohesion between the kernel and utilities layer and also the entire X Window system layer. I worked on X at MIT. I offered to contribute my time for free. I was shut down. At that time, I think RedHat had a very commercial vision. They wanted to capture the market. They were not hobbyists. They were an aggressive commercial endeavor.

I just don't think any system that's of the order of complexity as an OS should be done ad hoc.

So... I'm slowly taking my short list of distros and I'll play with them until I find one that suits me.... It would be more fun if I had more time... ;-)

1 members found this post helpful.

Oh wow. Looks like we have another Linux Sucks Thread!

As far as if we really need that many choices on DE's, WM's, pkg-mgrs, etc.

The answer in my opinion is yes. Think about it... If Linux would be "one size fits all" we'd all be using systemd and what's more, gnome is also a good example.

And if we only has one or two pkg-managers, we'd likely miss out on the innovative new designs such as Guix, Nix, etc.

It's also important to realize. Choice in pkg managers is only limited by two things. Proprietary software, and the demand for a certain piece of software.

Proprietary stuff can't be compiled for a new architecture or ported to a new system layout/pkg system and this severely limits it's effectiveness however it's NOT the OS or the pkg managers fault.

Demand is another factor. I use Void occasionally. It uses runit and xbps. Since it's an independent distro the repo selection is rather limited. However they encourage you to submit and maintain packages if you need something that's not there. So if the distro is lacking software you can change that, or just build from source.

I agree that innovation is good. And experimentation leads to improvements. But you can have interoperability along with different products. For example, if Microsoft Word used a standard file format such as ODF, then an .odt file produced by Word would be perfectly readable by OpenOffice, LibreOffice, FreeOffice, etc. The products themselves can be different and they could compete on those differences. And the community could choose which they liked best. I think it's the same with OSs.

In regards to DEs, WMs in particular, they can be "different" in look-and-feel, customization options, GUI tools for customization and so forth. But, in my opinion, they should follow a standard interface so that any DE could operate with the underlying X Window system, any WM would work with any DE, etc. I worked on the original X Window system, Xlib, Xt, etc. It was designed that way from day one.

OK, to be fair, I haven't worked down at that low level in a long, long time. But I don't know of any good engineering or technical reason why that cleanly designed architecture can't be maintained. Like the word processor example above, there is plenty of room for these products to distinguish themselves on features, sophistication, aesthetic appeal, etc. But they should, in my opinion, follow an interface standard that maintains the original design goals.

And Linux is not the only culprit. Apple intentionally hides things from users. So, I cannot make X work on my mac without seriously risking screwing up my entire system. They don't publish documentation on their modifications to their version of the desktop manager, window manager, etc. And, my previous statement holds: I could customize the behavior of my WM more back in 1989 than I can now on my Mac. No one else might care. But I'm just stating my opinion that I don't think it's a good design.

For Apple, hey, go ahead and keep the configuration but make it another check box in the GUI configuration panel. They chose not to I guess.

So, in my opinion, Linux is still immature. Maybe people like it that way. That's fine. I'm not criticizing anyone else's opinion. I am criticizing, however, the attitude that I experienced when I tried to get involved to help.

Sorry, I forgot to say. At some point in time, improvements in a technology or product require an upgrade to the basic architecture of things. So, to use @Mill J 's example, at some point if someone figures out how to make a package manager more sophisticated in general, I think that's an improvement. It might be the case that it's not technically feasible, not a good idea or just too much effort to try to maintain backward compatibility. In that case, I think that the improvements should spearhead a new "standard."

I have never worked on a package manager. But I do wonder whether all the different ones could use some base compatibility layer. Maybe there could be some kind of capability to "advertise" which features were supported including dependency checking. Just a thought. I don't think that innovation needs to throw out this morass of incompatibility. Maybe there are good technical reasons for package managers. But I don't think it's the general case that you need to throw out standards and interoperability in order to have innovation and distinguish yourself from the field.

What I was trying to say--and maybe I didn't say it well--is that I am personally against difference for the sake of difference. It's just my opinion, and no one has to agree with me. Going back to the word processing example, I'm against Microsoft's philosophy of making the document format proprietary in order to gain competitive edge.

If you remember the famous law suit between Sun and Microsoft... Microsoft intentionally changed the Java runtime so that Java compiled (or even Java source dynamically interpreted) would not run on Windows. Microsoft removed support for certain classes and methods. They did so to discourage frustrated developers so that those developers and users would just buy everything Microsoft and use Microsoft tools and platforms to build and run Java. I prefer the "make a better VM" to compete. Or "make a better-looking native UI" to compete.

OK, I'll be quiet now.

Some good points.

I've always been one of those guys who like to tinker with obscure OS(That's why I'm posting in a plan9 thread ). I occasionally compile OS just for the fun of it and to check out the development stack, So I'm by no means sold on Linux, there are projects that have way more potential but the biggest problem is always software/drivers. Even BSD is behind Linux in this area. That doesnt mean those OS aren't well written/designed. The same can be said for Linux as well.

As far as X. There's wayland that's gaining traction.


I can't really see where Linux is immature. I don't know of any other kernel that is complete enough to be used as a daily driver and yet can be stripped down to several megabytes, by anybody with a little time. And what's more I don't have to rely on some company to build the kernel to they're specs. It's fairly easy to configure if you use the included config tool.


Honestly by far my biggest gripe with Linux is the huge amount of commercial companies(including MS) scrambling to extend it. It's on borrowed time because of that.

@Mill J, you make some good points as well. Let me try to clarify because there's always some subjectivity in what anyone says despite best efforts.

When I say Linux is immature the context specifically is the ease with which I can just get done what I want to do. Naturally "what I want to do" and "how much effort is too much" is subjective. So I understand if others don't agree.

I have concluded that I need two systems. One that I can play with to learn as you do, and at least one that is my reliable daily system. In fact, I want to play, build systems, etc., but I'll do that after I get a stable, reliable, Linus platform going. I want to use some Linux distro to be my "play and learn" system, not macOS or Windows. I really don't care to use macOS or Windows for that, not at all.....

So, I don' t dislike Linux generally. My perspective was just from my "user's point of view" -- and me specifically, someone who wants a robust system without the need to tinker too much to get things "just the way I want it." By no means did I mean that Linux is immature from the general perspective of capability.

To get a Linux system that can be that daily reliable one to use professionally and personally would take me more time than what I can get from Windows or macOS. I mean, to be fair, my old Dell Inspiron laptop (at least 6 years old...!!) running Windows 7 has not crashed at all. I can use it to do all email, and I can use GIMP, LibreOffice, read PDF documents, etc. just fine. And, with cygwin, I can play with a shell to test scripting and stuff.

That being said, clearly the quality of Windows is ghastly compared to Linux or macOS. Outlook's or Windows' search indexers are terrible compared to the Mac. I use AVG anti-virus and it's good quality but chews up CPU like nobody's business. I can't use my system for 10-15 minutes after booting until AVG runs, indexer runs, instup.exe runs, etc., etc. It's terrible.

Further, to be fair to Linux, I haven't tried all mail applications. I know I don't like Thunderbird; it's immature, and many folks on LQ agree with me on that point...! But I haven't yet tried Mailspring. I'll try it when I find the Linux distro I like. I'm still on that journey. ;-)

Also, I do know that in recent years Linux has adopted some of the best features, designs, implementations of Solaris and solid commercial Unix-like platforms. I believe Linux (at least RedHat, but I'm not sure if it was adopted universally) took the SMP stuff from Solaris. So clearly it's improving all the time.

And thanks for the pointer to wayland. I just looked it up. It looks interesting. I think this is a good example of how things advance. Philosphically, I think new things should take the best and then build atop that if it improves things (I know, "improvement" requires qualification). If they take to heart the lessons learned from X, then it will be great. Great example of what I was trying to say previously.

I also want to add something for perspective. As much as I thought Solaris was (or still is) a robust, advanced, sophisticated Unix-like system, it is not without problems. I remember how our very excellent, skilled sysadmins at Sun would just roll their eyes in dismay whenever the topic of volume manager arose.

But overall, I thought both Motif and Open Look were pretty polished and user friendly. Motif was a bit of a pain to use to program GUI applications, and I thought Open Look's declarative style architecture was much better from a computer science perspective as well as from an ease-of-use perspective. You could easily edit the simple text file in your home directory to do some really fine-grained control of the desktop manager and window manager.

Honestly, I think Linux DMs and Was are not as sophisticated. I could be wrong. I can just say that using Debian and Ubuntu left me underwhelmed. Recently Manjaro was really, really disappointing as I posted earlier. I have to do a lot just to pin an application icon in the dock.

That's what I meant by Linux is immature. I was not referring to the kernel or the utilities. However, and again I could be wrong, I think there is still a long way to got towards standardization and interoperability. I don't intend it to be a criticism, just an observation.

I'm happy to have anyone correct me because I'd like to learn. Just point me to some article or documentation. I'd like to know for my edification. But it might also help me choose the distro that I will be happiest with.
;-)

OK, I promise, no more espousing.... I just wanted to add that I know this is a Plan9 thread, and that's what originally caught my eye. I'm interested to learn more about Plan9. So I'm all ears....
;-)

1 members found this post helpful.

While plan9 is a very interesting system, being immature would be a huge upgrade. Not bashing it at all though, since it never claimed to be more.

Sure, I understand. I would not be interested in Plan9 as a daily use system. But I am interested to learn more about it. I only remember a few things from that lecture that Dennis Ritchie gave at the Usenix conference way back in the early '90s. Hey, if I learn one piece of interesting information a day, I'm doing well.

Many times I have just sat and listened to someone who knew something I didn't. Sometimes years later, even if I didn't remember every technical detail, I could always recall the concept, and it enabled me to find the details when I needed them.
;-)

I'm watching this thread so you gentlemen continue...!

2020 will be the year of the Plan 9 desktop!

See, Word actually uses a free standard file format (OOXML) these days, giving the "X" in "DOCX". However, it is poorly supported by most competitors except SoftMaker. Openness/standardization does not seem to be the relevant problem here. ODF is not even supported well by LibreOffice, or at least I found huge problems when trying to use an ODF file in a different ODF-capable application. (Which was the day I stopped using LibreOffice.)

@YesItsMe, which word processor suite (document writer, presentation and spread sheet) do you use? I moved to LibreOffice when I discovered (I think here on LQ in another thread someone said) that most of the OpenOffice developers defected (perhaps because of the same in-fighting that occurs in every other project :-) ). I like it, but I'm happy to hear of others if they are doing a better job adhering to standards and if it's good.

I just think that in the long run standards will win. One of the big problems with OpenOffice was that it didn't do a good job with .docx, .pptx, .xlsx files. Probably not their fault because doesn't have an interest in interoperability or fair competition.

Hey, I'll be happy to see Plan9 take off...!

I wanna see glenda on the members profile on the left sidebar.

This from a guy stuck like slackware user for years. But not Slackware. Probably fewer members. I wonder what the browser string is for glenda/plan 9?

Cuz

https://www.zdnet.com/article/google...ngs-in-chrome/