Date: Tue, 28 Jul 2020 01:29:42 +0200 From: Polytropon <freebsd@edvax.de> To: freebsd-questions@freebsd.org Subject: Re: Ask stupid questions and you'll get a stupid answers, was: Technological advantages over Linux Message-ID: <20200728012942.2157eb85.freebsd@edvax.de> In-Reply-To: <20200727153926.2c8bbfc3@scorpio.seibercom.net> References: <CAGBxaX=Ktr-pqtT8FU37ajkYonVLYT_WhSenn23Tj5b=i0d-8g@mail.gmail.com> <20200726165212.a0ac28ce104b9dfd009cc4c5@sohara.org> <20200727025242.GA48950@admin.sibptus.ru> <04bb7095-3e62-22d7-9db7-cfba3c66247b@kicp.uchicago.edu> <alpine.BSF.2.23.453.2007270732330.893@fbsd.local> <b9046235-27f0-4305-fbe7-4ed431e79439@yuripv.dev> <CAGBxaX=jy5ZB6GxeB0tii-gd1rmS1h_mHH9g71puO%2Bud_h3kdg@mail.gmail.com> <alpine.BSF.2.23.453.2007270839320.972@fbsd.local> <CAGBxaXn=TORaWE0RMDv1Z6p3Vk48BHcYtF0RukrCHfieG_BQsA@mail.gmail.com> <20200727063036.31316d93@scorpio.seibercom.net> <CAGBxaXn_BV%2BZ50G1pk682D=WnYHYhMYD3%2BEdWZ6Orp5PyKo75Q@mail.gmail.com> <20200727082427.3dd26eeb@scorpio.seibercom.net> <CAGBxaXmAkmbJqzJZpQahHNYrKQcTv8h%2B-mKTRzHz6Zc%2BZ=aBig@mail.gmail.com> <20200727121736.4f95993e@scorpio.seibercom.net> <CAGBxaXmqP81uOZThWPQ64_L_F1pT7CzRXZYvzsF95=JCE3528Q@mail.gmail.com> <20200727153926.2c8bbfc3@scorpio.seibercom.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 27 Jul 2020 15:39:26 -0400, Jerry wrote: > There is absolutely no contradiction in what I said. Using FreeBSD for > "X" and "KDE" is NOT the optimal setup. FreeBSD is simply NOT the > optimal OS for that pairing. Now, if a user is not interested in or > requires optimal performance, then that is their decision to make. Using FreeBSD + X + KDE does not have to be less performant than a comparable solution. Take for example the port "mkdesktop" which delivers an "almost Linux-like completeness" of KDE for FreeBSD. However, I agree that you have a certain point: The "out of the box experience" of KDE on FreeBSD isn't nearly as complete and integrated as it is on Linux. Why? Because KDE is developed on and primarily intended for Linux. As I mentioned in an earlier post, this creates specific problems whose origin is simply admitting that FreeBSD != Linux, and that's it. You maybe do remember how PC-BSD (later TrueOS, now discontinued) started? They took FreeBSD + X, added KDE, added custom tolls to be able to configure FreeBSD from within KDE, added a KDE-based package manager for their PBI system. Neither was this less performant than a Linux distribution targeting KDE at that time, on the same or comparable hardware. You see: It was possible. And it is still possible. Of course you are right when you now argue that this requires time to set up. That's absolutely correct, and you (as the user) have to decice if this is worth your time, which might be money you're losing. There is a quite simple consideration: Am I happy with what another OS delivers me as a preinstalled and preconfigured desktop? Does it fit my needs? Does it fulfill my requirements? If yes, I'll probably just use that. But maybe do I need a custom-taylored solution that maps to a workflow no "out of the box experience" can cope with? Is this an O(1) action, i. e., I set it up _once_, then keep using it? If this boosts my performance, then I'll probably take the time to do that work. On FreeBSD, an OS which does not include (or come with) one specific desktop environment, the user can choose. > FreeBSD is an excellent OS for use with Apache, Postfix (or other MTA), > database programs like MySQL, etc. Again, I use, whenever possible, the > best OS for the problem at hand. That is the most natural, the most educated, and the most obvious thing to do. > Only a fool blames his tool. Any bad craftsman blames his "bad" tools. :-) > I know FreeBSD sucks with KDE, but I don't spend all day here bitching > like a little girl. I reserve that for a more appropriate forum. FreeBSD does not suck with KDE. Please don't generalize, it's simply not true. I mentioned the existing restrictions, but maybe they don't matter to you (i. e., you don't need a KDE- based interface to /etc/rc.conf, so it's not a problem that there is none). Again, let me emphasize that the more modern and advanced KDE gets, the bigger the gap originating from FreeBSD != Linux gets visible to the user. It doesn't matter that there are claims like "any POSIX-compliant source can be compiled on any POSIX-compliant system and will produce a comparable result" - it doesn't apply here because of the heavy accumulation of stacks of libraries of abstraction of libraries of stacks of interfaces of libraries of libraries of kernel calls. ;-) Expecting (!) a default KDE installation on FreeBSD to work just like on Linux is a case of "blame the tools", in my opinion. If you want a "just works" solution, Linux is often the way to go, except, of course, you actually need FreeBSD, which Linux obviously isn't. > I know > FreeBSD driver support is sub-optimal, but there is little that I can > do about that. Things are improving. If you're lucky and not depend on certain drivers, you are in the great situation that you don't have to care. For example, if you have a system that is intended to run Linux, where Linux has support for everything, but FreeBSD just causes problems on such a system, but you still wish (or need!) to run FreeBSD, using a VM solution gets you rid of all of those annoying problems. Plus, you can use the majority of software you know from FreeBSD because it's available on Linux anyway. > Aryeh, at some point in life, you have to learn to accept life as it is. But if you do that, you have no chance to make a change that leads to improvement. You'll have to eat whatever is fed into you. ;-) I absolutely understand mindsets like "time is money" or "don't fight when you cannot win". Accepting everything as given, and learn to live with the least annoying solution, instead of standing up and building something better has lead to significant problems in economy. It's not per accident that we're using stuff that is more or less "overclocked outdated steam machines" and mental concepts that are 50+ years old and don't match reality and current requirements anymore. People don't want to change. Sometimes they have to. And in the worse cases, it leads to cost increases, workflows breaking, data loss, redundant work, or simple despair. > I like utilizing two complete OS's. It suits my needs perfectly. And, > it is extremely simple to setup. If you are happy with the KVM, that is > fine with me. Having two physically separated fully equipped systems can be an advantage, just as using a KVM switch to interface with two systems using one "workstation frontend" - it depends on what you're doing. There is no general solution that works for everyone, no one size fits all egg-laying wool-milk-sow. Sometimes, having a Macbook next to your FreeBSD workstation is exactly the solution that helps you achieving your goals, and sometimes it's just a distant server that you connect to using SSH from whatever system you want. > But is it a 72" curved flat screen with surround sound? I kind of doubt > it. If it suits your needs though, that is all that matters. I was able to get FreeBSD + Mate work on such a system (even though not with sound, it didn't have any). :-) Follow Aryeh's experiment, and you will see that if this always leads to a less performant desktop than it did in the past, that's surely a problem. There is a convenient equation for "overall usage speed": resources provided by hardware++ -------------------------------- = const. resources demanded by software++ The quotient "const." isn't actually constant: In specific settings, the "overall usage speed" gets lower, read: some stuff is getting slower. That's what you can _observe_. The primary reason is increasing program complexity. Consider a "simple" thing like a web browser: it has become as complex as, or even more complex than a whole operating system. Why? Because prople treat it as one. That isn't just an increase of resources demanded, but also more chances to get things wrong, and lead to memory leaks and such. Now consider the libraries intended to be used, like Gtk or Qt. They expect things you find on Linux. They're not there on FreeBSD? Okay, let's create another layer that hides this fact and somehow delivers what's needed. Oh wait, Linux introduced something new that is now also expected? No problem, add another layer that provides a shim so it can at least compile, and we'll add the missing functionality later on. If performance is your goal, you maybe have to say goodbye to KDE and Firefox, and look for something else. On FreeBSD, you can actually do so. Choice is key. And if your choice is to _not_ use FreeBSD, that's fine - but it does not help solving the existing problems on FreeBSD. It's hard comparing things with all that growing complexity in mind. I still have a non-networked 300 MHz FreeBSD 5 PC here that runs XFCE 3 as a desktop, Opera web browser, and OpenOffice 1.1. It boots extremely fast and is a general joy to use. However, it doesn't match the requirements of today's web anymore, let alone security considerations. But it beats any newly bought "Windows" PC in startup time and what you can do with it (browsing the web is limited, of course, but it opens office documents, even broken ones, plays media, has network diagnostics, has a C++ compiler, can print to a laser printer, has games installed, can edit photos, and doesn't even have to phone home to obtain missing parts). Now compare _that_ to a system where you can't even use the calender "app" without being online and registered as a paying customer... ;-) -- Polytropon Magdeburg, Germany Happy FreeBSD user since 4.0 Andra moi ennepe, Mousa, ...
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20200728012942.2157eb85.freebsd>