Marak most people cannot afford a high end machine today. Me included. The great depression killed my IT career. 20+ years but I had no degree and was competing against people with the same 20+ years of IT experience who had degrees for jobs that wouldn't even pay my child support. Outsourcing and guest workers destroyed the pay scale in IT in the late 90s, early 2000s and it hasn't really rebounded. So I went back to college and switched careers. Currently working on my PHD in psychology.
I've been more than a back seat developer. Over the years I've contributed to the code base of several opensource projects and was part of the Fedora docs team for a couple years. I'm out of date with my coding skills, but could jump into some Python based projects today if I didnt' feel KDE was going the absolute wrong direction. My C++ skills have atrophied to the point I wouldn't feel confident coding in C++ any more. Not anything major. I spent 10 years writing code for a living, mostly VB & C++ but some Asm, Pascal/Delphi, Foxpro & other Xbase platforms, even odd stuff like a little COBOL (test app to see if the Asm apps I wrote to talk to PC Cobol from mainframes worked). Even dabbled in some Cold Fusion & PHP back when web programming was new. Converted Blowfish to ASP for one project as the M$ dlls were klunky and insecure. We needed native encryption for the app and Blowfish worked out well. Though about 5 years later they found a serious flaw in Blowfish. Hopefully the end users replaced that module lol.
I am not afraid to roll up my sleeves. I've actually volunteered for a couple KDE projects but never heard back. This was years ago when it was KDE 3.
As for the validity. It must be nice to have $2k to throw at a computer. Most people do not have that kind of money any more. If they can afford a desktop at all, it's the $400 Walmart special which you consider obsolete. The latest laptop I bought for example has only 4 gigs of RAM. It's not any faster than the 8 year old machine, also with 4 gigs that I am typing this on.
A lot of people used PCs only for email, browsing and short messages. Those people went to phones. They already had to have a phone anyway. The phone works for them. The people using PCs today are mostly gamers and people like us. There is however a big niche for low end PCs. People who cannot afford the high end PCs and the M$ insanity and the thousands of dollars of software to get a useful PC using Windoze, only to do it all over again 2 years later.
The economy is down world wide and there are millions of users in emerging nations like India and Brazil who could use a cheap, stable, easy to use OS/Desktop that runs on a hand me down or very low end machine.
Think about who KDE 5 is targeting? Wealthy and upper middle class older males. That's about it man. The resource demand is too high to run on low machines. Linux scares a lot of users because the Linux community has failed to develop and promote low end machines that do more than what a Chrome book does.
KDE is the perfect platform. Unlike Gnome where the keyboard shortcuts, look and feel and interface is dramatically different from app to app. Where the quality of the apps vary dramatically. KDE has a simple interface. The keyboard shortcuts are generally universal. Many KDE apps are best of breed. The worst of them at least do a decent job. It's a massive resource hog. You expressed exactly the sentiment that will kill KDE. The let them eat cake. If they are not rich like me they don't deserve KDE.
It's also pointless. Yes PIM IS simple. Seriously, how difficult is it to manage information? I've run million plus record MySQL and Postgress DBs on very low end hardware before. Even made them small office production DB severs and often they also served as the file server and maybe even print server for the office also.
Email is simple. Yes I have experience. Not specifically with Kmail but other email formats. I even wrote the world's largest Outlook conversion program at the time. I wrote the code that converted Walmart to Outlook from a defunct email format, that had no documention. Not that Outlook's documentation was all that helpful. I had to use hex editors to reverse engineer both formats, then object browsers to work out the Outlook API. It's NOT that complex. I did all that in 3 months.
I'd wager money you got a lot of young programmers who are so OOP fanatic they create methods for 1 line procedures that call other 1 line methods and similar spaghetti code. I've torn into so many of those over the years. Young programmers think it's impressive to do that kind of thing. Annoys the crap out of developers that have to follow later and understand that code and you have to chase down 10 methods to see what is going on in what could have been a 5 line piece of inline code or been a single method/function/property/sub routine if they hadn't gone all Hungarian on the thing.
I also see people who've never taken the time to understand the compiler and do the dumbest things, totally oblivious to what kind of end code is generated. All compilers have bugs and things you should avoid. The biggest part of learning a language is learning those little undocumented things about your language.
Know your libs, bypass buggy resource monster elements of your dependencies. If you have too, it's open source. Rewrite the section you need to use to fit what you are doing. It's a pain doing updates to keep it fresh for new releases of dependencies but at least you are not slamming your poor system with a known bug/problem that will never get fixed.
Thou shalt Malloc and frequently lol. Just b/c Linux has garbage collection doesn't mean you should rely on it to clean up your mess. If anybody started hunting memory leaks on KDE they'd be a very very busy person.
Just because there is lots of RAM on a developers machine doesn't mean end users have that much RAM or use that impossibly small font. I see so many apps today that are difficult to use in even medium resolutions because the button you need to hit is way off the bottom of the screen. As a developer I ran very tiny fonts also. But my eyes were 20 and 30 year old eyes not 50 year old eyes. I cannot read those fonts any more. I run much lower resolutions today. There are some who have to run 640x480 due to eyesight limitations.
Try to understand your end users. Most OSS developers write what they need and if that's good enough for others great. That's about the limit of their empathy for their user base. It works for me, does what I want. If anybody else finds it useful more power to them is the attitude on full display by so many. That and if it's not in the specs I'll not even stub it out or take it into consideration even if I know it'll be added later. That used to drive me nuts about other coders. They'd KNOW we'd need it. The least they could have done was stub it out and leave room to add functionality. Instead I'm rewriting their code from scratch because they precluded all expansion on what they were doing. I always left stubs for expected or potential expansion of functionality. It only takes a few mins, there's no memory or compilation penalty for doing it and if you create vars instead of hard coding stuff, you make it easy to change/expand the functionality. Also fix bugs. I hate people who hard code data with a passion. Almost as much as I hate people who use Hungarian notation lol.
My root points.
KDE can be the goto OS for Linux users. It can help bring millions of users to Linux.
The focus of KDE is on a dead end and dwindling group. It would be better to create a KDE app than focus the entire desktop on what KDE is trying to do. It's like building an OS to do PIM. Rather pointless unless you are using dedicated hardware and even then, look at how phones are used. They do a whole lot more than PIM. KDE can be the killer app Linux needs to put Windoze down for the count. It just needs to be more versatile and dramatically lower resource usage and Akondi must die. Turn Akondi into an app instead of a service. If you need those features, run it WHILE you are using those features. Me I'd LOVE to be able to open up Kmail and sort through my ancient KMail archives but I don't dare do so without inviting the Akondi monster in. Akondi does NOTHING for me. Kmail could be useful now and then. I might even set up connections to domains I manage, but not if Akondi is invoked. There are other similar services I'd wager. Akondi is the most obvious and most gruesome of the lot.
My effort is to get the KDE community to look at who exactly it is they are writing software for and why? What is the future of KDE? Why drive people away from KDE? What gain is there in that? The fewer people who use it, the less likely distros are to support KDE. The fewer spins there will be of KDE with different distros.
XFCE and Trinity are options yes, and more and more people who WERE KDE users are moving to those options. That Trinity exists at all is a sign of dissatisfaction with KDE and that a large base of users have moved away. KDE needs to reconsider what they are doing in my opinion. Who they are doing it for and what the future of the PC is. PIM is not the future of PCs. It's handy to support PIM, but to dedicate a desktop too it and devote so much resources too it makes zero sense.
On Thu, Oct 15, 2020 at 1:49 PM Marek Kochanowicz <sirherrbatka@xxxxxxxxx> wrote:
First of, i can't understand why you perceive akonadi to be a monster while on
my machine plasmashell uses far more memory then all of the akonadi + mysql
combined (more then the firefox at the moment actually). Secondly, backseat
developers don't get anything done. PIM (contrary to what you said) is not a
simple task and the main reason why akonadi exists is because the original KDE
PIM codebase became unwieldy.
Anyway, I can't understand why one would insist on running KDE on a obsolete
machine while there are "lightweight" alternatives like mentioned trinity
(btw: KDE 3.5 was considered to be a resource hog back in the day) or XFCE
(which is actually a better option then a trinity). They also run on linux.
This is end of topic for me.
