OS WARS MEGA THREAD (Now debating proprietary vs. open-source!)

[Artlav]

I'm somewhat shocked, but wish I could say I was surprised, to find myself in the position of likely switching to KDE in the near future.

Why not XFCE? Been using it since forever, it's clean and does not get in your way.

 

[Linguofreak]

Why not XFCE? Been using it since forever, it's clean and does not get in your way.

Also uses GTK3 these days: I don't like a lot of the GTK3 dialogs, in particular the file chooser dialog, and GTK also has a crapton of deprecated functionality deliberate regressions in theming.

MATE and XFCE haven't regressed in the non-toolkit stuff, but the system toolkit is half the reason for using a desktop environment, and GNOME has been driving GTK into the ground. If the MATE and XFCE projects aren't willing to fork GTK and maintain their own version, I can't blame them for going to GTK3, as GTK2 isn't maintained anymore, but someone really ought to fork GTK. A GTK2+i would salvage the GTK ecosystem, but at this point QT seems to be the trustworthy toolkit with a slow and steady, "don't break anything" development philosophy.

 

[martins]

Sometimes it baffles me just how inefficient Windows can be at the most basic tasks. I am currently moving (or trying to move) a couple of files (well, ok, about 10^5 of them, but still...) from one directory to another. The two directories are on the same file system, so no physical data transfer of the file contents should be required, only directory entry updates. I would naively assume that this should take a few seconds at most. Yet, for the last 3 hours or so, Windows is informing me that it is "preparing to move" and "discovering items" (apparently it has discovered zero so far). Now call me a slave-driver, but when I ask the OS to move files, I'd expect it to do so promptly, instead of embarking on a journey of self-discovery. Is there a way to tell Windows to stop discovering and just get on with the job?

I wonder if I should carry around a Linux installation on a memory stick so I can just plug it in and quickly switch over to Linux whenever I need to do heavy-duty file operations.

 

[dbeachy1]

It's probably having to do lots of physical disk head seeking to read the file info for all 100,000 files. According to this, you could try moving the files using the Windows Command Prompt, since it appears that an extra validation occurs when moving files using Windows Explorer (perhaps so it can count all the files up front and show a progress bar?). It does seem odd that it takes 3 hours, though. :blink:

 

[martins]

Good call! I abandoned my movement attempt via explorer and used the terminal instead. Took about two minutes, much of which may have been spent on just echoing the file names in the terminal window.

 

[Linguofreak]

Sometimes it baffles me just how inefficient Windows can be at the most basic tasks. I am currently moving (or trying to move) a couple of files (well, ok, about 10^5 of them, but still...) from one directory to another. The two directories are on the same file system, so no physical data transfer of the file contents should be required, only directory entry updates. I would naively assume that this should take a few seconds at most. Yet, for the last 3 hours or so, Windows is informing me that it is "preparing to move" and "discovering items" (apparently it has discovered zero so far). Now call me a slave-driver, but when I ask the OS to move files, I'd expect it to do so promptly, instead of embarking on a journey of self-discovery. Is there a way to tell Windows to stop discovering and just get on with the job?

I wonder if I should carry around a Linux installation on a memory stick so I can just plug it in and quickly switch over to Linux whenever I need to do heavy-duty file operations.

Ehhhhhh.

Yeah, keeping the files at the same place on disk and just updating directory entries is what *should* happen, but if it is trying to move data, it could end up spending most of its time seeking between two locations (if your drive is the "spinning rust" type and not an SSD), which could tank performance bigtime. I can't promise that Linux will do any better, but I don't recall any large intra-filesystem moves ever having behaved like that, but, OTOH, I can't recall the last time that I tried doing something like that (though it would be more likely to stick in my memory if it had gone poorly).

 

[Artlav]

I'm still baffled that there are people out there that use Windows Explorer for managing files. To me it feels like doing something through a borescope - you see nothing, you control nothing, basic operations are as complicated as possible and require third-party software, and apparently it's also not very efficient.

Then again, i've been using Norton (Volkov, Windows, Total, Midnight, etc) Commander since the day my dad got a PC back in 1990.

 

[Linguofreak]

So, I've been trying out OS/2 lately just to see what all the hype from the 90s was about. If its proponents are to be believed, it was 10 years ahead of its time and would have taken the world by storm if not for Microsoft's predatory business practices.

I won't argue about the business practices, but I am so far unimpressed by OS/2 v 4.0 itself. I couldn't find working drivers for the network hardware available with qemu/KVM, so I tried the serial route, with tcpser emulating a modem on the host side:

1) The system GUI periodically stops responding to mouse and keyboard input. The machine is still alive, and the screen is still updated, but the input-deadness forces a reboot.
2) The system dialer failed to even *dial*, let alone establish an IP connection (tcpser didn't receive a single byte coming down the line). At first I thought this was an issue with the VM's serial emulation, but it turned out that echoing "ATDT555..." to COM1 at a terminal did actually cause TCPser to dial. But it took me a while to discover that, because the dialer didn't release the serial port after failing to dial, so echoing directly to the port just hung the first time I tried it.
3) So I found a third-party dialer and was pretty much immediately able to get a connection. So then I tried the system FTP client (both of them!). They connected to the host, but fell flat on their faces as soon as a file transfer was attempted. I had to bring in the GNU inetutils to get a working FTP client.

So far, I don't think IBM can blame Microsoft for OS/2 failing...

 

[Urwumpe]

So far, I don't think IBM can blame Microsoft for OS/2 failing...

Still it lived extremely long in German banks and insurances. Not sure when they started switching their special software to Windows, I still had some people doing OS/2-support around 2015.

EDIT: And according to German Wikipedia, there are still companies using it AFTER 2015...

 

[jedidia]

So, I've been trying out OS/2 lately just to see what all the hype from the 90s was about. If its proponents are to be believed, it was 10 years ahead of its time and would have taken the world by storm if not for Microsoft's predatory business practices.

I tried that once back in the days... it was a jarring experience, though in retrospective I don't think it's fair to blame the OS for that.
I think what made it fail is not so much Microsofts business practices as that it was an extreme paradigm shift for PC-users (that would be the "10 years ahead of its time" part, which is admirable, but only profitable if you have the sales machinery to convince people that it's worth to learn all the new crap required for such a thing. Also, 10 years is probably exagerating it somewhat).
The main reason why Win95 made the run despite being less stable (allegedly... There's no telling how OS/2 would have performed if it saw such widespread adoption) was that Microsoft was in a position to *force* that paradigm shift onto users. And also because they could say "hey, look, DOS is still running in there, you can switch to DOS mode any time you want!"

And sure enough, I was doing a *lot* of stuff just in DOS mode in those early days.

EDIT:
Also, as for your specific criticism, you have to understand that seamless network connection was not yet a priority when OS/2 had its run. If it would have survived, those things would certainly have been improved upon in the early 2000s.

 

[Linguofreak]

EDIT:
Also, as for your specific criticism, you have to understand that seamless network connection was not yet a priority when OS/2 had its run. If it would have survived, those things would certainly have been improved upon in the early 2000s.

Yeah, the third party dialer I tried turns out to dial and connect, but there's still squirrely stuff going on. It turned out that the (g)new FTP client didn't immediately do better than the other. After changing some dialer and server-side PPP settings, I can now get transfers to complete, but the files get corrupted in transit. I have just noticed, and am rather worried about, the fact that this third party dialer doesn't have any settings for bit-width and parity.

 

[Linguofreak]

Just in time for Halloween: KDE never quite cut it. GNOME was a rotting corpse, and its dominance if the GTK ecosystem had compromised the viability of MATE and XFCE. In desperation, the mad scientist grafted the last few uninfected bits of MATE onto KDE, creating a monstrosity with overlapping MATE and KDE panels. Will the world survive the

NIGHT OF THE FRANKENDESKTOP?