EDIT: congratulations with LibreOffice port, that does make a difference in terms of usability.
Probably because you're on the VESA driver, which can only use whatever resolutions are embedded in your GPU. The `radeon_hd` driver works best, and on those you can scale to a lot more resolutions; I know some users who run at 1440p.
> and having no multi-user capacity
That's not quite true; we have POSIX multiuser already, so you can `useradd`, `chown`, ssh in to other users, etc. The GUI does not support multiuser yet, but that's partially for BeOS compatibility (on 32-bit), and partially because nobody has had any time to work on it (all other platforms.)
As to multi-user without GUI on Haiku, while I can see that this would technically work (and it's nice to know the capacity is there) it also leaves behind the strongest asset of Haiku which I think is the UI.
We do have an intel_extreme driver, but it's not as mature than the radeon_hd driver, mostly because Intel's graphics hardware is a lot more complex to handle than AMD's. It works on most chipsets up to Haswell and then the Atom SoCs, but anything after that (Broadwell and beyond), it's doesn't really. I have a Broadwell device it doesn't work properly on, and was trying to fix it, but haven't really gotten anywhere yet.
Patches wanted & accepted...
Depends on what you mean by "modern computing environment". I mean, my computers are mine, an no-one else uses them. I know very few people that share their computer with other users (but maybe it's just my bubble).
Another story entirely for those who use a smartphone/tablet for their personal computing (the target use of Haiku).
That's only half of what a multi-user system does though. The other half is about fine-grained privilege management.
One of the reasons (not the only reason by any stretch but a contributing factor certainly) why Windows has so many more problems with viruses than other OS's is that under windows you're basically either a user or an admin, there's no in between. Crucially, there's also nothing lower.
Under linux[1] systems, it's totally possible to define a user that can read and write from one specific folder in the file system tree but not it's ancestor nodes, read from the keyboard but not the mouse and absolutely nothing else. All of those problems that windows has with random programs being keyloggers are eminently solvable on linux because you can make a custom user that cannot use networking or listen to the keyboard as appropriate, and then launch the process using that user.
Anyway, how well used all these capabilities are depends a lot on the person administering the linux system, but windows is extremely reactive in this fashion. Desktop linux users tend not to run outgoing firewalls, for example, they just don't allow software that they don't trust to access the network.
[1] anyone know how to write star-nix on hacker news without it making everything italics?
Useful in a networked corporate environment, a lot less so on a personal computer.
> All of those problems that windows has with random programs being keyloggers are eminently solvable on linux because you can make a custom user that cannot use networking or listen to the keyboard as appropriate, and then launch the process using that user.
You can do that on Windows and have been able to since at least Vista. the only reason I disqualify XP is that I wrote a keylogger for it so I know for a fact that you can grab raw keyboard input from any user session. Even before that though NTFS offered a fine grained access control mechanism for files and folders.
> Desktop linux users tend not to run outgoing firewalls, for example, they just don't allow software that they don't trust to access the network.
My experience is that they just don't run software that wasn't sourced from their distribution's package repositories. That is certainly an advantage of the appstore model, but you have to admit it is significantly more limiting that being able to get the latest version of something by downloading it from the vendor's website.
It would be very limiting if it were reality, yes. I frequently download packages that aren't part of the default distro repo though. There's lots of PPA's, and I also install a lot of software just by extracting it into a directory in my home folder and running it.
But i disagree with the part about how Linux users behave: from my experience Linux users, at least home/desktop Linux users, behave pretty much the same way as Windows users.
Only have one single * in the entire post.
[X] Not sane
[X] Already complained
[X] Issue acknowledged
[X] Complaint filed probably over a year ago at this point
[X] Given up; incentivized to build alternative to HN (know what I want, but currently have no time unfortunately)
EDIT: I added Unicode checkmarks to my post, and they got eaten. [X] Cherry+icing on cake - seriously, wow.
When I grew up having a computer was a big thing and we certainly didn't have one for each person in my home we had a family computer which i think cost around 1K in 1995 or around 1700 adjusted for inflation.
I think that as computers have gotten progressively cheaper it ought to be increasingly likely to have a computer per person but I feel like the family computer ought to still to some degree to be a thing but have no stats to back this intuition.
However even with one user it would be good to have a login to protect data from at least casual snooping. Ideally since laptops have become more common than desktops full disk encryption ought to be sort of a minimum requirement. Maybe someone can port luks.
Harder but zfs would also be neat as well as it includes encryption.
Fair enough, and I see your point. The first computer we got arrived in my house the same year, but we used to run windows 3.1 on that thing, and it certainly didn't have different user for different members of the family :)
also, my father was so scared of me and my brother screwing up, that, when given the opportunity to have Windows 95 he installed it on a different partition and we dual-booted windows 95 and 3.11. Me and brother were only allowed on dos/win3.11. I doubt it was a sound idea, technologically-wise,but at the time I was like 8-9 years old, and was mostly interested in starting doom 2
So that was how my father implemented users at the time.
I've worked places where the roles of people was more important than the people, themselves. So if you were performing a particular role you'd sit in a different seat/at a different computer. We all had out specialties, but knew enough about each other's work to pop in and out.
This is often the case in time-sensitive operations, or when the computers are merely controlling something else larger or more important (mining operations machinery, steel mills, television stations, etc...)
In a more traditional office environment, your computer is your own; like your desk or your red stapler. IT can log in to a shell remotely. This is the scenario that Haiku seems to be focused on.
One is not necessarily better than the other. Just different. Different environments require different methods.
On the other hand, when I let someone else use my PC (wife/child/friend), I would like them to run in another environment. Also when trying random downloaded applications. In this scenario, multi user makes sense.
Multiuser is on the agenda for Haiku R2, and seeing how well they resolved the package management system, I expect multiuser R2 will be a thing of beauty.