I really appreciate the vision behind this project. I like the FreeBSD base system because it’s organized and understandable, but then adding a GUI gives me decision paralysis. A default UI based on proven usability studies sounds like what I want. Big bonus points for the bringing back the spatial Finder.
The 0.5 release is based on FreeBSD 12.2 and is progressing nicely. The release notes show a number of important fixes like “sudo su works now”, “fix wrong font sizes” and “fix menu and desktop on multi-monitor setups”.
But the rest of the article doesn’t have that tone.
I’m not sure what you mean: sudo su works now, thanks @crees is pretty far down on the page and not at the top of any section here. Unless you mean this article ignored the list on the release page and found a chronological order somewhere and used that.
I get the impression they cargo-cult a lot of their UI decisions to the point it feels more like a reification of neurosis; i.e. “must have global menu bar, because this specific reason” where the reason stated feels weak. (Global menu bars in the X11 world feel so flimsy because they’re just rehoisting of a menu, without any considering of if this is the “anchor” of the application and thus must be consistent, exists without a window, is consistent between windows, etc.)
The global menu bar has been proven to be superior.: Walker, N & Smelcer, 1990.
Referred to in the Macintosh Human Interface Guidelines with highest praise:
A seminal paper on why pull-down menus are superior to any other kind. Everyone who designs for the screen must read this paper. Everyone who designs for the screen must read this paper.
This paper provides empirical evidence on the superiority of global menus.
Systems that maximize the percentage of menu items with borders will have a decided advantage over other menu systems.
In this experiment the initial movement to the pull-down menus was larger that average distance required to reach the top of a 19 inch diagonal screen, yet the p&downs still significantly outperformed the walking menus. Therefore, it may be more efficient to place menus at the top of the window
Source: Walker, N & Smelcer, JB 1990, A comparison of selection times from walking and pull-down menus. in JC Chew & J Whiteside (eds), Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI 1990. Association for Computing Machinery, pp. 221-225, 1990 SIGCHI Conference on Human Factors in Computing Systems, CHI 1990, Seattle, United States, 4/1/90. https://dl.acm.org/doi/pdf/10.1145/97243.97277
What I like about the (your?) article is the notion that “real world” applications tend to be cross-platform. So in helloSystem we are not trying to get people to write applciations (beside desktop utilities) specifically for helloSystem, but accept Qt as the de-facto standard for cross-platform productivity applications. The job of the OS then is to run those existing Qt-based cross-platform applciations well and make them feel as “native” as possible. This is a fundamentally different approach from Gnome and elementary OS, which expect people to write applications speficially for those desktop environments (splintering the already small Linux/Unix user base).
I’ve read the HIG and a long while ago I skimmed your series on Linux desktop stuff. I feel like there’s a lack of coherent vision; what decisions are there feel like post-hoc justification for personal preference, usually ripped from the Macintosh, without having things that build upon each other or why those concepts worked together, instead citing literature out of context.
(With your regards to your reply to my article: Interesting idea, but the real cross-platform is Electron nowadays. I don’t know how you’ll get out of that.)
I wish you luck despite not really agreeing with said preferences (i.e. I think how Haiku does packages is better than AppImage), mostly for the bravery of bothering to actually execute.
I really appreciate the vision behind this project. I like the FreeBSD base system because it’s organized and understandable, but then adding a GUI gives me decision paralysis. A default UI based on proven usability studies sounds like what I want. Big bonus points for the bringing back the spatial Finder.
Some of it reads like sarcasm or just being mean:
But the rest of the article doesn’t have that tone.
That’s just El Reg in general; they very much aim to be the British tabloid of the industry.
Interesting. I’ve read them from time to time but never noticed anything before this article.
Maybe we should start writing our changelog in order of importance rather than in chronological order.
I’m not sure what you mean:
sudo su works now, thanks @creesis pretty far down on the page and not at the top of any section here. Unless you mean this article ignored the list on the release page and found a chronological order somewhere and used that.I get the impression they cargo-cult a lot of their UI decisions to the point it feels more like a reification of neurosis; i.e. “must have global menu bar, because this specific reason” where the reason stated feels weak. (Global menu bars in the X11 world feel so flimsy because they’re just rehoisting of a menu, without any considering of if this is the “anchor” of the application and thus must be consistent, exists without a window, is consistent between windows, etc.)
Oh well, another potemkin village.
Lots of rationale given in the 6-part article series on desktop usability.
The global menu bar has been proven to be superior.: Walker, N & Smelcer, 1990. Referred to in the Macintosh Human Interface Guidelines with highest praise:
This paper provides empirical evidence on the superiority of global menus.
Source: Walker, N & Smelcer, JB 1990, A comparison of selection times from walking and pull-down menus. in JC Chew & J Whiteside (eds), Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, CHI 1990. Association for Computing Machinery, pp. 221-225, 1990 SIGCHI Conference on Human Factors in Computing Systems, CHI 1990, Seattle, United States, 4/1/90. https://dl.acm.org/doi/pdf/10.1145/97243.97277
Reference: helloSystem User Experience Guidelines
What I like about the (your?) article is the notion that “real world” applications tend to be cross-platform. So in helloSystem we are not trying to get people to write applciations (beside desktop utilities) specifically for helloSystem, but accept Qt as the de-facto standard for cross-platform productivity applications. The job of the OS then is to run those existing Qt-based cross-platform applciations well and make them feel as “native” as possible. This is a fundamentally different approach from Gnome and elementary OS, which expect people to write applications speficially for those desktop environments (splintering the already small Linux/Unix user base).
I’ve read the HIG and a long while ago I skimmed your series on Linux desktop stuff. I feel like there’s a lack of coherent vision; what decisions are there feel like post-hoc justification for personal preference, usually ripped from the Macintosh, without having things that build upon each other or why those concepts worked together, instead citing literature out of context.
(With your regards to your reply to my article: Interesting idea, but the real cross-platform is Electron nowadays. I don’t know how you’ll get out of that.)
I wish you luck despite not really agreeing with said preferences (i.e. I think how Haiku does packages is better than AppImage), mostly for the bravery of bothering to actually execute.