“Usually, a distro’s most recognizable feature is the desktop environment. With SymphonyOS, the screenshots are the first hint that this is not your ordinary remix. SymponyOS is a bold step into a new desktop paradigm. It features such innovations as the Mezzo desktop environment, the Orchestra application environment, and the apt-plus software installation tool. Ryan Quinn, the lead developer of SymphonyOS, took some time out to talk with me about the distro and a few other subjects.”
It seems to me that they’ve taken the Dashboard/Konfabulator idea and put it a step higher. Rather than writing widgets in HTML, JS, and CSS, they’ve written an entire DE (and its applications) in HTML, JS, CSS, and Perl. It’s an interesting idea, but I can’t see much in the way of performance. Also, somebody needs to tell them that antialiased fonts are “in”. Adding a custom ROX theme (preferably matching their FVWM theme) would be a nice touch too. It appears that they’re not even using a matching Gtk+ theme!
dead link.
Link worked for me. I’m not so sure I could get used to working with that desktop.
connection refused
I’d hardly call this a _bold_ step. Croquet[1] or Archy[2] is bold steps.
The desktop paradigm using windows and having stuff behind them requires a WUXGA screen to get usable in the first place. Then having 24″ between the corners kills the benefits of fips law and corners. With this in view I don’t see how SymphonyOS improves on anything.
[1] http://www.opencroquet.org/
[2] http://rchi.raskincenter.org/aboutrchi/index.php
“The desktop paradigm using windows and having stuff behind them requires a WUXGA screen to get usable in the first place. Then having 24″ between the corners kills the benefits of fips law and corners. With this in view I don’t see how SymphonyOS improves on anything.”
The idea is that clicking on a corner button iconifies the windows so that the root window comes completely into view. I’m not completely sold on the corner idea, either. I would rather something based on viewports:
A W W W W….
D W T or A D S T
S
where A is applications, D is documents, S is system, T is trash, and W is the workspace. The first setup might be confusing for some users if the workspace is allowed to have virtual desktops. The second setup solves that problem by using only viewports (or only virtual desktops) at the expense of some spatial coherency. The second setup would benefit from some intelligent navigation controls (i.e. moving to panel row always drops you in applications no matter which workspace you were last on, and moving to the workspace row always drops you on the workspace you were last on).
Thus we can provide the desktop panel paradigm without interfering with the workspace paradigm.
My other idea is to have one panel viewport for all four panels, and use a tabbed interface.
Interesting, but I don’t see why they had to go to such great lengths to incorporate these ideas. Since they’re using Fvwm, much of what they wanted to do could simply be a customized configuration file running on any *Nix system.
“Interesting, but I don’t see why they had to go to such great lengths to incorporate these ideas. Since they’re using Fvwm, much of what they wanted to do could simply be a customized configuration file running on any *Nix system.”
I’ll try to help you understand. They are using FVWM as their window manager. Mezzo is a desktop environment. It works along with Orchestra, which provides a Gecko-based root window that supports a stack of web technologies. It also has some FVWM modules (the bitmapped button boxes in the corners) that control the automatic iconification of windows and the loading of a new “webpage” on the root window.
Mezzo and Orchestra will available for most *nix systems.
I’ve read over some of the Mezzo desktop stuff, it’s an interesting concept. It appears to basically try to apply PDA-ish UI conventions to a desktop WM. I imagine it would start to get pretty unweildy for users with large amounts of documents/programs, but I think many of the concepts it uses would be a lot more comfortable to typical home users.
There used to be another project applying PDA concepts to the Desktop … anybody rember EONE or EON … they had a DE with Filemanager, Mail Programm Calendar etc. based on Mozilla and Python running on top of RedHat, it loked very interresting but never lreally took off
I don’t think I could get used to a Desktop Environment like this. I still like my KDE/Gnome/XFCE 😛
But for someone like my grandma, whom uses like 1 or 2 programs at a time, this may turn out to be a real good choice for her.
Like GNOME, KDE, XFCE, and other attempts at putting a usable interface on Linux, I fear that this project is just wallpapering over the gaping cracks in the Linux user experience. It will work well for the two dozen or so usage scenarios that the authors have though of, but it will be a disaster once you step outside of those. For example, can I install the Windows drivers for my unsupported wireless card through ndiswrapper with just a few simple mouse clicks? No. Can I add any third party kernel module with just a few mouse clicks? No. Will I have to drop down to the command line, edit some rc file, which will break some GUI functionality? Most probably yes. All Linux “desktop environments” that I have seen so far have mostly been cute little sandboxes that instantly fall apart at the first encounter with an unexpected configuration.
and this is not for you as your a power user.
this is for the people out there that need a glorified typewriter and web/mail access.
most of those will have some relative or friend that they can call on if they need to have anything like what you said done. and they will then most likely go full screen cli and forget about the gui anyways.
but then i would say trying to install a third party kernel module or a driver thru ndiswrapper as to damned limiting. most likely you would have a dialog that would parse a gigantic list of text errors and commands behind the gui anyways.
im just glad one can reach the cli from anywhere in linux, and that its so powerfull. this means that one have a common way of fixing things when stuff get broken.
in comparison, if something goes badly broken in windows its reinsall time 9 times out of 10 as you most likely dont have to tool to fix it if you cant get to the gui.
hell, telling a recent windows to boot into cli mode basicly means booting up a stripped down gui enviroment(no explorer bar or anything like that) and then putting a cmd window on top.
ok so you have recovery console, but even that is limited if you go the route of a livecd linux
all guis are in the end sandboxes to hold the unknowing masses in their hand while they try to do stuff.
the problem is when you try to allso put the power user into the same sandbox. while the user will be happy with a bucket and spade and making simple sand castles. the power user will want his hydraulic digger and bulldozer. and dont forget the concrete and steel to build anything
im have a feel that trying to build a “one size fits all” kind of computer interface is hopeless. just like building one os that can be both a desktop computer and a cluster server without giving the admin the ability to recompile the kernel will be hopeless. the requirements are just to diffrent.
sure you can boil it down to some common stuff, but in the end your then limiting both ends of the spectrum.
Not really. It is stated on their website that it is meant for novice people, but it doesn’t do a very good job at it. Most of the time, a good user interface is usable by both novice and experts alike (with some exceptions). The fact that this interface can only really be used by novices is a big hint that it’s not that great.
all guis are in the end sandboxes to hold the unknowing masses in their hand while they try to do stuff.
The CLI is a gui, just a restricted one (it only uses 2d text, which is a graphical element). The fact that the CLI is generally more powerful than a bitmapped interface is because it’s more elegant than the current interfaces; it doesn’t mean that it’s better than all graphical interfaces, just the currently available ones. I believe that it’s possible to engineer a user interface as powerful and elegant as the CLI, and you should all know whether I’m successful in a few years.
-bytecoder
looking forward to being disproved.
and good luck with your plans.
and this is not for you as your a power user.
this is for the people out there that need a glorified typewriter and web/mail access.
This is totally against my experience with normal users. Joe User wants to do a lot more with his computer then just email and typing. He buys printers, web cams, iPods, digital cameras, and other devices. If he has to call somebody else to use the CLI and get any of these things working then the interface has failed. Every desktop user interface has to provide for an easy and foolproof way for the user to extend the functionality of his computer. This includes letting the user install drivers and modules for devices that didn’t event exist then the OS was released. If you think that only “power users” want to extend their computers then you are sorely mistaken.
I think the lack of quality in Linux GUIs stems from the distain that Linux people have for GUIs. Most Linux developers seem to treat GUIs as some shiny useless things that only morons play with. They see GUIs as limited video games instead of useful tools that are more powerful and efficient than CLIs. Linux GUIs are poor because their developers still do their “real work” under the CLI. They very rarely eat their own dog food.
with the perfect system drevers would not exist. atleast not in the normal sense.
instead the system would either use a standard protocol or upload info about how to interface with it to the os when its pluged in.
if you need a extra driver just to access a digital camera today then i would say the camera is flawed, not the os or the ui.
same goes for printers and all the other stuff.
a DAP (digital audio player, or ipod if you will) should just id itself as such to the os and then the os would tell the music app that comes with the os and you would then be able to select or make a playlist or group of songs and have it transferd over. or just drag and drop it to the removable storage device. go get a iriver or iaudio player and be enlightend as to how easy it can be done.
users that buy from lexmark or apple should be taken out and spanked for all to see. they keep pushing a corrupting idea called vendor lock-in.
give me reliable open standards any day over closed source drivers and solutions!
with the perfect system drevers would not exist. atleast not in the normal sense.
instead the system would either use a standard protocol or upload info about how to interface with it to the os when its pluged in.
Yeah, it would be nice if the system could figure out by itself how to talk to new devices (or the devices could tell the system how to talk to them). But such a system is only a figment of your imagination. It isn’t a PC running Linux. If you want to make a desktop environment on a PC running Linux, then you should make it very easy to install drivers.
“Joe User wants to do a lot more with his computer then just email and typing. He buys printers, web cams, iPods, digital cameras, and other devices. If he has to call somebody else to use the CLI and get any of these things working then the interface has failed.”
I appreciate your point, but all of the existing desktop environments solve these problems. The last three are all USB mass storage and pop up as desktop icons under GNOME or KDE (and they have for the past year and half at least). Printers are probably the hardware area that needs the most work, but most common printers are well-supported. KDE’s CUPS manager is a bit better than GNOME’s, but they both get the job done. The printer manufacturers have recently started to come on board in a big way. Epson released PPDs for all of their new printers to linuxprinting.org in July, for example.
“Every desktop user interface has to provide for an easy and foolproof way for the user to extend the functionality of his computer. This includes letting the user install drivers and modules for devices that didn’t event exist then the OS was released.”
Most distributions allow users to update their kernel as if it were a normal package. As for 3rd party modules, the only experience I’ve had was with ipw2100. I just installed the package like any other, and it made the modules available to hotplug. A couple kernels later, ipw2100 was in the mainline kernel. Have you had problems with 3rd party kernel modules?
“I think the lack of quality in Linux GUIs stems from the distain that Linux people have for GUIs. Most Linux developers seem to treat GUIs as some shiny useless things that only morons play with. They see GUIs as limited video games instead of useful tools that are more powerful and efficient than CLIs. Linux GUIs are poor because their developers still do their “real work” under the CLI. They very rarely eat their own dog food.”
You write this as if you think that the same group of people that develop the kernel develop the file manager. OSS is based on interested people hacking on what interests them, and it has been extended to include interested companies hacking on what promises a good return on investment. The KDE or GNOME developers, for example, are passionate about GUI technology. Those desktops are more consistent and more comprehensive than anything Microsoft has put together, and you get more desktop software with your distribution than you get with Mac OSX or Windows.
I do think your main argument is flawed, though. There’s no question that the PC is becoming more and more like an appliance. Peripherals plug into it, but they more or less conform to a standard. 9/10 PC periherals are USB mass storage devices. This has become a ubiquitous standard on any platform.
Printers will eventually be predominantly network devices. There will be a tremendous growth in network satellites, whether they be phones, media players, or even tradition home appliances, that use embedded linux (or some other OS) to abstract their interface over IP and open standards. And the GUI of choice for all of this will be in the browser, not on the desktop.
I appreciate your point, but all of the existing desktop environments solve these problems. The last three are all USB mass storage and pop up as desktop icons under GNOME or KDE (and they have for the past year and half at least).
Yes, but what about the Foo devices which will be released X years from now by company Bar that don’t run on USB (or Firewire or any other existing standard) because it isn’t fast enough to handle the data requirements.
Epson released PPDs for all of their new printers to linuxprinting.org in July, for example.
So can I double-click on them in Nautilus and have them automatically installed and detect my attached Epson printer?
Have you had problems with 3rd party kernel modules?
No, but I was able to install my Broadcom-based wireless card using ndiswrapper from the command line in FC4. I was able to do a iwconfig wlan0 and everything worked properly. But when I went to the GUI networking tool provided, no wlan0 interface was listed. The GUI networking configuration tool was now essentially useless bloat to me because I had a non-standard configuration.
You write this as if you think that the same group of people that develop the kernel develop the file manager. OSS is based on interested people hacking on what interests them, and it has been extended to include interested companies hacking on what promises a good return on investment.
I don’t believe that the kernel and GUIs are from the same developers. I don’t understand how you could ever read what I wrote and come to that conclusion. My point is that the people writing the GUIs don’t seem to care about exposing all the power and functionality that exists in the CLI tools and kernel underneath. They don’t make it so that when you edit an rc file from the command line, the appropriate option boxes will be updated in the GUI control panel app too (and vice versa). They just treat the GUI as this pretty abstract art project. It’s pretty to look at, but when it comes down to getting some “real work” done, the CLI is still better.
if device foo shows up then most likely its using some variant of firewire or usb. or a technology thats very similar but with a high bandwidth. hopefully the designers of said technology have learned from the usb storage device standard and have developed similar ones for their tech. i would belive not doing so would be a make or break factor for their technology on the market as people get used to only plug something in and it works, no matter the brand of computer or device.
as for exposing the power of the cli thru a gui, there is no point. it will only lead to bloated config tools, something both gnome and kde are getting ridiculed for.
its funny realy. most of the time the same people that dont want messy dialogs are the ones that want the same power in a gui as on the cli. sorry but that cant be done.
for every -i -o -p of a cli tool there have to be a check box or similar, with a descriptive text for what it does and maybe a tool tip that expands on that text.
compare that to the cli where most of the time -h or –help will give you info about every one of them. and if you want more info there is allways man appname to the rescue.
hopefully the designers of said technology have learned from the usb storage device standard and have developed similar ones for their tech.
Yeah, hopefully.
/me rolls eyes
its funny realy. most of the time the same people that dont want messy dialogs are the ones that want the same power in a gui as on the cli. sorry but that cant be done.
Yes it can be done. Windows and MacOS have been doing it for years now.
for every -i -o -p of a cli tool there have to be a check box or similar, with a descriptive text for what it does and maybe a tool tip that expands on that text.
No. The GUI dialogs should provide the same functionality as the CLI tools. That does not mean mirroring all of their command line switches.
yes yes i know that hoping dont help. but if company A didnt and bombs then company B and C will be quick to learn from it.
hell, look towards bluetooth. every revision adds more and more profiles to make sure everything works with everything. and i dont think it will change any time soon.
windows have tryed to put the power of a cli into a gui and failed. having to dig down 3-4 layers of gui just to change the ip of a network card is anything but clean.
on the cli in linux its ifconfig plus the right chain of variables. this can be done right there on the cli as long as your access is root or equivalent.
how its on mac i cant comment as i have not used it…
and the routing system on windows server is famous for being a mess. there are so many dialogs and options hidden in rightclick menus all over the place that trying to find them all, much less recall where they all are is a nightmare.
doing a similar task with a cisco box and a terminal app connected by serial cable (or by telneting into it) is just a matter of some “simple” commands.
so yes you can hide away the power user optiojns under a second layer of gui or in some registry like enviroment. but thats not the same as delivering the same functionality as a cli in a clean and simple gui.
your basicly hiding the dirty and maybe faulty pipework behind a nice wall. and pitty the man that have to knock down said wall to fix the leak…
windows have tryed to put the power of a cli into a gui and failed. having to dig down 3-4 layers of gui just to change the ip of a network card is anything but clean.
I wouldn’t say that Windows has failed at this. The GUI may not be the most efficient, but at least you can get to all the functionality using the GUI. That alone makes it better than any KDE or GNOME control panel app.
It will be interesting to see how the next MS Office release will work. They claimed to have gotten rid of these “million option boxes and dialogs” by changing the interface.
i tought we where talking about a simple gui with the power of the cli, as in being able to access all the option of the cli within a clean layout. the windows layout is anything but clean.
as for kde/gnome and network dialogs. im maybe geting to used to using the cli but last time i looked for a gui network setup for my mandrake box they had supplyed their own network gui. same deal with the last fedora i interacted with.
in both cases it was just a wrapper for the ifconfig cli tool and the config files from what i noticed. and was realy just there to allow you to set basic stuff like ip address or dhcp server and default gateway.
thing is that its about all you need when you know that the cli is just a icon away. why put every last config option into the gui when an advanced/power user will bring up the cli anyways?
as for that office gui. i dont know about the million option boxes and dialogs, but from the looks of it they have gathered all the diffrent buttons into clearly labeld groups and will most likely exchange those groups depending on what elements of the document your working on.
question is if it will have the same effect as the “intelligent” menus that they put into win2k and later…
I think the problem is that people think that only one interface device is allowed. In CLI only keyboard is used (with GPM providing some exception), and in GUI everything has to be doable with a mouse.
The truth is that booth devices is good at some things and bad at others. There is nothing wrong with having a CLI insidea GUI app f.ex.
I remeber the huge usabillity boost that AutoCAD got once I realized that I could use the CLI to do the same operations I could do with the mouse. Now creating a line with exact measurements was easier with the CLI while hitting edges and stuff was easier with the mouse.
I think a good GUI has to realize this too. In this respect I think Archy has got some fundamentals right.
Yes. You are perfectly right. I was always looking for a combination to work well.
Eg:
Using Win98 if i have to copy ji12.txt ji45.txt etc i have to locate each of them, select with click while pressing and holding control down.
If at the bottom of the explorer window i had a small CLI i could just issue the command copy ji*.txt <target>
Wish .. !!!
If you are using Windows now I can recommend TotalCommander[1]. It proveds some fetaures along these lines. And frankly, in my windows days I couldn’t imagine using the sytem with out this software.
[1] http://www.ghisler.com/
true. like say doing image editing on the pixel level may well be simpler to do with a mouse. but when wanting to do filters of some kind then hitting say ALT, then type the name of the filer and the settings you want. hit enter and see the result. then if your not happy maybe hit undo, or just type the same command but now tell to move the setting +/- from what you told it before. thats mostly what i see people do with sliders anyways, make the slider active and then move it up and down with the arrow keys.
in the end its basicly the right tool for the job.
like say how i browse in firefox these days, a combo of mouse and keyboard.
i have search on type active, so that i can look for any text on the page. but if i want to activate a link i prefer the mouse. maybe in combo with ctrl to make it a new tab. and to browse backwards i use alt+left arrow.
hmm, it would be interesting to see rachy being used for a browser or a spreadsheet
or maybe a mail program.
to: [email protected]
title: testing
body: this is a test of inline mailing in archy.
then mark the whole mail, hit alt and type send?
attach: foo.jpg, bar.zip maybe?
This looks really great, especially when talking about regular desktop users (i.e. grandma). I mean, it can’t get much easier than that without a neural interface, right
. Seriously, though, I don’t know if I would want to use this as a developer, but that’s mainly because I’ve not tried it. When this thing becomes stable, I’m definately going to be checking it out.
Great Work, Guys!
I downloaded the LiveCD of this back a few months ago. I showed it to my mother (who is a grandma), who has learned within the past year how to browse the web and write emails. She can type up a paper in a word processor, print things… the normal stuff older people use a computer for. Nothing fancy, no games (aside from the occasional web flash game) no devel. I asked her opinion of this, and she said it would be easy to get used to – much easier than her Windows XP she has now. The other few people I showed this too that used a computer about like her felt likewise. So I think this project really has potential. It is great for people who just browse the web, read emails, maybe type a few papers, that type of thing. I don’t think it is meant for geeks like us.
the great thing about linux is that it have a power user interface hidden away anywhere (and one can say the same for osx i guess) as you can access all setting for a computer using a cli.
so if your over and trying to help set up something then you just call up the cli and you dont have to work within the limits of this gui, while at the same time dont have to install a second gui for you to use while your there.
all the new OS are utilizing the rendering capabilities of recent video cards and symphonyos just ignores it and uses html/js/css and calls it innovation? come on!
all the new OS are utilizing the rendering capabilities of recent video cards and symphonyos just ignores it and uses html/js/css and calls it innovation? come on!
And transparency, “genie effects”, and drop shadows ARE innovation? What you are referring to is called “eye candy.”
If you really look at the architecture of the DE, and at the GUI layout of SymphonyOS it is quite a departure from the typical MacOS/Windows clone UIs that are currently available for Linux. That makes it innovative.
Being a long time Unix/Linux user and being somewhat of a geek, I don’t think I’d necessarily like this environment, but I’m not going to dismiss it before I try it. (I’m fetching the ISO now.) It does look like it might be a good environment for Aunt Tillie, though.
—
Signals42
OSX rip off!
Where do you see OSX rip off here???
Except same colors on title and background?
For your information before you name Dashboard like widgets. GDesklets , Karamba, Super Karamba and others existed far before OSX.
Before you mention dock, well I had dock like OSX long before OSX was mentioned in pre1.0 Gnome and Afterstep.
Navigation in SymphonyOS is completely different than anything before.
If anything then OSX is a rip off of Enlightenment, NextStep and Karamba.
I don’t see what all the fuss is about. It’s got a horrible UI (passed off as ‘revolutionary’), and horrible integration. I tried the live cd a while back, and I’d die if this was the future of computing.
-bytecoder
The battlefield for next generation Linux desktops is shaping up. The veteran Enlightenment in one corner, with its powerful C libraries, EFL. The community challenger SymphonyOS, with its Gecko renderer, Orchestra. The up-and-comming commercial contender Komodo with its impressive .NET framework, Emotion. And the not-to-be-underestimated futureware HeliOS, with its unfathomably brilliant BeOS/MacClassic feel with a completely different UI.
Did I do your project better justice this time than last?
I think what’s happening right now with the Linux Desktop is great!! Instead of the two DE giants duking it out, we have all of these little projects springing up out of the infinitely fertile development environmnet. We’re been working on all of these backend technologies like dbus, gecko, mono, cairo, xine, and the like for a long time. We have 1200 pieces to a 500-piece puzzle that we haven’t yet solved. Let’s start putting these pieces together and evenually we’ll see the picture start to take shape.
The battlefield for next generation Linux desktops is shaping up. The veteran Enlightenment in one corner, with its powerful C libraries, EFL. The community challenger SymphonyOS, with its Gecko renderer, Orchestra. The up-and-comming commercial contender Komodo with its impressive .NET framework, Emotion. And the not-to-be-underestimated futureware HeliOS, with its unfathomably brilliant BeOS/MacClassic feel with a completely different UI.
Did I do your project better justice this time than last?
Wow, it took me a while to stop laughing after that
I don’t remember you mentioning my project before, though, so I can’t comment on that last sentence.
-bytecoder
“Navigation in SymphonyOS is completely different than anything before”
actually its exactly like everything before, it’s got a program menu in the lower left corner, a trash can in the lower right. it’s got a link to a file browser in one corner and a link to system in another, that takes, oh 10-15 seconds to do in gnome or kde and probably not too tricky in OSX or windows either.
difference is they draw the menus/trashcan… to the desktop instead of to a menu or a new window.
me i kinda take issue with it.
fitts law, yeah they use it, for all of one layer/four buttons then completely disregard it, no app menus on rightclick/screen edge, all the icons are in the middle of the screen (conviently covered by windows), window controls are brought in from the corners
and windows minimizing every time you hit one of the corners, then staying minimized when you hit close this menu (which seeing as it’s drawn directly onto the destop is probably a poor choice of words) sure as hell isn’t my first choice of behavior.
“and windows minimizing every time you hit one of the corners, then staying minimized when you hit close this menu (which seeing as it’s drawn directly onto the destop is probably a poor choice of words)”
Really? I got the impression that windows would restore when the desktop panels were dismissed. If they stay iconified then that’s clearly a bug or a bad design choice.
“”and windows minimizing every time you hit one of the corners, then staying minimized when you hit close this menu (which seeing as it’s drawn directly onto the destop is probably a poor choice of words)”
Really? I got the impression that windows would restore when the desktop panels were dismissed. If they stay iconified then that’s clearly a bug or a bad design choice.”
it’s a design choice forced by using the desktop as your interaction point. if i hit files it iconifies everything, then i open a file and hit “close menu” problem is, as it is set up there is no way to indicate whether i wanted to go back to the favorite programs/recent documents default desktop OR if i wanted to go back to using all the things that are iconified.
the solution is two options a resume work option and a back to favorites option, but then if i don’t have anything open they both exhibit the same behavior (or one does nothing) which is probably counter to the manifestos desire for consistency. as well as being yet more objects that pay fitts law no mind, which is like the first point in the symphonyOS manifesto (not that i care about fitts law i want logical flow more than speed out of my buttons and widgets, but they seem hip on it)
I absolutely love the idea of being able to use HTML to design my environment.
I hate Linux as a desktop OS, and this is something I am actually interested in.
As a web developer for over 3 years, I can tell you right now that would be a nightmare; HTML is a horrible layout language. This is one of those very rare instances in which XML would actually be useful.
In any event, arguing over this is useless, because the whole idea is flawed–a good UI reuses as much as possible. To quote Antoine de Saint-Exup’ery, “A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away.”
-bytecoder
HeliOS sounds great. I love classic MacOS/BeOS. Is there a web site?
Not yet. I still have to finish the language, since I’m basically writing the entire thing in it (I’m not rewriting the kernel/CLI tools/X11, so it should be fine).
-bytecoder
Not yet. I still have to finish the language, since I’m basically writing the entire thing in it (I’m not rewriting the kernel/CLI tools/X11, so it should be fine).
-bytecoder
Are we talking about this HeliOS:
http://en.wikipedia.org/wiki/HeliOS
?
Nope. Guess I’ll have to think of another name. It was so good, too
-bytecoder
The menu layout faintly resembles the Start Page in early builds of Microsoft Whistler. Is it similar in concept perhaps?
Simon.
I love the passion from lovers and haters. New approaches always bring both. I do see a lot of misconseptions of the mockups though since there is not solid implementation of Mezzo yet and it’s hard to make behavior clear from a few pictures.
To clarify one point, the corner targets minimize everything, then once something is activated it takes you right back to where you were. With everything exactly as you left it. Exactly. This behavior is no different than a “start menu” or other menu, it just fills the screen. If you close it without activation, it takes you to exactly where you were. That kind of predictability that is essential in a simple UI. The UI doesn’t try to guess the user’s needs “Did she want to go back to one of the other targets? Did he want a certain window left up?” The state of the interface is exactly as the user left it before they clicked the target.
If they want “out” of that target choice, they have several options. 1. Click an icon (take action in the target) 2. Click close (no action taken) 3. Click the target again (no action taken) 4. Click any of the iconified windows at the bottom to jump to a different work path.
If the user wants to get to a different target, ie from Programs to Computer, then they simply click the target they wish to go to. No “back to Computer target” button, since “back” can have several meanings, and implies drill down in some cases.
The Computer target always means, the Computer target, and that’s where the user knows to find it if they want it. By giving the person multiple places to click to get the exact same thing, but not ensuring that it will always be there, you trick the user. For example, the “back to Computer” icon/button where ever you would put it, would magically become a “back to Programs” button if that’s where you came from. My rules say NO to that kind of thing as much as possible, especially for the major desktop locations.
Thanks for your interest, or dis-interest. Peace and love.
Jason
not in the alpha4-99 release. as it behaves in vmware if you hit programs everything minimizes, you pick say xterm, the desktop goes back to the favorites screen (while, inconsistantly, the file management and system desks don’t drop back to the default after a selection) xterm pops up and all the other programs stay minimized.
also i didn’t mean a general back button but a button that went from whatever you currently had back to the default desktop. which is purely for the semantic difference between wanting exit the file management desktop and have the programs all pop up versus wanting to returning to the default desktop with the programs remaining minimizing (which assuming the windows actually did maximize like you claim they should becomes an issue). and yes i already said that it more or less violates your cute little manifesto.
Well, to help you out, I guess I’ll post some concrete reasons why I don’t like it:
First off, there’s way too much context switching. This requires a complete change in thought, and should be avoided like the plague.
Secondly, extra abstractions are bad. Application menus, and abstracted lists in general, are horrible because they look like the applications on your computer, but you can only perform one of the operations one would normally do with applications–execute them. You can’t move them, you can’t delete them, you can’t install them. Instead, you have to use completely unintuitive, separate interfaces because it was easier to program that hack than create a useful tool. That sucks.
Finally, I’d like to point out that it seems kind of slapped together at the last minute. It lacks quite a bit of integration, and just seems poorly thought-out. The four-corners idea was ok, but you didn’t even extend it to allow for putting things on the sides of the screen, etc. In my opinion, the best way to save this project is to treat this as a mockup, learn what you can from it, and try again.
-bytecoder
BTW, I’d like to see an OS implement radial menus. Much nicer than normal menus, though harder to implement.
Does anybody have any good names that they wouldn’t mind me using?
-bytecoder