This is the first installment of Think Secret’s “Inside Panther” series, covering Mac OS X Version 10.3. Check back to their site frequently for additional Panther coverage. Today, they begin by taking a look at updates to the Finder and System.
Inside Panther: A Look at the Finder and System, Part I
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
How exactly is Finder a media application, qualifying it for the Brushed Metal look? Oh, and everyone bitching about Linux apps looking different from each other can shut up right now
My browser (Konqueror) and my file manager (Konqueror) looks exactly like my email client (KMail) and my IDE (KDevelop). C’mon Apple! Where is the superior integration Macs are famous for?
> Oh, and everyone bitching about Linux apps looking different from each other can shut up right now
Not really. I did note that Apple has the consistency issues you mentioned, in my UI review a few months ago. But Linux as a desktop has these issues as well. At least for Apple, all their UIs look good.
They *still* haven’t modified Finder to allow “cut”ting of files. So, moving files in complex directory structures is still a PITA process involving multiple windows open or tedious and annoying drag/drop operations.
Expose looks *very* useful, however Can’t wait for KDE, GNOME and Windows to copy it.
How exactly is Finder a media application, qualifying it for the Brushed Metal look?
I agree. Apple has clearly decided the brushed metal look should be used for more than they originally intended. My guess is the original HIG guidelines are an excuse for the discontiguous brushed metal applications existing in the first place, and now Apple has decided that look should be extended throughout the system.
C’mon Apple! Where is the superior integration Macs are famous for?
I don’t particularly like the brushed metal look, but the more Apple integrates it across the system the less it bothers me. Apple’s use of brushed metal doesn’t particularly bother me now, and I think they don’t revise the HIG because 3rd parties who horribly abuse the brushed metal look, like this:
http://perversiontracker.com/archives/000070.html
I think Apple doesn’t revise the HIG to keep application developers willing to respect the HIG from abusing the look.
Looking at some of the newer screen shots of the Finder, it actually looks pretty nice:
http://www.thinksecret.com/cgi-bin/pic.cgi?i=/archives/images/panth…
Combine this with the fact that it’s a Cocoa application (which means CocoaGestures will now work) and is fully integrated with the kqueue mechanism for things like live queries, it looks like most everyone’s Finder woes will soon be gone.
Seems to me that Apple is using brushed metal in all of its “iLife and inspired” apps. Now that the Finder is more like iPhoto/iTunes, it makes sense to sport a new look. I agree that Apple isn’t following its own interface guidelines, but then again, is it that big of a deal? The UI is the same in terms of usability. It just looks a bit different.
Jared
> Expose looks *very* useful, however Can’t wait for KDE, GNOME and Windows to copy it.
To have a direct clone of the feature, the way it looks and works on OSX, you need 3D support (a’la QuartzExtreme). And KDE/Gnome on Linux/Unix don’t run on a 3D system, they run upon the traditional 2D X11. So, it will take a few years before you see this feature on your favorite distribution.
Yeah, true Apple seems to be changing their own guidelines as they go. I guess it has to do with the move (as of the G5’s) to that aluminum case on most of their hardware? I don’t know. Like maybe they’re thinking that a consistency between the software and hardware’s look is somehow warranted.
Anyway, the very word guideline implies that it isn’t written in stone. I think they can certainly change their own guidelines over time if they think they have a reason to do so.
I personally don’t hold as much disdain towards the brush metal as a lot of Mac users. The stripes of the Agua look can hurt my eyes sometimes, so in that sense the brushed-metal helps, but I would like to see this look evolve beyond what it is.
I’m excited by the new finder, it does seem to be better organized.
>so in that sense the brushed-metal helps, but I would like to see this look evolve beyond what it is.
I agree. I have trouble reading the Safari tabs sometimes. Grey-ish/black text upon dark grey-ish background, is literally unreadable.
Hey, you took away the smiley when you quoted me! I was clearly speaking tongue-in-cheek when I made the Linux crack
Anyway, my main point is I’m skeptical of the “slippery slope” that Apple is heading down. A lot of Mac users believe that OS X is a step backwards in the usability department. From my limited exposure to both OS X and OS9, I’m inclined to agree.
PS> What is this “Expose” feature? I can’t find any info about it on the page.
>What is this “Expose” feature?
Check the Steve Jobs WWDC video, it is there…
And a more static version: http://www.apple.com/macosx/panther/expose.html
I suggest you watch the video about it though, as it is difficult to visualize the feature without seeing it in action.
Ah, this one should be enough to see it in action: http://www.apple.com/macosx/panther/exposetheater.html
It uses QExtreme to zoom in/zoom out the windows on the fly.
It’s not the flashy zoom-in/zoom-out part of Expose that makes it useful, it’s the way it best-fits all your open windows onto a single display and lets you pick one to make active. Quartz allows that to look pretty cool, but it isn’t integral to the functionality.
They also picked a horrible default key to use for it – F9. I hope that gets fixed before release (it really needs to be something on the left hand side of the keyboard). At least it’s configurable (you can even set it to a mouse button).
Yeah, that is true Eugenia, text on that dark grey is a problem too.
Did they make the striations (is that the word for it?) in the brushed metal subtler recently, I think I noticed they pulled it back a little as of Jaguar. I’d like to see that grey a bit lighter or something – more contrast would help. Doesn’t bode well for the vision-impaired.
Not to stray too far from the topic, but it’s also kind of weird how dropdowns or pulldown menus within a brushed metal interface carry the standard stripe background – just strange is all.
To have a direct clone of the feature, the way it looks and works on OSX, you need 3D support (a’la QuartzExtreme). And KDE/Gnome on Linux/Unix don’t run on a 3D system, they run upon the traditional 2D X11. So, it will take a few years before you see this feature on your favorite distribution.
DIRECT CLONE? You don’t have to do a direct clone, the idea AFAIK is to show the user a overview of all windows. So in X you have to write an application which renders all open windows as small images and reacts on shortcuts/mouse->edge events which brings you to a separate (normaly hidden) “workspace” where all the small images are shown, clicking on one of them brings you to the desired window. voila! I am currently working on a detailed draft for that (for GNOME).
Offtopic: Hey Eugenia, I am going to greece for three weeks :-), do you know maybe some nice places? *me loves greece*
It’s too big. It’s got too many stuff in it. It’s the same as the one from Windows XP. IMO there is no need for Back/forward buttons(we’re not browsing web pages) and views buttons. And *any* file content options should belong in the Preferences menu item.
A good Open/Save dialog box should be just a little bigger than a message box. With not more than 2 or 3 widgets(excluding the Open/Save/Cancel buttons).
Actually all you need is the ability to get a pixmap of a window area without having to raise it. This is also necessary for transparent windows, which is on the XFree 5 TODO list. Hopefully they’ll get it in (it requires driver changes, which is why it’s been put off for so long).
Once you can get this information, it becomes a matter of grabbing each window and uploading it to texture memory, at which point you can do Expose style layout.
However, I seriously doubt that feature would be useful. Most Linux users have very few windows open thanks to virtual desktops, and when they do they are often maximized. Between having virtual desktops and a real window list/switcher, even though it’d be not that hard to implement on XFree 5 (you can switch into and out of 3D mode very quickly), it wouldn’t match the workflow patterns of your typical desktop user.
A better application of eyecandy would be slick but fast virtual desktop switching, aka 3D Desktop (but it still needs some work before being really usable, at least for me).
To those asking about the HIG, I’d say yes, it does matter that Apple are violating it. One of the Macs biggest selling points has been the “usability” factor, which was reinforced by a strong HIG. If Apple are now ignoring it in favour of shiny things and worse retrofitting it to make them look less guilty, what kind of message does that send to developers? That shiny things are more important than consistancy? That cannot have a good effect on the platform in the long term.
>> Most Linux users have very few windows open thanks to virtual desktops, and when they do they are often maximized.
Don’t speak for other users please.
>you need 3D support (a’la QuartzExtreme)
Quartz Extreme is not a specifically 3D graphics engine. It works in the style of “Display PDF”.
To those asking about the HIG, I’d say yes, it does matter that Apple are violating it. One of the Macs biggest selling points has been the “usability” factor, which was reinforced by a strong HIG. If Apple are now ignoring it in favour of shiny things and worse retrofitting it to make them look less guilty, what kind of message does that send to developers? That shiny things are more important than consistancy? That cannot have a good effect on the platform in the long term.
Well, human interface guidelines aren’t that strict you know. Consistency isn’t automatically usable, and usable isn’t automatically consistent. Don’t follow those rules blindly, but apply some common sense to it. Does brushed metal and aqua make applications unusable? No.
Consistency is meant to help users learn new applications. If a user knows how to work with iTunes and the Address Book, he won’t have any problems with the Finder. That is what consistency is about. Heck, in my opinion they improved consistency with the new finder.
The inconsistency in Linux isn’t in the look. It’s allot deeper. If I click on a link in a Gnome application, what will happen? If I click on a link in a KDE application, will the same thing happen? Will I find the “Preferences” menu in the same place in KDE and Gnome? Can I see the same fonts in all the X11 applications? Use CUPS? Save a file to a samba network from in an application, regardless wether it is KDE, Gnome or OpenOffice? That is the inconsistency of the Linux desktop, it’s not about the pretty themes like BlueCurve.
Surely the Finder gets a brushed metal theme because, as an application, it provides interaction with yuor hard disk.
isync –> mobile devices
iphoto –> digital camera
safari –> servers
finder –> hard disk
Not that hard to see now is it?
The HIG specifies peripheral devices, I believe. Your reasoning would be correct if there was an application to interface with external hard disks, but that’s really a redundant issue.
Most Linux users have very few windows open thanks to virtual desktops, and when they do they are often maximized.
I don’t know what Linux users you watch, but most *unix* users I know run massive resolutions (1600+) and have dozens of windows open at once (usually xterms) over multiple virtual desktops. I suspect they’d find Expose (or functional equivalent) rather useful – I know I will.
I don’t care about Mac OS X
Why should I bother with a non-free OS when they are
free (as in speech) ones available.
For one, you know what is going to be available in your “free” OSes in a couple of years.
I like the basic idea f the left panel. In the basic shots of the Finder, there is a huge amount of space in the metallized area at the top where the toolbar used to be.
For one, you know what is going to be available in your “free” OSes in a couple of years.
Yes, but the reverse is also true.
Sorry for my last post, I was just kidding after having
seen yet another “not ready for prime time” BS.
I stop the pollution of this forum right now.
Expose looks *very* useful, however Can’t wait for KDE, GNOME and Windows to copy it.
I wish there was a way for developers to protect these little “features” that they invent. Cause it seems like most GUIs are relatively the same (mainly KDE, Windows, and Gnome). This would force developers to be more inovative, and would create more unique operating system GUIs….I think…
I wish there was a way for developers to protect these little “features” that they invent. Cause it seems like most GUIs are relatively the same (mainly KDE, Windows, and Gnome). This would force developers to be more inovative, and would create more unique operating system GUIs….I think…
patents for the feel, trademarks for the look
Yeah, so do I (well, gnome-terminals, same thing). However, they generally aren’t overlapping to any great extent or minimized, which is what Expose would be most useful for.
I wish there was a way for developers to protect these little “features” that they invent. Cause it seems like most GUIs are relatively the same (mainly KDE, Windows, and Gnome). This would force developers to be more inovative, and would create more unique operating system GUIs….I think…
But if only Macs can use Exposé-like things, then they might just settle with that and try and use that as their main selling point for five years. Whereas if everyone can have it, then Exposé isn’t a selling point and you need something else.
I’ll think it’s right to protect ideas like apples when swapping ideas and swapping apples has the same result.
I understand that QE can leaverage 3D accerleration for fast/smooth scaling of geometry, but to mimic the functionality of Expose we don’t necessarily need to animate it. The difference is in pure functionality (unscattering windows) versus making it look cool.
I don’t know the inner workings of Linux window/display managers, so I can only wonder if there’s an object that owns all windows, even applications’ main window (I’m thinking kdesktop for example). If there’s such at thing, then it might be able to move/resize all windows on the desktop. First it takes a snapshot of every window’s current position and size, so it can restore it later, then it figures out how to lay them out in a thumbnail fashion and allocates the screen regions (by that I mean determine what the rectangular dimensions will be) and then assign each window to the allocated regions and send move() and resize() commands to each window.
Your desktop becomes a large MDI-like application where each running application with a main window is a sudo MDI child window. With regular MDI apps, we can tile or cascade windows. What Expose does is just a fancier way of tiling windows but not in a MDI app, it’s the whole desktop.
“A good Open/Save dialog box should be just a little bigger than a message box. With not more than 2 or 3 widgets(excluding the Open/Save/Cancel buttons).”
Personally I don’t think that a good open/save dialog box exists. To me they’re a pointless throwback to singletasking OSes, where you couldn’t access the file manager while in an app. When you have a fully featured file manager that can be accessed at any time, why have a crippled mini file manager for saving and opening files?
You can open files using drag and drop, all it needs is a way to save files without a dialog and they’re totally redundant. In RISC OS you saved a file by dragging an icon to a folder, personally I found that much faster and more elegant than navigating a save dialog.
But if these dialogs have to exist, crippling them even more is a terrible idea. If they’re little bigger than a message box, how am I meant to see the file information, or even read a long filename? I’d need to switch to the Finder to see the existing files and folder hierarchy, then switch back to the save dialog to save the file.
Making the dialogs more like the main Finder is a very good thing. Even if you don’t need the extra features it makes the UI more consistent, it means the user doesn’t have to learn two different file managers.
You can open files using drag and drop, all it needs is a way to save files without a dialog and they’re totally redundant.
In OSX, you can click and drag & drop the icon in the title bar of a window. I never tried but that’s one possible way of saving files, if they have/will code it that way.
I think it would be sensible to assume that Apple is currently revising the standard.
It appears that all of the little apps that are included as part of OS X (Finder, calculator, browser, mail, address book) are going to Metal. That may not agree with the current HIG document but it is hardly inconsistent.
I don’t mind Apple changing the HIG as they go for usability reasons. This is because as time proceeds, applications (and thus the UI) becomes more complex. No one in the 80s would have imagine the complexity of applications today.
However, what Apple is doing is not following the HIG. And Apple changes some parts of it for no usability reasons (please, someone, tell me what usability gained with brushed metal?) In addition to the fact that most, if not all, of Apple applications go against the HIG at the most basic points.
[i]I agree. I have trouble reading the Safari tabs sometimes. Grey-ish/black text upon dark grey-ish background, is literally unreadable.[i]
LOL
So how do you read this site?
>> Most Linux users have very few windows open thanks to virtual desktops, and when they do they are often maximized.
Don’t speak for other users please.
He did not say all rather most. i use virtual desktops. And on both Windows and Linux, I keep all my windows maximized (not for managability purposes per se, rather the fact that the monitors are small…). Sure, I can live without it, but nontheless I still use it.
And I normally have very few windows open (Opera, baby! 🙂
rajan r,
If I recall from past discussions, you were never one that cared for UI consistency. Why are you getting on Apple for it?
(Unless it is YOU that is just being consistent, and are doing it for the sake of… well, getting on Apple for it.)
Ah, this one should be enough to see it in action: http://www.apple.com/macosx/panther/exposetheater.html
It uses QExtreme to zoom in/zoom out the windows on the fly.
Since this is the second time you’ve mentioned it I thought I should respond.
Quartz Extreme is the graphics acceleration. All it does is make Quartz faster.
Quartz Extreme was an innovation of sorts, but the technology underlying it all that allows these things to happen is Quartz.
Expose could have been written for 10.1 it just would have been slow.
From the screenshots, I can see, one of the beauty of Expose is that it shrinks the window contents so it would fit the screen arranged with other windows. I can’t see how this is technically feasible on XFree86 ot X11 for all that matters. Why? The content of the window is in the appication’s hands, not the window manager.
EXACTLY. I was going to say the same thing.
Its the fact that the interface is vector-based which allows this feature to occur. I don’t see how it could be mimicked until other OSes make a rapid departure from bitmap graphics like Apple did.
They also picked a horrible default key to use for it – F9. I hope that gets fixed before release (it really needs to be something on the left hand side of the keyboard). At least it’s configurable (you can even set it to a mouse button).
But if I set it onto a mousebutton on a mac I used up all my available mousebuttons!
(and for the ususal dull “buy a Logitech mouse” answer: if I buy a laptop I don’t carry a mouse around, k thx)
EXACTLY. I was going to say the same thing.
Its the fact that the interface is vector-based which allows this feature to occur. I don’t see how it could be mimicked until other OSes make a rapid departure from bitmap graphics like Apple did.
It is not just bitmapped graphics that make the difference because bitmaps can be scaled.
What really makes the difference is that in Windows (not sure about GNU/Linux) only the visible portion of the window is buffered by the OS. The non-visible parts get a paint message when they become visible.
Quartz on the other hand actually does store the full view of every window whether it is visible or not. This does require more memory but it means you can do fancy things with the window without requiring the application to do any work.
In RISC OS you saved a file by dragging an icon to a folder, personally I found that much faster and more elegant than navigating a save dialog.
If you use X, you might want to look at ROX-Desktop at http://rox.sf.net/ which includes a library for GTK2-based apps to allow them to use drag-and-drop save. Unfortunately, not many apps are written to take advantage of it. But if you’re a coder…
Ok I have another crazy idea 🙂
In Windows/Linux you can press ALT+PRINT_SCREEN to take a snapshot of just the active window. This creates a flat bitmap of the window and it’s contents. What if the window manager (or what ever object that would be in charge) did that for each window then used the snapshot bitmap and scaled it, then render the thumbnail images and arrange them on the desktop? As for the actual windows themselves, they all receive a message to hide so all you see is the tiled thumbnails. When a thumbnail pic is click on that window is restored, or all windows are restored and the one selected will be brought to top.
DJ Jedi Jeff wrote:
> Expose could have been written for 10.1 it just would have been slow.
Or atleast slower, I have heard that exposé works fine on a mac without QE. It might not be as slick but it is usable and useful.
> If you use X, you might want to look at ROX-Desktop at http://rox.sf.net/ which includes a library for GTK2-based apps to allow them to use drag-and-drop save. Unfortunately, not many apps are written to take advantage of it. But if you’re a coder…
Well, actually drag-and-drop save almost works on os x too.
I guess it needs to be a cocoa application but test it with TextEdit.
You can drag the icon next to the filename, unfortunately the feature is quite buggy and it is impossible to make the initial save with this feature.
ROX rocks!
>Hey Eugenia, I am going to greece for three weeks :-), do you know maybe some nice places?
Email me about it.
” In Windows/Linux you can press ALT+PRINT_SCREEN to take a snapshot of just the active window. This creates a flat bitmap of the window and it’s contents. What if the window manager (or what ever object that would be in charge) did that for each window then used the snapshot bitmap and scaled it, then render the thumbnail images and arrange them on the desktop?”
For some reason there’s some notion that we can’t allow Apple to have all the fun… This is ridiculous. Rather than copy Apple’s idea, they should try to make a new/better idea and help the computing world progress even further.
“A good idea is a good idea, regardless of who came up with it”
I don’t care about Mac OS X
Why should I bother with a non-free OS when they are
free (as in speech) ones available.
Because a “free (as in speech)” OS can’t run applications like this:
http://www.propellerheads.se/products/reason/reason25/frame.html
http://www.steinberg.net/en/ps/products/music_production/cubase_sl_…
http://www.apple.com/finalcutpro/
http://www.quark.com/products/xpress/
http://www.adobe.com/products/illustrator/main.html
If the only use for your computer is IRC and web browsing pretty much any OS will fit the bill…
Is anyone else tired of the “Why should I pay for an operating system when you can get one for free?” trolls
“A good idea is a good idea, regardless of who came up with it”
But if an entity is taking more good ideas than contributing, I think we’d all agree that it doing the computing community an overall disservice.
“A good Open/Save dialog box should be just a little bigger than a message box. With not more than 2 or 3 widgets(excluding the Open/Save/Cancel buttons).”
Which is why there is a button to display a simple view which only displays current location, a field for a file name, and save/cancel buttons.
You can toggle back and forth between this simple view and the expanded mode. Prior to the changes, this was still possible.
Have you used OS X?
First of all, I think that whoever pointed this out first was quite right… Consistency doesn’t meant that everything has to look the same. I have the feeling that Apple is making a split between applications simulating real devices (compare the web browser to a TV set) and generic applications like document handlers. As long as all the interface elements still work the same way and you can easily learn the differences between a Metal and an Aqua application, I don’t see nothing wrong with that.
It is just reality that not every possible computer program has the same requirements, so developers should be free to make their application look unique in the name of usability. Beeing consistant with other applications should only be excercised when their actually is another application that performs a similar task. In my opinion.
The look of interfaces is really overrated actually. No computer user looks at a Gtk or Qt application and wonders how to use those scrollbars or menus because they look different.
Equal looking controls are only for looks, not for usability. Equal behavior is for usability.
As for maximized windows, expose and such, I think this topic is rather interesting. Users with large resolution certainly never maximize their windows. Use 1600×1200 on a rather large monitor for a while and you will understand… It just makes no sense. Windows handling becomes a lot more convenient if you don’t have to maximize your windows and something like expose could make it even more convenient. It would just rock if I could just middleclick my desktop (for example) and all windows would smoothly sort themself to expose hidden windows. I wouldn’t need this often because I don’t tend to cover my windows, but it would be very nice to have available.
Working with maximized windows is completely different again, but I think that’s only practiced anymore by people on small monitors and of course laptops. A good window switcher like a taskbar becomes a lot more important in those working environments.
Some people mentioned workspaces, but as much as this typical X feature is usually praised, I don’t think it’s as good a solution as many are thinking. Switching workspaces is not very efficient or convenient and it’s not always easy to remember which windows is where. It becomes even more of a pain if you want to do drag and drop between windows on different workspaces.
I’m not saying it’s a bad feature and I use it constantly when I need a clean workspace but can’t close other windows, but I think it’s not a solution to the general problem of window managment. The more windows you can manage on one workspace, without loosing control, the better.
As for open/save dialogs, I agree that it should be possible to do most open/save tasks using the file manager. Especially opening files. The integrated open file dialog should probably look rather similar to a file manager window, just simplified. As a matter of fact, most popular open file dialogs are MORE complicated than a Nautilus window for example. I don’t think that’s the right idea.
It’s a bit different for saving files. What I dislike most about Windows/KDE/etc is, that the save dialog looks just like the open dialog. So I think Apple has the right idea here. I don’t think that the RISC way is perfect either. Using the icon and dropping it in a filemanager window is certainly neat, but not always the most convenient. I want a save dialog that allows me this, but also to enter a name and select a directory. Maybe not even a directory, if we enter the metadata-world, using directories might not be necessary anymore. The type would be used to put it into a category (like documents) and more metadata could be added as needed, for example the current project name could be attached.
“Working with maximized windows is completely different again…”
Actually, isn’t it just as useful? Expose will scale all windows, ALL windows, down to the necessary size that they are all visible. If you have 20 apps or docs that are all maximized, Expose is still going to EXPOSE all of them to you visually. When you select the appropriate window, they all return to their maximized size below the foregrounded window.
If anything, Max’ed window users (more likely PC users) have more use for it than those that stack min’ed apps and docs for maximum visibility in the first place (usually Mac users, particularly graphics people).
Considering that Quartz Extreme works by turning windows into textures, when Expose resizes windows there is likely to be image quality degredation, such as blurry text.
When Longhorn arrives with DCE, this won’t be a problem because windows are more than just textures.
Considering that Quartz Extreme works by turning windows into textures, when Expose resizes windows there is likely to be image quality degredation, such as blurry text.
Not necessarily. If you click and drag the resize box, only the window frame resizes, the contents don’t scale along with it silly!
Ok serious now. QE composites the window and it’s contents into a texture, the window and contents aren’t flattened into one texture then scaled. The keyword is “composite.” Window frame and background drawn first, then the contents drawn on top. Think “double buffering” back in the days of sprites.
Sorry I didn’t mean to sound like I was picking on you.
The only time fonts look ugly in OSX is if they’re not anti-aliased or if they’re small and on top of the stupid grey stripes in dialogs or menus. Font rendering is another step in the window compsiting so they won’t look bad at all.
“Considering that Quartz Extreme works by turning windows into textures, when Expose resizes windows there is likely to be image quality degredation, such as blurry text.”
And the point is to get a quick look at them, not to do proofreading while scaled down. [I’ve been complaining to Apple that I can’t read the text in windows that are beeing minimized with the genie effect even if I do it in slomo.] –That’s a joke if you didn’t get it.
“When Longhorn arrives with DCE, this won’t be a problem because windows are more than just textures.” I love how you say they are “more than just textures” because the fact is you have no idea. Everything leads me to believe that DCE is doing a late stage composition as well. But no one knows. (Longhorn disciples act as if a QE composited window cannot be interacted with–this is bogus).
Anyway, a pathetic criticism for a pointless reason.
How about an organic wood look instead of a brushed metal… Brushed metal is to gimicky.
However, what Apple is doing is not following the HIG. And Apple changes some parts of it for no usability reasons (please, someone, tell me what usability gained with brushed metal?)
Well, it is Apple’s HIG after all.
If I was a programmer and wanted to change the look of my software, I’d be within my right to do so. That’s why they don’t call the Ten Commandments the Ten Guidelines. Oh no, a religious reference – anyway…
Point is, you ask what is to be gained in usability by brushed metal? I think the question is: What is being hurt or detracted – usability-wise – by using it? It’s just a look – widgets are in the same place and perform the same functions, everything still works the “Apple way” as an user might expect. That’s usability and it’s still consistent. Unlike say the difference between opening Windows Media Player or something of that ilk, where it’s all just an unnecessary splash of cold water.
Despite the inference of focusing on interaction design usability of Human Interface Guidelines, they also describe Apple’s specific look and feel.
It is not guided solely by usability.
Despite some legit complaints, the HIG has been an inconsistent and evolving document since it was the FIRST OF ITS KIND. Somehow the attention and the criticism it receives has skyrocketed despite the fact that this is an ever present reality of the HIG.
Since this is JUST a “Developer’s Preview” and there hasn’t been a revision of the HIG to account for Panther’s release (which is unknown and in the future), can we chill with the lack of adherence? For all we know a revised HIG explicates the issue for those who simply don’t like it. (Personally, I see interpretations where it IS consistent, but let them FINISH the product and REVISE the guidelines first.)
“In OSX, you can click and drag & drop the icon in the title bar of a window. I never tried but that’s one possible way of saving files, if they have/will code it that way.”
I can export files from iTunes to MP3 format (that is also how they are stored internally) by dragging and dropping them into the Finder (ie Desktop).
It is a nib-based Carbon app.
The brushed metal Finder in 10.3 is not a Cocoa API based application.
Being a user of Cocoa Gestures, I wish it were.
Actually, isn’t it just as useful? Expose will scale all windows, ALL windows, down to the necessary size that they are all visible. If you have 20 apps or docs that are all maximized, Expose is still going to EXPOSE all of them to you visually. When you select the appropriate window, they all return to their maximized size below the foregrounded window.
You are right, it would be useful of course. But different.
I think the main difference would be, that when you are working with unmaximized windows, you usually don’t watch the windowlist but the windows themself, so in the rare case you “lose” an application, using expose to make the windows glide into better positions would be much faster then looking at the windowlist and trying to find the one you want.
But when working with maximized windows, you will probably look at the windowlist most of the time to keep an overview of your windows. Just clicking the item on the windowlist would then probably be quicker than to press the expose button and search the window you are looking for, unless you have a lot of windows with equal names and icons so you can’t find them easily in the windowlist (or you just have way too many windows open).
Of course it’s a bit different with OS X, considering that the dock only shows minimized images of the windows anyway. I’m thinking more about desktops with a Windows-taskbar-like windowlist.
I use multiple windows on a single desktop when on Unix. I don’t like virtual desktops. Most often, the windows I have open are ones I’m using. I need to see multiplicity of windows at the same time. I.E. programming docs and terminals. But the specific combination of windows I have on front changes. Sometims I have a browser and a terminal. Somtimes I have a browser and two terminals, sometimes I have just two terminals, somtimes just three terminals.
For this virtual desktops are not usefull. I’d love an expose like feature. I really don’t think all Unix users use virtual desktops. Especially now that more ‘regular’ users are trying Linux.
And really, regardless of whether it is usefull for everybody or just a few people, expose is a cool feature. Anyway you look at it, it’s a extra point for Apple. I really doubt Xfree will be able to do this for a looong time.
But if an entity is taking more good ideas than contributing, I think we’d all agree that it doing the computing community an overall disservice.
Generally speaking… No, I don’t agree. Taking ideas from someone else doesn’t hurt the community, unless you think that nobody would ever come up with new ideas if other people can use them.
Windows has certainly taken most interface ideas from Apple (and Apple from others before), but yet Windows brought many new ideas to the computing world and so did free desktops.
If OS wouldn’t be allowed to copy from each other, it would hurt consistency, compatibility and the general user who just wants the best system, integrating all the goodies.
Expose is a great example for something that may NEVER be legally patented or we will be in deep shit. Because what we are talking about is just a method to tile windows temporarily until one is selected. You don’t need to be a genius to come up with this idea and just because Apple are the first to have the technology available to make this work well, doesn’t mean that nobody else should ever be allowed to do something similar.
Besides, it is only the large software companies who can really effort such patents and how would you feel if companies start to patent every freakin’ idea, even those that aren’t technically possible yet but might be in a few years?
However you turn it, patents on methods are just horrible. Imagine someone would have a patent on tabs in user interfaces. Or popup menus! How about a patent on progressbars? Don’t laugh, such a patent EXISTS in Europe and it might become valid if patents are declared valid in Europe. Have fun then!
Sorry for this rant but I get a little bit worked up when someone is actually asking for method patents.
See here for another fine piece of patent that would screw the Gimp and other applications pretty badly:
http://swpat.ffii.org/patents/samples/ep689133/index.en.html
“Just clicking the item on the windowlist would then probably be quicker than to press the expose button and search the window you are looking for, unless you have a lot of windows with equal names and icons so you can’t find them easily in the windowlist (or you just have way too many windows open).”
This is exactly how it MOSTLY is for me ALL THE TIME on WINDOWS. I mostly have IE windows and Access windows (which spawn mutliple icons for each window and dialog — yuck!).
“(or you just have way too many windows open)”
There is no such thing. If it affects performance, then yes, but that’s a hardware issue. In terms of UI, if anyone ever tol me, you just have too many windows open for the features to assit you, I would have to laugh in your face. It’s standard operating procedure for me to have 20 or more windows open, and I wish I could handle a lot more. Now maybe I can.
I don’t see how alt-tabbing, using the task bar, etc… is easier than a visual representation when I have ten IE windows open (eaching having the same icon and a long file name — or worse a file address name or query string) and/or ten different Access icons representing different tables and queries (again with names that are hardly distinguishable). Once you are dealing with the same icons or similar ones and/or long or meaningless file names, existing methods aren’t good for navigating windows.
Features can be used many ways, but I would say a definite rule is: if a window is easiest to find based on its visual contents, Expose will be the fastest method of reaching it if your environment is relatively complex (multiple windows, apps open).
In fact, a corollary to this is: if you know the location on the desktop of the window you want, Expose can be the fastest method of acquiring it if your environment is relatively complex (multiple windows, apps open). (Since Expose provides the visual cues of moving in and out from their location — you may see a window in the “right spot” before you even glance at its contents.)
I’m not digging at you. I think you are conceding it is useful. What I would point out is: this is another tool on top of the methods of navigating you discuss.
But your perspective is particularly skewed to the application window metaphor… Needlessly so…
“you will probably look at the windowlist most of the time to keep an overview of your windows.”
And? Twenty icons with text cut off doesn’t really provide clues as to what I’m looking for…
“Just clicking the item on the windowlist would then probably be quicker than to press the expose button and search the window you are looking for”
What if your task bar collapses and I have to look for an icon, select the group, scroll through the list for the file name I am looking for… What if I don’t even want to move to a particular window, but I want to compare the status of three different windows?
“unless you have a lot of windows with equal names and icons so you can’t find them easily in the windowlist (or you just have way too many windows open).”
If you have a set up that is just slightly complex (complex enough to require more than one method of navigation), the odds are high that this is EXACTLY the situation you have.
“Of course it’s a bit different with OS X, considering that the dock only shows minimized images of the windows anyway.”
How is it different? YOu were specifically talking about programming yourself to know where min’ed windows are because you need to use the windowlist as a constant reminder… I find I can hardly every read the text labels in the taskbar, they are useless. You are talking about clicking on a button, what’s wrong with scrubbing it for a text label?
“I think the main difference would be, that when you are working with unmaximized windows, you usually don’t watch the windowlist but the windows themself”
Here your own predispostions really show. I am a Mac user but spend more time on the PC. I prefer my Mac spatial orientation. I work with many apps and docs at the same time. I don’t organize windows and JUST focus on them. I focus on their arrangement, the status of open apps (and minimized docs) in the Dock, use the application switcher fequently, hide-and-show apps, use the Windows menus within apps and to navigate through dirs, AND I stay aware of these different methods SIMULTANEOUSLY and AT ALL TIMES.
Expose will be an ADDITIONAL and OPTIONAL method.
“so in the rare case you “lose” an application”
Yes, it is rare. I lose track of windows when they’re max’ed. I have fewer cues to where it is.
“I’m thinking more about desktops with a Windows-taskbar-like windowlist.” Yes, you are. Unfortunately.
“(or you just have way too many windows open)”
There is no such thing. If it affects performance, then yes, but that’s a hardware issue. In terms of UI, if anyone ever tol me, you just have too many windows open for the features to assit you, I would have to laugh in your face.
Hmm I once thought the same, but today I would disagree. There are very good reasons why having loads of windows open at the same time is bad. The more windows you have open, the longer it will take to switch between them, that’s a fact. Better window managment can improve this, but it will always be like this.
Practical reasons to have 20 windows open at the same time are pretty rare. Most include either startup time (don’t want to wait for an application to load each time) or some irritating login procedure (for example logging in to a remote server via SSH in a terminal). Those issues should and can be fixed. Memory caching for example makes re-opening applications lightning fast, so this is already not a big issue anymore. I also notice, the more usable the application becomes that I use, the more likely I am to close them when I don’t need them anymore. jEdit (one of my favorite applications) is a very _bad_ example for both of this. It starts slow and is rather complicated, so I usually keep it open even if I’m not actually editing text right at this moment, just for this reasons.
I’m not saying that you shouldn’t have twenty windows open if you want to, just that it isn’t exactly the most efficient way to work in an “ideal” desktop environment.
I don’t see how alt-tabbing, using the task bar, etc… is easier than a visual representation when I have ten IE windows open (eaching having the same icon and a long file name — or worse a file address name or query string) and/or ten different Access icons representing different tables and queries (again with names that are hardly distinguishable). Once you are dealing with the same icons or similar ones and/or long or meaningless file names, existing methods aren’t good for navigating windows.
Yeah like I said, expose would help in those situations but it wouldn’t be faster than using the windowlist with only a few open windows.
This means that even with expose, you wouldn’t be as efficient as you could be with fewer windows. Many IE windows open is actually a very typical case which IMO rather leads to a problem with how IE is used. Users like to keep websites open for various reasons, but most of them are UI issues again or technical issues (loading time of websites). There are many browsers which are much less likely to make you end up with ten windows. Safari seems to do pretty well on this (from reading the featurelist) or this iRider thingy. Basically every browser with tabs should do, unless you dislike tabs.
I don’t know about this Access behavior (is this new?) but it doesn’t sound sane to me. I like what Gimp 1.3 does it in Metacity. Only the main window and document windows are shown in the window list, utility windows are not.
Features can be used many ways, but I would say a definite rule is: if a window is easiest to find based on its visual contents, Expose will be the fastest method of reaching it if your environment is relatively complex (multiple windows, apps open).
Most probably. It has the disadvantage however that you have to press an extra key and wait for the overview to appear, you don’t have it constantly on the screen for quick access. That’s a similar dilemma to the GNOME windowmenu. Basically it is superior to the windowlist, but because it only appears when you press a button, it’s less efficient most of the time. In fact, I use the windowmenu button only in those cases where expose could come in handy. It would be mucho kickass if this button would temporary tile my windows instead.
In fact, a corollary to this is: if you know the location on the desktop of the window you want, Expose can be the fastest method of acquiring it if your environment is relatively complex (multiple windows, apps open). (Since Expose provides the visual cues of moving in and out from their location — you may see a window in the “right spot” before you even glance at its contents.)
Right, that’s why I think it would be less useful if all your windows are maximized, because then there isn’t any position to know and remember.
Sorry, I can’t really follow you. Where exactly are you disagreeing with me now?
You seem to be a Mac user, so I should point out that I really don’t have a clue how well expose works on a Mac and wasn’t criticising it at all (if that’s what you are thinking), I was just commenting on how well it might work (IMO) in a Windows/GNOME/KDE-like environment. And I think it would work great, but mostly when windows are _not_ maximized or in those cases where someone completely overblows his windowlist (I know this is common amongst Windows users…).
BTW, would it be possible that someone converts this Quicktime movie into another format like mpeg? Or would this be illegal/impossible? I somewhat wonder how I’m supposed to view this and get jealous when I can’t view this on my platform.
The brushed metal Finder in 10.3 is not a Cocoa API based application.
Being a user of Cocoa Gestures, I wish it were.
Are you certain? What is your source on this? Do you have a copy of Panther yourself?
I’ve never even heard of making NIB-based Carbon applications. I believe it should be theoretically possible to use the Carbon Appearance Manager to create a textured window, but I am not familiar enough with Carbon to know if that’s really the case.
Other brushed metal Carbon applications, such as iTunes, have their own internal brushed metal renderer…
There is no such thing. If it affects performance, then yes, but that’s a hardware issue. In terms of UI, if anyone ever told me, you just have too many windows open for the features to assist you, I would have to laugh in your face.
“Hmm I once thought the same, but today I would disagree.”
Did you read what I said? If your hardware can’t handle it, you shouldn’t do it. But there is no UI argument that says you should only have a couple fo windows open. All you provided was hardware issues for why you wouldn’t open large numbers of windows. You didn’t provide a single usability/UI reason for not doing so.
“Practical reasons to have 20 windows open at the same time are pretty rare.”
You’re clearly a techie. Try putting together a 50 page brochure without having a need to view/edit 20 or more images, text files, different apps… Even if you aren’t actively editing, any visual design process can be augmented by viewing “stock” which you aren’t using at the moment, to view various revisions, and mock ups, etc…
“Yeah like I said, expose would help in those situations but it wouldn’t be faster than using the windowlist with only a few open windows.”
That’s a retardedly obvious point. Why not say if we didn’t have multitasking we wouldn’t need it at all? What you are saying is: you don’t have a use of a feature because you limit the way you use your computer. I don’t need expose on the Mac if I only have two windows open either.
However, I’m restricting MYSELF by limiting myself to only two windows. That’s regression. And definitely not a FEATURE.
“Many IE windows open is actually a very typical case which IMO rather leads to a problem with how IE is used.”
Exactly, but aren’t you the one defending the taskbar as a superior method?
“Users like to keep websites open for various reasons, but most of them are UI issues again or technical issues (loading time of websites). There are many browsers which are much less likely to make you end up with ten windows. Safari seems to do pretty well on this (from reading the featurelist) or this iRider thingy. Basically every browser with tabs should do, unless you dislike tabs.”
Actually no. Again, the way I work, I want to and need to see multiple windows at the same time. I do use tabs, but I still use multiple windows–each group is separate — I don’t want MAC NEWS with WINDOWS SITES (when I said 20 before, I meant it, 20 IE windows with 12 of them having between 5 and 10 tabs). And again, you were making an argument for the taskbar: but tabs don’t show up in the taskbar! How do I navigate to tabs now?
“It has the disadvantage however that you have to press an extra key and wait for the overview to appear.”
And the Taskbar collapses now so you have to click and hold. If you turn this behavior off, you have tiny little tabs and can’t tell what is what. If you alt-tab, you have a click. Expose shows up as quickly as the alt-tab selector. Do not see how one click is all that “disadvantageous” or much more of a “wait.”
“you don’t have it constantly on the screen for quick access.”
You work with just a handful of windows. This isn’t a feature you would use much. But I don’t want a list of every window I have open!!! That would be crazy! Again, just because it’s not how you work, doesn’t make it a disadvantage. The doc and app switcher show me every running app. That’s enough.
“Right, that’s why I think it would be less useful if all your windows are maximized, because then there isn’t any position to know and remember.” Uhh, please. What does that have to do with anything. My point: if icons and text and lists are your primary means of navigating, that is available in OS X as well. If you want to have the EXTRA methods of navigating (visual content and location), expose provides that. I don’t see how this is negated by arguing from a NARROW and LIMTIED user style.
And there’s no need to mention how ANTIQUATED a maximized window perspective truly is. IT IS EXTREMELY ANTIQUATED!
“It would be mucho kickass if this button would temporary tile my windows instead.” Probably very easy to implement, by why would a “temporary” tiling be useful? Do you know 100% the layering of every window? In this case, it would be the same effect but you’d only get a visual cue from the top most window. Expose allows you to see the window name and its contents.
“Sorry, I can’t really follow you. Where exactly are you disagreeing with me now?”
Clearly not. My point was to make observations that you weren’t. That you were making general arguments from your personal style. Not a disagreement. I was hoping for a conversation if you could follow.
” You seem to be a Mac user, so I should point out that I really don’t have a clue how well expose works on a Mac and wasn’t criticising it at all (if that’s what you are thinking), I was just commenting on how well it might work (IMO) in a Windows/GNOME/KDE-like environment.”
Well, actually as I said, I’m more a PC user but consider myself a Mac user. I see you are one of those people who talks about somethign they have no experience with. Talk about it from a limited and incorrect perspective, only to try to adopt it but with an incorrect and limited perspective of how it was originally intended. Oh well.
“And I think it would work great, but mostly when windows are _not_ maximized or in those cases where someone completely overblows his windowlist (I know this is common amongst Windows users…).”
Again, there is no overblowing a windowlist. Your argument is that I should work like YOU, not using my OWN style, and in a LIMITED fashion. Hardware is the only thing that limits multitasking. UI doesn’t. It can complicate it or make it easier… But it doesn’t constrain it in a real world way simply because you don’t do things as such.
Bascule writes: “I’ve never even heard of making NIB-based Carbon applications. I believe it should be theoretically possible to use the Carbon Appearance Manager to create a textured window, but I am not familiar enough with Carbon to know if that’s really the case. Other brushed metal Carbon applications, such as iTunes, have their own internal brushed metal renderer…”
I just poked around and found the canonical answer to at least your conjecture. In the (apparently much-debated) Apple HIG, the section on textured windows concludes with:
“To create a window with this appearance, Cocoa developers can apply the NSTexturedBackgroundWindowMask to a titled window. Avoid using a borderless window, which won’t assume rounded corners. Carbon developers can use the new window type defined in MacWindows.h.”
So there we go.
There is such a thing as a Carbon nib, AFAIK; you can use Interface Builder to design UIs for Carbon apps. The sample application ‘Converter nib’ discusses this.
The link on the Apple Developer’s site to this (compacted by the nice TinyURL folks): http://tinyurl.com/fs2c
I’ve got an idea for a better taskbar floating around in my head.
You know how Windows XP (and now Gnome and KDE) can collapse taskbar buttons for the same app into a single button with a menu when space runs low on the taskbar? I want to do that with arbitrary windows. I want to be able to select multiple taskbar buttons and then right click on one and pick “make group” to group them. I’d also want to be able to drag more buttons into that group or just rearrange the order of the buttons on the taskbar (like you can with quicklaunch/launcher buttons).
PS: While trying to figure out what the UI for this would be I discovered something. In Windows you can ctrl+click multiple taskbar buttons and then right click to get the context menu. The action you select will apply to all of the selected windows. Neat.
When last I looked, the HIG hadn’t been updated in about a year (June 2002). OS X development techniques have not stagnated in a similar fashion.
See Andrew Welch’s comments at this MacNN forum thread:
http://forums.macnn.com/showthread.php?s=0cc9c0d7f7985facee066ebf2d…
I paraphrased his comment – I have not personally used Panther.
Woah. So if I understood you right now, you flame me to crap pieces because I made one single side remark that inclines that I might find the taskbar superior to a dock in one single occasion. =)
Geez, we must be at a serious misunderstanding.
I wrote some larger response to this, but scrapped it. No point in arguing at this point.
I will try to bring it to the point:
I hate the taskbar. I believe that maximizing all windows is primitive and severally limits development of Windows/Unix based desktops. So I don’t quite understand what you are flaming me for.
I would love to get rid of it and I believe that direct access to windows (clicking them instead of some obscure index) is way superior and this is also how _I_ work. Expose would fit in perfectly into this model and I’m not at all questioning its usefulness.
I also believe that OS X has it right in dropping the taskbar and basically forcing a more logical way to manage windows on the user.
However, I can also understand that this just isn’t an option for users on low resolutions. If you have any ideas how to solve this problems or you know how OS X users cope with this problem (like notebook users who run 1024×768 or even 800×600 resolutions), then we can very well discuss this further. This would at least be interesting as opposed to a silly discussion whether foo is better than bar.
My comments about limiting the number of open windows were mostly targeted to users who run all windows maximized, which you seem not belong to (judging from your comments). Of course it still holds that having many windows open at the same time will make it more difficult to access them. Retardedly obvious as you put it! That’s why I always try to keep my number of windows down as much as possible, as long as it doesn’t make me less efficient of course. I did not use any hardware argument to back this up, I don’t know where you are taking this from. All I’m saying is, that in an ideal desktop world you would always only have as much windows open as you currently need, and not keep windows open just so you don’t have to load or open them again once you need them.
I feel not like arguing for the sake of arguing. Please consider next time that I might not be a PC nor a Mac boy and I’m very well able to notice good usability when I see it. If you are interested in discussing the aforementioned problem of users on low resolutions, then go on. If you still think that I’m trying to throw mud at your beloved operating system, then please spare me.
See Andrew Welch’s comments at this MacNN forum thread
Ugh
Sounds like the “All-New Finder” isn’t all new after all… there’s no compelling technical reason to keep the Finder a Carbon application (built with PowerPlant, even) besides the fact that they’re still clinging to their legacy codebase.
Guess Apple is lying about more than Dell’s SPEC scores
Last I knew, the Carbon APIs were still under active development.
As for your comment about SPEC scores – real world application performance is all that matters.
Last I knew, the Carbon APIs were still under active development.
By “legacy codebase” I’m referring to the Finder application itself, not the Carbon APIs. Were they to start from scratch there is absolutely no reason for them to use Carbon, and especially not PowerPlant.
As for your comment about SPEC scores – real world application performance is all that matters.
Yes, I agree completely, but application benchmarks only makes Apple’s publishing of essentially invalid SPEC scores for Dell systems seem all the more unnecessary and stupid. There was no reason why Apple even had to publish any numbers for SPEC, simply showing numbers for real world application performance are enough.
And if they hadn’t published SPEC results at all then you can bet people would’ve been swooping down on them like vultures for that as well. No win either way.
Those SPEC scores are important for the scientific workstation market. In other real world applications it’s the busspeeds that matter, not the processor.
And guess what, scientific workstations use GCC and a Unix like operating system such as Red Hat Linux. (primarily because of better portability).
It’s typical that you assume that Dell spec tests are correct, not Apples. In reality it’s Apples test that are allot more relevant.
It’s typical that you assume that Dell spec tests are correct, not Apples. In reality it’s Apples test that are allot more relevant.
This is dumb. Dell did their own tests, but they were verified by SPEC and are available on the SPEC website. Why would Dell make their own machines look slower than they really are? Unless you’re going to accuse Dell of somehow boosting the speed of their machines in a way they can’t do normally, and in such a way that the SPEC team themselves couldn’t tell, lay off the conspiracy theories.
There is, from what I have read, also no real reason for them NOT to use Carbon. Each API set has strengths and weaknesses.
The perception of Carbon as a lesser API is probably the result of the substantial number of Carbon applications which were clumsily Carbonized [or Carbonated
] for OS X. Cocoa applications initially felt more seamlessly integrated with OS X.
I agree that there is no way Apple could win with any benchmark(et)ing.
Look everyone, there are different testing methodologies. The tests submitted to SPEC’s website follow a certain methodology. This does not mean that methodology is the best. Apple used a different methodology for performing their tests. The fact that it was different does not make it invalid.
It is on this last point that many people here fail. They don’t provide reasons. They don’t provide analysis. They just preach the Gospel that all tests must be done a certain way and Apple didn’t follow what The Good Book says. Praise Jeebus. Amen.
Those SPEC scores are important for the scientific workstation market. In other real world applications it’s the busspeeds that matter, not the processor.
Okay, but this has little to do with Apple not using Dell’s SPEC scores…
And guess what,
What?
scientific workstations use GCC and a Unix like operating system such as Red Hat Linux. (primarily because of better portability).
Oh they do, do they? Well, I happen to work for an atmospheric research group. We run an atmospheric modelling program called the Regional Atmospheric Modelling System (RAMS) which is implemented in Fortran 90 and C.
We don’t use gcc on any operating system we run the model on. We use Forte 7 on Solaris and the Portland Group’s High Performance Fortran and C compilers on Linux. See http://www.pgroup.com/
It’s typical that you assume that Dell spec tests are correct, not Apples.
That is how SPEC is designed to work. Vendors compare their scores with those generated by other vendors. Vendors don’t generate scores for their competators and then compare to scores they generate for themselves. Remember Mindcraft?
In reality it’s Apples test that are allot more relevant.
No, in reality if you are running a scientific modelling program on a $10,000 cluster or computing system then shelling out another $400 for a good compiler to ensure the best code optimizations isn’t a problem.