Um. Duh. I've been saying all along that I'm not a programmer. I work in a different group entirely.
But aside from that, the question was whether the combined title bar-tool bar appearance is a third window style. It's not. It's Aqua. Check out section 13 of the new HIG.
Now where's the little X button next to the progress bar, under the spotlight menu, that lets me STOP AN INDEXING IN PROGRESS on a volume that I'm only connecting temporarily, that may contain sensitive document data?
I don't understand the question. Spotlight indices are stored on the volume itself. It's not like you're copying data from the removable volume to your system disk.
But if you want to exclude a volume, all you have to do is drag it to the privacy pane of the Spotlight prefs window.
Whatever name you choose, it is definitely one thing: INCONSISTENT.
A foolish consistency is the hobgoblin of little minds. If you're looking to Apple never to change anything ever, you're using the wrong company's products.
Also: Wailing about textured windows has precisely as much effect today as it did five years ago: none at all. Don't like 'em? Buy a PC. It's a free country.
Here's another bug for you to check off: Even if you enable the root user, you cannot drag executable files between folders in the System tree.
Do me a favor and help me come up with a reason why you would ever want to do that.
In March of 2004, we got together in Tahoe for a three-day weekend retreat, the whole staff. We had day-long sessions where we divided up into functional teams and discussed all the various ways in which we could piss off some random and utterly inconsequential Slashdotter. This was what we came up with.
Are you fucking kidding me? Like anybody in his right mind would believe for a second that you're sitting there weighing the options. "Hmm. On the one hand, system-wide search that will fundamentally change the way I work forever. On the other, a different-looking toolbar control. Decisions, decisions."
I don't want to make an enemy here, but you've hit on one of my personal hot buttons.
The core vision of the company I work for is to make IT as you know it obsolete.
Seriously. Right now, computers fucking suck. Seriously. All of them, even the ones we make. Computers are absurdly unreliable, and ridiculously hard to operate. The mere fact that we've raised an entire generation of people who think that IT is a valid career choice is testament to how we've dropped the ball for the past forty years.
We're just now -- literally, just this week -- starting to get to the point where computers are beginning to understand two vital things: inference and implication. If I e-mail a document to somebody in my address book, my computer can now infer that that document is related to that person; when I search for that person, I get that document, or vice versa. That's just the tip of the iceberg.
Servers should be entirely self-configuring, entirely self-adapting. Can you believe that just a couple of years ago, people had to sit down in front of servers and key in lists of IP addresses to enable things like print services? You had to actually sit down and tell your computer about the printer sitting next to it.
No more. Now, with Bonjour (née Rendezvous, and please don't ask) computers and services are auto-configuring. This is, again, just the tip of the iceberg.
You're probably going to hate me for saying this, but IT employees contribute absolutely nothing to an organization. They produce nothing, they transport nothing, they collect nothing. They're an expense. One we hope to render completely obsolete.
Will we still need computer repair men? Sure! We need air-conditioner repair men. We need electricians. We need plumbers. But the idea that a small business should be expected to keep an air-conditioner repair man or an electrician or a plumber on staff full time is absurd. Someday, hopefully sooner rather than later, the idea that a small business should have its own computer repair man will be equally absurd.
That's our goal. That's where we think we're headed.
Yes, that explains why companies like Apple, and even Microsoft in their own, glacial way, are innovating on a fundamental level while Linux is...you know. Not.
I'm being totally serious now: Linux is easily twenty years behind Apple. Seriously. Think about where all the attention is going: Human-user interface design. That was Apple in 1985. Today, Apple is doing no-shit innovation.
Even little things make a huge difference. Linux, being almost a file-by-file clone of Unix, is crippled by a vast and interdependent web of system watchdog services. There's init, there's inetd, there's watchdogd, there's cron, all separate and overlapping services whose job it is to start services. All complex, all in need of configuration. What did we do? We scrapped it all, replacing the whole mess with launchd. A single service with XML (meaning self-checking) configuration files.
Do you know what happens on a Unix machine if your inittab file contains garbage data? The system refuses to boot! With XML configuration files, a config file that fails to validate will simply be ignored. The system will run in a degraded state until the file is corrected.
It's stuff like that. Yes, we're doing big-time flashy innovation with things like Core Data and Spotlight. Those are no-shit world-changing things. But we're not just glomming new services onto old infrastructure. We're evolving the operating system, replacing things that are dumb with things that make more sense.
So tell me again, oh please, how Mac OS X is a 20th-century concept.
That's what NSToolbar looks like now. Check out any application with an NSToolbar and you'll see what I mean. NetNewsWire is a good example; I happen to have it running in a background window right now.
"DVD quality" my ass. I mean, the effort you put into it was neat and all, but that file looks like crap. One megabit per second is not enough at that size and with that particular codec.
Activate Dashboard, the iChat/Volume/Battery/Clock menus in the menu bar still work and pop up over the Dashboard layer.
Not correct. A click outside a widget dismisses Dashboard.
Open a.pdf in Preview, activate Dashboard, and the cursor will change to a hand as it floats over the (dimmed and unclickable) document.
Not correct. Outside widgets, the cursor is an arrow regardless of context.
Dashboard Translation widget, click the 'swap' button several times and the focus ring will flicker madly
I wasn't able to reproduce this. I don't know what you meant by "several." I clicked it 20 times. No error.
Finder, start renaming a file and the insertion caret will flicker twice on each keystroke until the name wraps to the second line
That was an occasional bug in 8A425. Are you using a pirated copy?
System Preferences/Mail, now showing the third major window style on the system (Aqua, Metal, and now Plastic)
No, that's Aqua.
Spotlight, randomly fails to index non-boot-drive partitions
Obviously not reproducible. Spotlight will not index a volume if there's insufficient free space available. We look for about 1/10th of one percent, if I remember correctly.
Your response may be "oh well, they're all minor"
No, my response is "Please stop using pirated copies of Tiger that you download off the Internet and then complaining about them."
I'm glad somebody else pointed this out. This made the rounds internally under the headline "What's wrong with this picture?"
Look, I'm not gonna criticize Microsoft for showing early, very rough code and having it look...well, early and very rough. If you go back and look at the Mac OS X public beta, or even the 2004 WWDC demo of Tiger, you'll find that our early builds differ significantly from the final releases of our products.
But the thing is...every single one of us, to a man, would be ashamed to show something like that in public. Seriously, we'd hang our heads in embarrassment.
Microsoft's position, of course, is, "Don't look at the icons or the controls. They're not important. We're demoing underlying technology." Which is fine. But that's not how we do things. If you're going to take the time to put a UI on a demo product at all, take the time to do it right. Don't just slap something on there and say, "Oh, this'll all come out before we ship." That's not fair to your product or your customers.
It's just another sign of the difference between our philosophy and Microsoft's philosophy. I don't think either one is objectively right or wrong, but I won't hesitate to tell you which one I think is better.
Calling something disastrous "a train wreck" is a long-established idiom that isn't going to just go away because a train wrecks. And frankly, I think calling it "an unfortunate choice of words" is just a big, steaming load of language-police bull crap.
I don't know what this "*nix" thing is, but if you're talking about Unix, by that definition any currently shipping operating system is Unix. That's a dumb definition.
Mac OS X has moved beyond Unix. If you can call Mac OS X and Linux by the same name, then that name has no real meaning.
Oh and you can do a whole bunch of authoring stuff I don't care about.
Sigh. Once again, our marketing department falls flat on their faces. That's exactly like saying, "I know what a Lear Jet gives me. I can taxi from hangar to hangar, and I can use the engines as the world's biggest leaf-blower. Oh, and there's a whole bunch of flying stuff I don't care about."
But I do care that software that I purchase follows the UI guidelines for the operating system it runs on.
We're obviously not losing any sleep over this. As you've already explained in detail, you'd be a dumbass to buy QuickTime Pro anyway.
That's poorly worded. I'll send out an e-mail about that and see if I can get it fixed.
Here's how it works: People who want CDs are not going to get us to mail them for free. That's just not going to happen. The $9.95 charge covers the cost, to us, of manufacturing the CD (which we already did, of course), putting it in a box with an address label on it, and shipping it. The $9.95 figure does not cover the costs. We lose a little money every time somebody mail-orders Tiger on CD from us.
I'll see if I can't get those Web pages cleaned up.
If you're really a programmer, then the problem here is that you just haven't been reading the documentation.
Spotlight consists of two databases linked by a query service. These are a content database which is a new-and-improved version of V-Twin, and a metadata database which is something entirely new.
I won't bother describing the content database. I assume you understand it already. It indexes all the content in every file on your computer. And it's fast. Faster than anything like it.
The metadata store works like this: Whenever a file operation happens anywhere on the computer, a task called "mdimport" gets kicked off. This task examines the file to see what type it is (using a database of UTIs, or universal type identifiers), then hands it by path to a tiny piece of code called an importer. That importer examines the file and populates a data structure of key-value pairs, which it then returns back to mdimport.
This data structure consists of a set of key-value pairs corresponding to attributes and values. The attributes are defined by the importers themselves. There's a set of basic attributes (kMDItemContentType and such), but each importer can define its own attributes. The attributes can be any basic data type: CFString, CFBoolean, CFDate, CFNumber, that kind of thing. The importer populates this data structure and hands it back to mdimport, which inserts the key-value pairs into the data store.
Now, here are two very important things. First, Spotlight is entirely modular. Importers can do anything at all, and they can be added to the system at run-time. Because of our UTI typing system, an importer can be written for any file type at all, including generic file types like public.text.
Second, Spotlight is entirely extensible. There's no central registry of attributes; anybody can create an attribute at any time. We use the same structure for attribute names that we use for type identifiers; basically the same structure the Java people use for packages. It's a reverse-domain-name scheme. We trust that people will subscribe to this scheme and keep from trampling each other's attributes.
Add all this stuff up, and you've got something entirely new. Not a little bit new, not just like something else, not even a little evolutionary. It's completely, entirely new.
You know, everything we do is documented at developer.apple.com. Have you ever considered checking it out so you can be informed before dismissing world-changing technologies out of hand?
You do not remember the demo right. Spotlight doesn't do that. Though it sounds kind of neat...in a way... I guess.
Also, Spotlight requires a "plug-in" to read all files. They're called importers. We include 21 of them with Tiger. They're not really linked to a specific file format. For example, we have one importer that handles all Microsoft Office file formats. We have one that handles all text formats, including stuff like HTML and XML. We have one importer that handles all image file types, from one-bit PICT files from 1984 up through OpenEXR and DNG files that practically nobody is using yet.
Yes, the Office format importer can index Open Office files. The Text importer indexes WordPerfect files. The ZIP importer archives ACE files, and practically any other kind of archive you can think of. These three importers, and a bunch more, are bundled with Tiger. And third-party importers will start appearing on Friday to cover more obscure file formats.
Don't forget our Address Book integration. If you search your computer for "John Doe," you'll get back John Doe's address book entry, of course, but also all the e-mails that you've sent John and that he's sent you, all the attachments he's ever sent you (the "where'd this file come from" metadata attribute at work), all the chat transcripts from all the iChats with John. It'll find all the files that John Doe created (the "author" attribute), and all the appointments in your calendar that include John as an attendee.
Translation: Show me all the audio files composed by Bach in the key of G#.
You don't have to type that by hand, of course; you use the Spotlight user interface to construct it. And you save it off as a saved search, and it appears in the Finder as a folder with dynamically updated contents.
Slashdot Mirror
Um. Duh. I've been saying all along that I'm not a programmer. I work in a different group entirely.
But aside from that, the question was whether the combined title bar-tool bar appearance is a third window style. It's not. It's Aqua. Check out section 13 of the new HIG.
Now where's the little X button next to the progress bar, under the spotlight menu, that lets me STOP AN INDEXING IN PROGRESS on a volume that I'm only connecting temporarily, that may contain sensitive document data?
I don't understand the question. Spotlight indices are stored on the volume itself. It's not like you're copying data from the removable volume to your system disk.
But if you want to exclude a volume, all you have to do is drag it to the privacy pane of the Spotlight prefs window.
Whatever name you choose, it is definitely one thing: INCONSISTENT.
A foolish consistency is the hobgoblin of little minds. If you're looking to Apple never to change anything ever, you're using the wrong company's products.
Also: Wailing about textured windows has precisely as much effect today as it did five years ago: none at all. Don't like 'em? Buy a PC. It's a free country.
Here's another bug for you to check off: Even if you enable the root user, you cannot drag executable files between folders in the System tree.
Do me a favor and help me come up with a reason why you would ever want to do that.
In March of 2004, we got together in Tahoe for a three-day weekend retreat, the whole staff. We had day-long sessions where we divided up into functional teams and discussed all the various ways in which we could piss off some random and utterly inconsequential Slashdotter. This was what we came up with.
Are you fucking kidding me? Like anybody in his right mind would believe for a second that you're sitting there weighing the options. "Hmm. On the one hand, system-wide search that will fundamentally change the way I work forever. On the other, a different-looking toolbar control. Decisions, decisions."
Whatever, dude.
I don't want to make an enemy here, but you've hit on one of my personal hot buttons.
The core vision of the company I work for is to make IT as you know it obsolete.
Seriously. Right now, computers fucking suck. Seriously. All of them, even the ones we make. Computers are absurdly unreliable, and ridiculously hard to operate. The mere fact that we've raised an entire generation of people who think that IT is a valid career choice is testament to how we've dropped the ball for the past forty years.
We're just now -- literally, just this week -- starting to get to the point where computers are beginning to understand two vital things: inference and implication. If I e-mail a document to somebody in my address book, my computer can now infer that that document is related to that person; when I search for that person, I get that document, or vice versa. That's just the tip of the iceberg.
Servers should be entirely self-configuring, entirely self-adapting. Can you believe that just a couple of years ago, people had to sit down in front of servers and key in lists of IP addresses to enable things like print services? You had to actually sit down and tell your computer about the printer sitting next to it.
No more. Now, with Bonjour (née Rendezvous, and please don't ask) computers and services are auto-configuring. This is, again, just the tip of the iceberg.
You're probably going to hate me for saying this, but IT employees contribute absolutely nothing to an organization. They produce nothing, they transport nothing, they collect nothing. They're an expense. One we hope to render completely obsolete.
Will we still need computer repair men? Sure! We need air-conditioner repair men. We need electricians. We need plumbers. But the idea that a small business should be expected to keep an air-conditioner repair man or an electrician or a plumber on staff full time is absurd. Someday, hopefully sooner rather than later, the idea that a small business should have its own computer repair man will be equally absurd.
That's our goal. That's where we think we're headed.
Yes, that explains why companies like Apple, and even Microsoft in their own, glacial way, are innovating on a fundamental level while Linux is ...you know. Not.
I'm being totally serious now: Linux is easily twenty years behind Apple. Seriously. Think about where all the attention is going: Human-user interface design. That was Apple in 1985. Today, Apple is doing no-shit innovation.
Even little things make a huge difference. Linux, being almost a file-by-file clone of Unix, is crippled by a vast and interdependent web of system watchdog services. There's init, there's inetd, there's watchdogd, there's cron, all separate and overlapping services whose job it is to start services. All complex, all in need of configuration. What did we do? We scrapped it all, replacing the whole mess with launchd. A single service with XML (meaning self-checking) configuration files.
Do you know what happens on a Unix machine if your inittab file contains garbage data? The system refuses to boot! With XML configuration files, a config file that fails to validate will simply be ignored. The system will run in a degraded state until the file is corrected.
It's stuff like that. Yes, we're doing big-time flashy innovation with things like Core Data and Spotlight. Those are no-shit world-changing things. But we're not just glomming new services onto old infrastructure. We're evolving the operating system, replacing things that are dumb with things that make more sense.
So tell me again, oh please, how Mac OS X is a 20th-century concept.
That's what NSToolbar looks like now. Check out any application with an NSToolbar and you'll see what I mean. NetNewsWire is a good example; I happen to have it running in a background window right now.
There's a difference between an idiom and a joke, dude.
"DVD quality" my ass. I mean, the effort you put into it was neat and all, but that file looks like crap. One megabit per second is not enough at that size and with that particular codec.
Then they shouldn't have made a huge deal out of demoing it in front of a giant audience.
And yes, I think you hit the nail right on the head when you said they don't give a shit about UI.
There's a difference between internal builds and public demos. Our internal builds are pretty hairy, too. But we'd never show them.
I just checked each of these on my machine.
.pdf in Preview, activate Dashboard, and the cursor will change to a hand as it floats over the (dimmed and unclickable) document.
Activate Dashboard, the iChat/Volume/Battery/Clock menus in the menu bar still work and pop up over the Dashboard layer.
Not correct. A click outside a widget dismisses Dashboard.
Open a
Not correct. Outside widgets, the cursor is an arrow regardless of context.
Dashboard Translation widget, click the 'swap' button several times and the focus ring will flicker madly
I wasn't able to reproduce this. I don't know what you meant by "several." I clicked it 20 times. No error.
Finder, start renaming a file and the insertion caret will flicker twice on each keystroke until the name wraps to the second line
That was an occasional bug in 8A425. Are you using a pirated copy?
System Preferences/Mail, now showing the third major window style on the system (Aqua, Metal, and now Plastic)
No, that's Aqua.
Spotlight, randomly fails to index non-boot-drive partitions
Obviously not reproducible. Spotlight will not index a volume if there's insufficient free space available. We look for about 1/10th of one percent, if I remember correctly.
Your response may be "oh well, they're all minor"
No, my response is "Please stop using pirated copies of Tiger that you download off the Internet and then complaining about them."
Actually, IT is exactly where people like that belong. I think your opinion of IT might be a little out of touch with reality.
I'm glad somebody else pointed this out. This made the rounds internally under the headline "What's wrong with this picture?"
...well, early and very rough. If you go back and look at the Mac OS X public beta, or even the 2004 WWDC demo of Tiger, you'll find that our early builds differ significantly from the final releases of our products.
Look, I'm not gonna criticize Microsoft for showing early, very rough code and having it look
But the thing is...every single one of us, to a man, would be ashamed to show something like that in public. Seriously, we'd hang our heads in embarrassment.
Microsoft's position, of course, is, "Don't look at the icons or the controls. They're not important. We're demoing underlying technology." Which is fine. But that's not how we do things. If you're going to take the time to put a UI on a demo product at all, take the time to do it right. Don't just slap something on there and say, "Oh, this'll all come out before we ship." That's not fair to your product or your customers.
It's just another sign of the difference between our philosophy and Microsoft's philosophy. I don't think either one is objectively right or wrong, but I won't hesitate to tell you which one I think is better.
There's this really cool new thing now. All the kids are doing it. It's called "amusing hyperbole."
You probably haven't heard about it because it's so new.
I think by "people" he meant "my friends and other members of my peer group."
Calling something disastrous "a train wreck" is a long-established idiom that isn't going to just go away because a train wrecks. And frankly, I think calling it "an unfortunate choice of words" is just a big, steaming load of language-police bull crap.
I don't know what this "*nix" thing is, but if you're talking about Unix, by that definition any currently shipping operating system is Unix. That's a dumb definition.
Mac OS X has moved beyond Unix. If you can call Mac OS X and Linux by the same name, then that name has no real meaning.
You mean kinda like CrashReporter?
Oh and you can do a whole bunch of authoring stuff I don't care about.
Sigh. Once again, our marketing department falls flat on their faces. That's exactly like saying, "I know what a Lear Jet gives me. I can taxi from hangar to hangar, and I can use the engines as the world's biggest leaf-blower. Oh, and there's a whole bunch of flying stuff I don't care about."
But I do care that software that I purchase follows the UI guidelines for the operating system it runs on.
We're obviously not losing any sleep over this. As you've already explained in detail, you'd be a dumbass to buy QuickTime Pro anyway.
That's poorly worded. I'll send out an e-mail about that and see if I can get it fixed.
Here's how it works: People who want CDs are not going to get us to mail them for free. That's just not going to happen. The $9.95 charge covers the cost, to us, of manufacturing the CD (which we already did, of course), putting it in a box with an address label on it, and shipping it. The $9.95 figure does not cover the costs. We lose a little money every time somebody mail-orders Tiger on CD from us.
I'll see if I can't get those Web pages cleaned up.
If you're really a programmer, then the problem here is that you just haven't been reading the documentation.
Spotlight consists of two databases linked by a query service. These are a content database which is a new-and-improved version of V-Twin, and a metadata database which is something entirely new.
I won't bother describing the content database. I assume you understand it already. It indexes all the content in every file on your computer. And it's fast. Faster than anything like it.
The metadata store works like this: Whenever a file operation happens anywhere on the computer, a task called "mdimport" gets kicked off. This task examines the file to see what type it is (using a database of UTIs, or universal type identifiers), then hands it by path to a tiny piece of code called an importer. That importer examines the file and populates a data structure of key-value pairs, which it then returns back to mdimport.
This data structure consists of a set of key-value pairs corresponding to attributes and values. The attributes are defined by the importers themselves. There's a set of basic attributes (kMDItemContentType and such), but each importer can define its own attributes. The attributes can be any basic data type: CFString, CFBoolean, CFDate, CFNumber, that kind of thing. The importer populates this data structure and hands it back to mdimport, which inserts the key-value pairs into the data store.
Now, here are two very important things. First, Spotlight is entirely modular. Importers can do anything at all, and they can be added to the system at run-time. Because of our UTI typing system, an importer can be written for any file type at all, including generic file types like public.text.
Second, Spotlight is entirely extensible. There's no central registry of attributes; anybody can create an attribute at any time. We use the same structure for attribute names that we use for type identifiers; basically the same structure the Java people use for packages. It's a reverse-domain-name scheme. We trust that people will subscribe to this scheme and keep from trampling each other's attributes.
Add all this stuff up, and you've got something entirely new. Not a little bit new, not just like something else, not even a little evolutionary. It's completely, entirely new.
You know, everything we do is documented at developer.apple.com. Have you ever considered checking it out so you can be informed before dismissing world-changing technologies out of hand?
You do not remember the demo right. Spotlight doesn't do that. Though it sounds kind of neat ...in a way ... I guess.
Also, Spotlight requires a "plug-in" to read all files. They're called importers. We include 21 of them with Tiger. They're not really linked to a specific file format. For example, we have one importer that handles all Microsoft Office file formats. We have one that handles all text formats, including stuff like HTML and XML. We have one importer that handles all image file types, from one-bit PICT files from 1984 up through OpenEXR and DNG files that practically nobody is using yet.
Yes, the Office format importer can index Open Office files. The Text importer indexes WordPerfect files. The ZIP importer archives ACE files, and practically any other kind of archive you can think of. These three importers, and a bunch more, are bundled with Tiger. And third-party importers will start appearing on Friday to cover more obscure file formats.
What else you got?
Don't forget our Address Book integration. If you search your computer for "John Doe," you'll get back John Doe's address book entry, of course, but also all the e-mails that you've sent John and that he's sent you, all the attachments he's ever sent you (the "where'd this file come from" metadata attribute at work), all the chat transcripts from all the iChats with John. It'll find all the files that John Doe created (the "author" attribute), and all the appointments in your calendar that include John as an attendee.
It's pretty amazing.
Here's a typical Spotlight query:
(kMDItemContentTypeTree = 'public.audio') && (kMDItemComposer = '*Bach*'cd) && (kMDItemKeySignature = '*G#*'cd)
Translation: Show me all the audio files composed by Bach in the key of G#.
You don't have to type that by hand, of course; you use the Spotlight user interface to construct it. And you save it off as a saved search, and it appears in the Finder as a folder with dynamically updated contents.
This is obviously just an example.
Can you do that with Grep?