Apple-X discussing the authenticity of the Panther shots:
Quote:
I have seen a variety of people immediately assume the images are fake because of the copyright date on the "About The Finder" window. First off I would just like to clarify for you all that Apple began working on Panther back in November of 2002. According to our source the images are said to be from an early build, how early he did not say. Also, those of you running any version of Jaguar should note that the "About The Finder" window also has a copyright date through 2002, despite the fact several updates of Jaguar have been released in 2003. Another thing our source has told is that Apple has been having some trouble with Quartz, which is why the piles icons and such aren't as graphically appealing one would suspect. Lastly, someone noted that there is no cursor in the screen-shots; this is irrelevant because screen-shots taken using Apple, Shift and the number 4 provide a selection tool eliminating the cursor. Plus if an item is in motion or contextual menu is being displayed it is common for the cursor to not show up.
While I am still not expressing my personal opinion on the validity of these images. I just want to provide everyone with as much information as possible to allow you to make your own decision.
I strongly suspect a lot more will worm its way out into the public domain after WWDC. In the meantime, I'd hesitate to make a call one way or another on those screen shots.
Of interest, Panther has not yet seeded outside the development team. As a result, it seems unlikely that the "leaked" screenshots were legitimate.
(In additional speculation... while I hate to encourage speculation based on a lack of Apple Legal action. Apple has traditionally pursued legal action against pre-release Screenshots, Movies, and Photos.)
Those screenshots sure look ugly...I mean that desktop wallpaper?!?!
To me a " New Pile" action would have to be quite different from a " New Folder" action to be of use and to not be confusing.
To me the "New Pile" selection in an Apple menu would have an arrow next to it opening a nested menu with selections, such as:
"...of Files" or
"...of Applications" or
"...of folders," etc.
Then each of these might have a nested submenu suggesting:
"...accessed yesterday" or
"...authored by me" etc.
Too late at night....not making sense...
That would make the whole idea way too complicated right from the start. The OS can figure out itself whether those objects are files, apps or folders.
That would make the whole idea way too complicated right from the start. The OS can figure out itself whether those objects are files, apps or folders.
Exactly. And if you use all these submenus you're really going to mess it up. The point I earlier tried to make about making a new "empty" pile. Yes you heard me, empty pile. If you could make a empty pile, you could after that select smart preferences or access the smart pile option in the submenu, where you could set the kriteria for the smart folder "search" - click update/ok and the pile would be filled with the resulting files/docs/prgs. It's not smart to have smart folders (pun intended) that have content in it before you make it smart. Another possibility is to have a piles submenu with "new pile" and "new smart pile". But if you had a pile search for documents created/edited today, it would have to be empty before you have actually edited/created a document that day. Complicated, I think not. Empty piles would have to be possible if smart piles is going to see the light of day. Yeah, yeah - "sum of it's content" I know, but to be usable in smart-mode it has to have the ability to be empty.
Chucker, you are right that those nested submenus I suggested were needlessly dumb and inelegant.
What I was thinking though is that a possible significantly new use for producing a "pile" might be to make a graphical and intuitive way of creating a query.
I know that may seem very unsexy and something only an accountant would love, but once file metadata is to a sophisticated state, then more advanced ways of searching are of course possible (as seen in your everyday databases business uses all of the time). However running searches and reports and queries is needlessly complicated and uses a specific vocabulary such as, filters and "operators" ("not", "and", etc.) and fields and others that the consumer would HATE.
The generic Finder does a good job of "finding" things, but is doesn't organize things well or find them intelligently without using advanced options. The Finder requires too many mouse clicks and text fields, plus it puts things in a window as a list that you still have to visually look through and confirm that it included objects that you want and then click on them to open them and THEN physically move them to the desktop with a mouse. This is okay for one or two files at a time or files with names of which you are already reasonably sure. But it is not really elegant for finding a large number of files or objects with similar metadata.
"Piles" could simply be ways of querying all data on the computer without ever using the term "query" and automaticaly having them available on the desktop.
So making a new pile could be a quick and easy selection of a few parameters via nested menus or text boxes that forms a temporary collection of apps, folder, files, whatever onto one place with the ability to be named and viewed with simple mouse controls.
This isn't much different than what many people have suggested, except that it needs to be grounded in the boring and technical structure of databases, while executed with little thought. AND if the pile only held aliases then the pile could be trashed or the entire desktop cleared without ramifications to any real files or apps. An Apple menu item named, "Last ten Piles" could be a way of quickly assembling piles that you already produced and named.
I wish I had the time and talent to make a Flash demo of what I'm thinking.
This may be exactly what a smart folder 'a la iTunes could do, but unlike in iTunes, these would need to have a very specific "feel" as a desktop object.
MacGregor: Hmm, interesting. So basically, you're connecting the "piles" rumors to the "metadata" rumors in BeOS's "Smart Queries" way. Gotta think about it.
I think piles would require use of some sort of metadata, some interpretive system for collecting stuff. How would the sytem know or learn how to associate files with one another with nothing more than file types to go by? Content indexing in software like the iTunes Music Store search, Mac Help, Finder searches and of course Sherlock is one obvious means. But there would probably either be a more enhanced version of that system, or additional systems akin to ID3 tags in MP3s, or even user-supplied metadata like labels extreme to help the system manage its own piles plus the user's piles. Come to think of it, the way Mail now "learns" what is junk mail is just the kind of system piles would need behind them, to have the computer observe the user's patterns until it can be made automatic (though of course not opaque to the user in case of errors). The system has to be expected to help in sorting this stuff, reducing clutter and labor on the part of the user. That's the trade-off of piles: slightly messy but quick versus the rigid but in-dubious nature of folders.
MacGregor: Hmm, interesting. So basically, you're connecting the "piles" rumors to the "metadata" rumors in BeOS's "Smart Queries" way. Gotta think about it.
I don't know anything about BeOS, but it sounds right. Queries are different from Searches or Browsings. Queries, in my mind, are how people think. It is a series of associations...patterns with imbedded memories and trials and errors of associations.
Humans don't in reality "search" their memories - like with keywords. Nor do we "browse" our memory banks like we do the internet. What we do is think of ideas and bring up associated memories in groups that are automatically evaluated by time and emotion and assumed relevance. That takes alot of fuzzy logic and intuition to be effective and of course it can also be dead wrong, but it is powerful.
The problem is that people are not aware of how they think and to get them to make a computer do something similar is as complicated as programming a robot to pick up an egg without crushing it.
It should be possible for today's PC's to be able to "learn" enough to at least be helpful. I don't mean to implement MIT sophistication, I just think there is still ALOT of room for improvement in the desktop GUI. And I hope "piles" is part of that next evolution.... an intuitive/icon-based query capability that becomes as transparent as double-clicking is now.
MacGregor: Hmm, interesting. So basically, you're connecting the "piles" rumors to the "metadata" rumors in BeOS's "Smart Queries" way. Gotta think about it.
For this kind of thing to work--which would be lovely--they'd have to make some deep changes. I mean, as far as I've seen iTunes and iPhoto do not deal with file-metadata. Of course, that's probably because the file system does not support that kind of metadata--yet. For example, the ratings in iTunes are stored in a separate file--I'd guess in the iTunes Music Library file. Trash that file and all that data is lost.
Another side-effect of this is that the ratings metadata is not accessible system-wide, only iTunes can. For the Finder to have access to this kind of metadata I suppose it should be stored in a prefs file that even the Finder and other apps can read--which would be quite anal--or using the new BeOS-like FS the metadata info should be attached to each file. I hope what I wrote makes sense, sometimes I wish I had a better grasp of English...
Yes, ZoSo, that does make sense and that is the reason for all of the talk of adding such metadata as a "core" component of the OS.
For the above kind of queries to be of really any revolutionalry use, the Finder would need the same set of metadata for everything from pdf's, jpegs, Word files, Excel files to html pages, mp3's and maybe even iCal dates and Address Book contacts.
That's a bit much, but that is what sets the human brain apart from current computer Finders, and it is not out of the question. It is just another layer of data with a bit of AI involved.
It is the file management version of Quartz Extreme. Add lots of functionality at the root level that you may not need for a few years, but you're glad it is there when someone makes the killer app to use it!
If you take metadata in its boradest sense, you could easily deal with files that hold extra data about other files rather than forks or embedded data in the files themselves. Remember that MP3, AAC, QT, TIFFs, and other file formats do allow embedded extra data in them anyway. But a database-like system could work too. If they use an xml-based markup system and standardize it, essentially having dictionaries a la applescript, and allowing aribtrary mark-ups, I imagine it could be done. It would take a generation fo software to take advantage of it, especially third-party offerings, but the foundations could be laid in 10.3 with some critical Apple apps taking advantage of it immediately with more expansion to come later.
Exactly, BuonRotto. I don't mean to minimize the amount of work or the difficulties for 3rd party apps, but the technology itself, as you said, isn't that difficult and has been used in a limited sense and in a proprietary sense before.
QE uses the graphic processor to do some of the extra work, with a new OS structure on a new chip, maybe the cpu can be efficient about keeping metadata updated.
Comments
I have seen a variety of people immediately assume the images are fake because of the copyright date on the "About The Finder" window. First off I would just like to clarify for you all that Apple began working on Panther back in November of 2002. According to our source the images are said to be from an early build, how early he did not say. Also, those of you running any version of Jaguar should note that the "About The Finder" window also has a copyright date through 2002, despite the fact several updates of Jaguar have been released in 2003. Another thing our source has told is that Apple has been having some trouble with Quartz, which is why the piles icons and such aren't as graphically appealing one would suspect. Lastly, someone noted that there is no cursor in the screen-shots; this is irrelevant because screen-shots taken using Apple, Shift and the number 4 provide a selection tool eliminating the cursor. Plus if an item is in motion or contextual menu is being displayed it is common for the cursor to not show up.
While I am still not expressing my personal opinion on the validity of these images. I just want to provide everyone with as much information as possible to allow you to make your own decision.
Originally posted by NETROMac
Phanter, Phanter, Phanter ...
No, no, no! I'm Panther...
Panther Screens Fake?
Of interest, Panther has not yet seeded outside the development team. As a result, it seems unlikely that the "leaked" screenshots were legitimate.
(In additional speculation... while I hate to encourage speculation based on a lack of Apple Legal action. Apple has traditionally pursued legal action against pre-release Screenshots, Movies, and Photos.)
Originally posted by NETROMac
MacRumors think the Panther screens are fake.
Yea, all those screeenshots are fake.
To me a " New Pile" action would have to be quite different from a " New Folder" action to be of use and to not be confusing.
To me the "New Pile" selection in an Apple menu would have an arrow next to it opening a nested menu with selections, such as:
"...of Files" or
"...of Applications" or
"...of folders," etc.
Then each of these might have a nested submenu suggesting:
"...accessed yesterday" or
"...authored by me" etc.
Too late at night....not making sense...
Originally posted by MacGregor
Those screenshots sure look ugly...I mean that desktop wallpaper?!?!
To me a " New Pile" action would have to be quite different from a " New Folder" action to be of use and to not be confusing.
To me the "New Pile" selection in an Apple menu would have an arrow next to it opening a nested menu with selections, such as:
"...of Files" or
"...of Applications" or
"...of folders," etc.
Then each of these might have a nested submenu suggesting:
"...accessed yesterday" or
"...authored by me" etc.
Too late at night....not making sense...
That would make the whole idea way too complicated right from the start. The OS can figure out itself whether those objects are files, apps or folders.
Originally posted by Chucker
That would make the whole idea way too complicated right from the start. The OS can figure out itself whether those objects are files, apps or folders.
Exactly. And if you use all these submenus you're really going to mess it up. The point I earlier tried to make about making a new "empty" pile. Yes you heard me, empty pile. If you could make a empty pile, you could after that select smart preferences or access the smart pile option in the submenu, where you could set the kriteria for the smart folder "search" - click update/ok and the pile would be filled with the resulting files/docs/prgs. It's not smart to have smart folders (pun intended) that have content in it before you make it smart. Another possibility is to have a piles submenu with "new pile" and "new smart pile". But if you had a pile search for documents created/edited today, it would have to be empty before you have actually edited/created a document that day. Complicated, I think not. Empty piles would have to be possible if smart piles is going to see the light of day. Yeah, yeah - "sum of it's content" I know, but to be usable in smart-mode it has to have the ability to be empty.
[preparing anti-rant shield
(
What I was thinking though is that a possible significantly new use for producing a "pile" might be to make a graphical and intuitive way of creating a query.
I know that may seem very unsexy and something only an accountant would love, but once file metadata is to a sophisticated state, then more advanced ways of searching are of course possible (as seen in your everyday databases business uses all of the time). However running searches and reports and queries is needlessly complicated and uses a specific vocabulary such as, filters and "operators" ("not", "and", etc.) and fields and others that the consumer would HATE.
The generic Finder does a good job of "finding" things, but is doesn't organize things well or find them intelligently without using advanced options. The Finder requires too many mouse clicks and text fields, plus it puts things in a window as a list that you still have to visually look through and confirm that it included objects that you want and then click on them to open them and THEN physically move them to the desktop with a mouse. This is okay for one or two files at a time or files with names of which you are already reasonably sure. But it is not really elegant for finding a large number of files or objects with similar metadata.
"Piles" could simply be ways of querying all data on the computer without ever using the term "query" and automaticaly having them available on the desktop.
So making a new pile could be a quick and easy selection of a few parameters via nested menus or text boxes that forms a temporary collection of apps, folder, files, whatever onto one place with the ability to be named and viewed with simple mouse controls.
This isn't much different than what many people have suggested, except that it needs to be grounded in the boring and technical structure of databases, while executed with little thought. AND if the pile only held aliases then the pile could be trashed or the entire desktop cleared without ramifications to any real files or apps. An Apple menu item named, "Last ten Piles" could be a way of quickly assembling piles that you already produced and named.
I wish I had the time and talent to make a Flash demo of what I'm thinking.
This may be exactly what a smart folder 'a la iTunes could do, but unlike in iTunes, these would need to have a very specific "feel" as a desktop object.
Originally posted by Chucker
MacGregor: Hmm, interesting. So basically, you're connecting the "piles" rumors to the "metadata" rumors in BeOS's "Smart Queries" way. Gotta think about it.
I don't know anything about BeOS, but it sounds right. Queries are different from Searches or Browsings. Queries, in my mind, are how people think. It is a series of associations...patterns with imbedded memories and trials and errors of associations.
Humans don't in reality "search" their memories - like with keywords. Nor do we "browse" our memory banks like we do the internet. What we do is think of ideas and bring up associated memories in groups that are automatically evaluated by time and emotion and assumed relevance. That takes alot of fuzzy logic and intuition to be effective and of course it can also be dead wrong, but it is powerful.
The problem is that people are not aware of how they think and to get them to make a computer do something similar is as complicated as programming a robot to pick up an egg without crushing it.
It should be possible for today's PC's to be able to "learn" enough to at least be helpful. I don't mean to implement MIT sophistication, I just think there is still ALOT of room for improvement in the desktop GUI. And I hope "piles" is part of that next evolution.... an intuitive/icon-based query capability that becomes as transparent as double-clicking is now.
Edited for terrible spelling....
Originally posted by Chucker
MacGregor: Hmm, interesting. So basically, you're connecting the "piles" rumors to the "metadata" rumors in BeOS's "Smart Queries" way. Gotta think about it.
For this kind of thing to work--which would be lovely--they'd have to make some deep changes. I mean, as far as I've seen iTunes and iPhoto do not deal with file-metadata. Of course, that's probably because the file system does not support that kind of metadata--yet. For example, the ratings in iTunes are stored in a separate file--I'd guess in the iTunes Music Library file. Trash that file and all that data is lost.
Another side-effect of this is that the ratings metadata is not accessible system-wide, only iTunes can. For the Finder to have access to this kind of metadata I suppose it should be stored in a prefs file that even the Finder and other apps can read--which would be quite anal--or using the new BeOS-like FS the metadata info should be attached to each file. I hope what I wrote makes sense, sometimes I wish I had a better grasp of English...
ZoSo
For the above kind of queries to be of really any revolutionalry use, the Finder would need the same set of metadata for everything from pdf's, jpegs, Word files, Excel files to html pages, mp3's and maybe even iCal dates and Address Book contacts.
That's a bit much, but that is what sets the human brain apart from current computer Finders, and it is not out of the question. It is just another layer of data with a bit of AI involved.
It is the file management version of Quartz Extreme. Add lots of functionality at the root level that you may not need for a few years, but you're glad it is there when someone makes the killer app to use it!
QE uses the graphic processor to do some of the extra work, with a new OS structure on a new chip, maybe the cpu can be efficient about keeping metadata updated.