Slashdot Mirror
The Death of Folders?
saintlupus writes "There's an interesting article on Wired about the interface changes in Tiger being a precursor to the demise of the classic folder-browsing Finder." From the article: "Users type search queries more or less as they did pre-Tiger, but 'the quality, scope and presentation of the results are significantly better, so users get good benefits without having to change their behavior.'"
Microsoft purposed the death of folders back when they announced the WinFS system. The idea of an SQL or Database file system where queries are performed more often than direct references isn't new. While Microsoft is not releasing WinFS with longhorn, much of their search capabilities and ability to group files into multiple spots and 'death of folders' will still be occurring. Obviously apple is the first to give a solid attempt at implementing this, hopefully it will make organization far easier;)
There's an interesting article on Wired about the interface changes in Tiger being a precursor to the demise of the classic folder-browsing Finder.
Call me when Folders become saved queries, and then we'll talk about the semi-demise of Finder. Actually, Finder wouldn't leave us at all. In a properly designed database file system, folders/directories should be replaced with standard queries. An example of this is the Labelling system in GMail. You can add a meta-data label to any email, which will then cause that email to appear in a virtual folder of the same name as the label. But if you pay attention to the search bar, you find that the folder is nothing more than a stored search on a key piece of meta-data.
This concept has massive implications for File System Usability. Under the folders-as-search concept, the same files can be organized under multiple folder groupings. This labelling data not only assists users in doing future searches for their information (i.e. A real reason to fill out meta-data other than "It might be useful."), but it also provides the user with a way of organizing ALL data for a given project under one folder without forcing the user to make a copy. It may not seem all that revolutionary, but I think you'll find that a lot of GMail users have already grasped the real power of the concept.
That being said, WHAT'S TAKING SO DAMN LONG?! This stuff was figured out 10+ years ago, and pieces of it were even included in BeOS. NTFS has had many of the necessary features since its inception (just turned off for some bloody reason), and ReiserFS is bringing the same design to Linux. So what is everyone waiting for? The next guy to scoop you on it?
*sigh* Dear Mr. Jobs: Will you please demonstrate to everyone how you do this properly with a file system? Thanks. Kudos to your NeXT development team who's made this possible.
Javascript + Nintendo DSi = DSiCade
Long live the directory!
That's funny, I thought Gmail's labels system was supposed to be the death of folders.
"Da ist ein Technölüst in mein Unterpanten!"
I just put everything in the C:\ drive and know that I can find it using Windows XP's sweet search capabilities!
err...yea...
It will become less prevalent, maybe. I worked for a company that made software that tried to do that. No one bought it. It's a nice idea, letting the computer manage things for you, you not being a file clerk, etc. A lot of users tend to want more control than that. It is difficult to do a lot of things in a purly search based environment, like archiving.
The only shocking part is that there will be millions of people that have been using computers since the 1980s, who never noticed that there ever was such a thing as folders/directories.
I'm sorry, but I like to categorize things. I like to know where they are, in this logical space. If this loses a document, can you dig it out? Or did it just never exist?
frog confirms it,
"The Finder has been dying for a long time," said frog creative director Cordell Ratzlaff.
What a load of Bullshit
Spotlight is really good, but that hasnt stoped me from being anal about setting up files so i can find things.
What really pisses me off is out iTunes reognized all my music when it was inported into the libary. I spent years putting together music in such a way that i can find it. Now i have the seach for it b/c itunes had to mess things up.
Mikey
I've always been the kinda guy to fall for the girl dressed like an eskimo.
If you have your work organized in a defined folder structure, your memory will be faster than any Spotlight search -- especially given Spotlight's annoying habit of searching before you complete the search term.
Strange women lying in ponds distributing swords is no basis for a system of government.
But the very concept of having millions of files just scattered about in a completely flat heirarchy, well, doesn't seem like a really good way to handle your company's data.
In other news, it was recently announced that due to the widespread use of email, street addresses would soon become obsolete. Out with the antiquated, in with the new!
I've always find that folders are big and in the way. ls works for me. If I use the Finder, then I have it in "list view" which is pretty darn handy and similar to ls.
Yes! I listen to NYC Speedcore and do math at 3AM. I suggest you try it too.
I think it's a good thing. I mean, originally folders/directories were designed to have an order in your files. With the advanced searching technics this whole issue is solved.
Of course there are more things folders are handy with, but they've not disappeared, have they?
You insensitive clod!
Do not try to read the dupe, thats impossible. Instead, only try to realize the truth
What truth?
There is no dupe
From the article: "The way Searchlight transforms the computing experience is akin to Google's effect on the web"
And Google has made bookmarks obsolete, right? So Searchlight will make folders obsolete.
Better search is always very cool. But proper organization and categorization is better yet. The problem is not that the latter is a bad system but that people don't do it very well. I think a system that helps people organize their stuff will be even better than a better search. The "labels" which are used instead of folders in gmail seem like a step in that direction.
<sig>Guvf vf abg n frperg zrffntr
I'm still waiting for the time when I can "see" the computer code, via a green monitor that displays a shower of code. Then, I will have a plug that connects to my spinal column and allows me to "enter" the computer and manipulate the code using my brainwaves.
It'd be very efficient, I could then just think of finding a file, and there it would be. Or better yet, I could imagine a beowul...NO CARRIER
IGB: More fun than eating oatmeal!
While I love the idea of a decent search system, the time honored forlder hierarchy works because thats how people think. For instance, pictures. For these meta based search systems each picture needs to have a comment attatched (if not searching by date).. and who really does that? I tried adding notes to my pics in iphoto but after a while it gets tiresome.
And backups.. in a workflow.. every project has its own file and subfolders, makes it easy for backup and finding files.
Anywho... folder hierarchy works great and is here to stay for most people. (except for those people who just save everything to the desktop.)
The idea of a folder as a visual reference for a directory may well be on the way out. There's still plenty of need for directories and hierarchical organization, though, for managing the contents of a system from the standpoint of software. OS X's Unix base is pretty heavily dependent on the basic Unix filesystem structure, and lots of software is built with a deeply ingrained assumption that it's there and the way files are organized.
Spotlight is great for users, but there will be a need for something like the Finder indefinitely.
Disinfect the GNU General Public Virus!
I applaud this effort. I think it's about time an OS implemented techniques similar to what Google has been using. Actually, it reminds me of Google's Desktop Search tool.
The age of linearity within computers is coming to a close.
That that is, is.
I must admit, I really like Tiger's Spotlight. It has improved file management on my machine considerably.
Having said that, how can this apply to removable media? I would like to see a feature on the next MacOS that automatically indexes removable storage.
Let's say I burn a CD of some data. The finder should keep track of which files I burned to that CD, long after I erased the actual files from my hard drive. That way, I can perform spotlight searchs on my data, even if it really isn't present on my local drive.
Find the file that you want and the machine prompts you to insert the proper CD.
I read this as the death of Folgers. I almost fained since Folgers is The Best Part of Wakin' Up(TM).
More
What's the point? Folders and subfolders work for me...one of GMail's sorest problems for me is that you can't have sub-labels (something solved by me using Thunderbird to do my gmail...yay POP access). I assume this would have similar problems.
By summer it was all gone...now shesmovedon. --
What I'm wondering is what is broken with the whole directory/folder design? I wasn't aware that there was a problem. And what's the alternative... every file is stored on the hard drive in some arbitrary location, and a query is needed for each and every file access? That seems like a *ton* of overhead to fix a problem that just doesn't exist.
And what about file systems? I know that modern file systems like NTFS are much better at optimizing file storage for large drives with millions of discrete files, but are all of the modern ones ready to handle a drive with millions of files all at root?
I don't respond to AC's.
I installed f-spot (a mono photography application for Linux -- it's in Debian unstable) on my wife's Thinkpad and it went out and thumbnailed over a 1000 pix on her system.
With the tagging system (you highlight and tag photos with tags like family, favorites, Mexico Trip etc) it was so easy to navigate through the huge collection without knowing where any of the pictures lived within the file hierarchy.
Something similar for navigating the whole system would be amazing: thumbnails of documents, meta information... the sample really made the possibilities come alive.
How about it? Can we take the f-spot engine and create a file browser with tags?
Comment removed based on user account deletion
All I can say is the linking of Google Desktop Search and the program called GDSuite which makes GDS work like the "search" function from windows has already changed how I get to things on my machine. If I know a chunk of code from a certain filetype is what I am looking for, it is extremely straightforward to just type that information in and get a response immediately.
The only thing I can hope to see is for Google Desktop Search to add a "label" functionality to GDS so that I can label things that are "games" and "code" etc, to help narrow down searches or even use virtual directories where it brings up a windows like link to all executables labled for games on the hard drive without having to individually organize.
This way you could make folders that consist of multiple labels and or focus them down to less labels etc at a click of a button.
If you don't vote, you don't matter, so don't waste your time telling me your opinion
It's all very well to talk about the death of folders because of intelligent indexing and searching of file systems, but this is in the context of retrieving data. Where a hierarchical structure is so useful is when you are saving information in the first place. It's important to remember that a hierarchy divides the file system into a number of logical namespaces.
A completely flat filesystem sounds all very well in principle, but how do you find names for all of those files? I have loads of files on my computers with the same names but in different namespaces. Or are we going to throw away filenames as well?
Something is happening here but you don't know what it is, do you, Mr Jones.
Ha! Wishful thinking.
Folgers is to coffee as aparagus is to chocolate.
TWW
"Encyclopedia" is to "Wikipedia" what "Library" is to "Some people at a bus stop"
Actually....can someone give me a more concrete idea of how a soup file system works?
when Push Comes to Shove
Or on MacOS take a look at all the pfiles and see what they can control and what they can't.
Or say you want to find a way to make the dock transperent and you search for Dock Transperance. While the real term that the search will find is Dock Clearness. Or that file you saved way back when you don't know the date you did it or what it is about but once you see it you know that is the one you need.
Sure I like spotlight but there are some cases where it just fails me mostly because I am absent minded.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
That plus there is still a large group of folks in the business world for whom computers are still fairly recent (the managers and partners who have been working since the 70's and 80's). Granted their numbers are starting to thin, but there are still a great many folks, in relatively high positions, who like the folder system because it replicates a filing cabinet- they get it. Trying to educate the entire generation on a "whole new way" of doing something "easier and faster" will frighten them off.
If brevity is the soul of wit, then how does one explain Twitter?
"Ooh, we're so cutting edge we're not going to use a capital letter at the start of our company's name." Pretentious twats. I bet they all have poncy rectangular tinted glasses and soul patches and ride around their offices on scooters as well.
You must think in Russian.
I have a gmail account for a variety of mailing lists. I am getting a bit fatigued with its interface. Yes, I can label all the mailing list stuff, but the oddball stuff is a pain to deal with. I wish I could route to folders instead of labeling the mailing list stuff, so the regular view is strictly stuff that is unlabeled.
ostiguy
Better yet, instead of the death of folders, why not something which sits alongside side, like som sort of brilliant search capability? But seriously, while its a good start - does it need to go any further than apple or google have taken it? Do we really want power to be hard to get at?
Finally... I've been waiting for the end of folders to come. I just hope that they are replaced by directories instead of libraries.
The Tools Of Ignorance wanna be a tool?
- The category(s) of the object are encoded in the metadata of the object
- The search is sufficiently broad enough to collect all items that belong in a search (false negatives)
- The search is sufficiently narrow enough to exclude items that do not belong in the set (false positives)
Some simple examples of hard-to-define-by-search collections of documents include:- All the documents for a given client (stock photos used for that client, invoices, emails) -- not ever document will have the name of the client in it.
- The set of "final" versions of files for given client (with no intermediate or working versions) -- might be doable if EVERY office file in the system were in version control system.
- "good" documents from a web search for further workflow processes (requires subjective evaluation)
- the set of spams (the imperfection of spam filters is a microcosm of the problem of the inaccuracy of search)
I have used (and created) system that do continuous/recurring searches to create collections. Its a great idea, but it can be awkward to form the right search and even the best search has exceptions (falsely excluded documents and false included documents). Direct manipulation (i.e., putting a file in a folder or removing a file from a folder) is far simpler and faster than typing metadata or refining a search.Folders are still useful.
Two wrongs don't make a right, but three lefts do.
I don't think the directory as we know it is dead, it is a nice way to hierarchically (word?) organize our data (but wait, Documents and Settings???). Seriously, directories are intuitive enough and most people get comfortable with them quickly.
But, there are some problems with directories:
However, this article I think shows the way technology will take us and I like the abstraction and "flattening" of the storage universe. I've already become less neurotic about how to organize and store photos, etc., especially now with photo organizers and desktop search software like Google desktop. For me it makes more sense to "ask" my computer where something is and have it return the top twenty most likely responses (with the ability to drill deeper if necessary).
Directories served a good purpose, but weren't they mostly artifacts anyway? Aren't they kind of an opaqueness of underlying technology? Directories as far as I remember were a way of implementing pointers and references to blocks of data on a drive, albeit a nicely abstracted implementation at the time (except for DOS, ick... (why no ".xxx" extensions allowed for DOS directories, huh?)).
Reasons? Well, first of all Spotlight won't search the whole of your drive. Can't remember if it was in /usr/local/bin or /usr/bin? Tough. Spotlight won't help you, it doesn't look in those hierarchies.
Made a mistake typing your search term into Spotlight and on an older machine? Don't even think of hitting that backspace key, or the Finder may go into a spinning beachball hell whilst it tries to live search everything for you.
Want to find just files and nothing else (ie. no meta-data or content-related stuff, just filenames)? Well, you can use the undocumented start-your-search-with-a-double-quote feature, but that doesn't work well because it doesn't understand wildcards (so "*.java won't work, for example, whereas ".java will but would include *.java.backup).Also it seems to lose its idea of filename-only as soon as you hit backspace and try to re-edit it. In other words, typing ".java will find me *.java*, but typing that, then hitting backspace, then typing hte final 'a' character again will start finding me things with java in the content instead of just the name.
It also has poor resource usage - some seem to be lucky, but search the forms and you'll see many people complaining about processes called mdimport or similar hogging large amounts of CPU. Then there's the indexing it does every time you connect a firewire drive - if I reboot my Powerbook in target mode and hook it up to the Power Mac, a large amount of indexing is initiated which slows down my performance on that drive. I can set it to not index, but then it slows down search on that drive. What's needed is for the indexing stuff to be really low priority or user-ppausable perhaps.
Sorry, Spotlight is ok but in the Finder it's a pain more than a help for me. I wouldn't have minded it in addition to Panther's more straightforward 'find a file' bit, but as a total replacement for that it's rather lacking. I'm not even contemplating using it as a complete replacement for a normal directory structure.
Cheers,
Ian
Pie-in-the-sky. Please spare me the deep-think prognostications of people who obviously are unfamiliar with how the facility actually works (or doesn't) in the real world.
When it is good, Spotlight is very, very good. And when it is bad, it is horrid. So far, in my experience, Spotlight has been very, very good about 50% of the time I've really used it (i.e. to find something I wanted to find, as opposed to playing around with it). And horrid the other 50%.
Spotlight has several big problems.
a) It doesn't find things reliably. This isn't like using Google on the Web, where you're happy with the results you find, and mostly don't know about what relevant hits Google missed. You have a very good idea what's on your hard drive, and it is incredibly annoying when Spotlight does NOT find a file you know is there.
There is ongoing discussion of why Spotlight doesn't find things reliably, and, of course, many people who say "It works for me," but the number of users reporting that Spotlight is not finding files they know are there is very significant.
There are various reasons for this. One is that Spotlight has a fairly long built-in exclusion list of directories it doesn't think you really want to search, but, unfortunately, it does not explicitly show you what they are. This is not, however, the only issue.
b) It doesn't find things quickly. Wags are starting to call it "stoplight." Frankly, I'm scared to type anything directly into the search field. I've gotten to the point where I type the search target into a text editor and paste it into the edit field.
The problem is that Spotlight oh-so-cleverly gives real-time live updating of the partial query as you type it in. So if you type in "Slashdot", for example, by the time you have typed in two characters it is trying to display every file on your computer that begins with "sl". For reasons that aren't clear to me, this frequently locks up the Finder's UI with a spinning pizza wheel. The entire Finder becomes unusable--you can't even activate another window and search for the file manually--for big fractions of a minute.
c) A signficant number of users are reporting frequent occasions when Spotlight causes their whole system to slow down. And, in at least one case, I've pinned down a situation in which Spotlight, for some reason, actually causes another program to fail with file I/O errors unless it is prevented from accessing the directories that program is using.
So, Spotlight is sometimes wonderful... but other times is unreliable, slow itself, slows down the rest of the system, and makes other programs unstable.
But aside from that, Mrs. Lincoln, how did you like the play?
"How to Do Nothing," kids activities, back in print!
No.
This is going to suck. How will the system account for spelling errors? Poorly, I'll bet. Also, what do you want to bet that this will lead to a completely guided view of the contents of your hard drive, in which OEMs now decide what we can search for and what we can't. It will be like that "These are the system files! Don't f*ck with these!" warning page on windows only much, much angrier.
I say, screw these guys. If you want to get that restrictive with my machine, I shouldn't have to pay for it. I guess it will be "Linux, here I come" time.
Users get "good benefits", eh? Since when are benefits bad?
Anyhow, the Volume->Folder->File metaphor has always been a bit strained, and I'm glad to see that quality search tools and filesystem metadata are chipping away at the average user's need to use such a metaphor as a crutch. Hopefully, this the first step toward establishing a new filesystem metaphor: one where the data can be somewhat independent of the logical location on a disk, and that doesn't treat the user like a cripple hobbling toward his/her data.
The addition of metadata especially interests me, as it opens the door to having the filesystem exposed as several different metaphors at once -- the logical layout of the filesystem need have nothing to do with its navigation metaphor.
We may not imagine how our lives could be more frustrating and complex—but Congress can. – Cullen Hightower
We need something to help that is clear from the number of digital objects we have lying round on our computers these days. Some method of collecting these objects into conceptual sets or classifications (apart from file extensions which is not always the most useful) could be really useful - I have read some interesting stuff by people who are Metadata crazy (seem to have lost the links though - the tiger review of metadata writer was really interesting...) Maybe the answers are somewhere there.
But for most people, some method of grouping data, adding categorical schemes, visually and texturally organising and generally making files/objects more plastic in the way that we store them would be a great step forward.
But in any case, nested folders *do* still have uses. And I think we need --in addition to-- rather than --instead of--.
---- Posted anonymous as bloody slashdot is banning IP
---- The Open Source Record Label : : LOCARECORDS.COM
Comment removed based on user account deletion
Will this model be limited to userland? Or will it also change the way the OS itself is organized? eg, instead of having config files scattered all over the place, could an attribute be added to those files called "conf" or what have you? Maybe these are stupid questions, but could someone here enlighten me?
Not an A/C, just too freakin lazy to create an account. CS
How can I hide things from my kids or wife?
Ten levels of folders with scary names like "SYSX.dat" keeps them
the hell outta my pr0n.
Search engines become massively popular. Google now biggest media company
Computer storage capacity nearing 1 TB; people stop worrying about running out of space, and save all files.
Microsoft announces the WinFS file structure; which will do away with folders.
Google releases a PC search tool, which is widely used.
Other companies have also release PC search programs. People debate which is the best
Apple follows suit. Anyone surprised?
I don't know about you, but when I am looking for a specific website, I am more likely to type in the URL than to search for it in Google.
My point is that Spotlight/Google is fine when it comes to vague requests (I want teh funny Star Wars spoof!), but when you know exactly what you are looking for and exactly where it is, you don't want a big pile of options to choose from.
My other processor is big-endian.
I've been using Copernic Desktop search and it has largely changed the way I interact with my laptop.
My 'My documents' folder is a mess. I am too disorganized to keep it straight as projects/information evolve over time, since my classification scheme will necessarily also, and it is a royal pain in the ass to go through and create new directories, transfer files etc... Nobody but the most anal retentive among us with loads of free time bothers with it.
But with the destop search, I just type in a few words and the info I'm interested in pops right up. I've found all sorts of information I thought I had lost, only to have the desktop search find it buried in some deeply nested folder, a victim of a previous noble effort I had made to classify things into neatly labeled folders.
Here's an example. I am a researcher and I have probably downloaded about a 1000 different pdf research papers over the years. It was becoming so difficult to keep track of these papers on my hard drive, that I would just re-download the paper again instead of trying to figure out which damn directory I had stuck it in two years ago, when the way I thought about my research was different than the way I think about it now. Now I've given up on the directories - I've stuck all the pdf papers in a single directory and just do a Copernic pdf search with a couple of the keywords I'm interested in and the paper pops right up. It's great and it's the way of the future.
Yes, but in a system that uses meta-data and saved searches, you have an interface that acts like a file and folder system where there is a folder for each case, a folder for each of your attorney and paralegal's work (useful if you have to check over the work of that incompetent fool you just fired), a folder for all of your tort works, a folder for all files involving certain clients, a folder for all files involving certain kinds of filings, a file for all of your official correspondance, a folder for all of your billing etc etc etc.
And rather than having to make a duplicate copy of each file, one file can appearin multiple folders.
So it's like the current system but better.
I have a lot of opinions about Cyborgs and Architects
While I'm all for the idea of a more database-esque file system, I see this as bad for consumers.
If only applications manage particular files than you become "locked" to that application for managing those files.
Think of it, it makes it far easier to force DRM (I'm sorry, you can't even SEE your files unless you use the approved application, citizen). Plus you're condemned to THAT application because it doesn't necessarily have to release its data store to other applications.
There are times when searches are ideal for grouping disjoint sets of information. There are many, many more times when a best guess is completely insufficient. Searches to augment folders? Sure. Searches to replace them? No way.
Dewey, what part of this looks like authorities should be involved?
I hate managing drive space. I just want to connect a drive and label it as removable or not. If it is not removable it becomes part of the storage pool, and the FS handles where to put my file. I don't want to have to remember what drive that big file is on. I should not need to know that there are 4 drives inside my computer. Just space, no matter what form it is in.
FYI. Folders should be part of the storage pool, and hold files that reside anywhere in the pool.
When (if) the OS gets rid of folders, we'll need (and have) 3rd party apps to put things back in a heirarchy of folders. It's a fast, logical way to group things that many people are not going to give up for a search or tag based system.
"And rather than having to make a duplicate copy of each file, one file can appearin multiple folders."
Haven't you heard of "links" (as in UNIX, Linux, and - yes - even Windows)?
Something is happening here but you don't know what it is, do you, Mr Jones.
grep: argument list too long
Maybe I'm alone in this, but I really hate tagging metadata as the sole means of organizing large sets of files. I tend to prefer the physical metaphor, a place for everything and everything in it's place, over the vast sets of forgettable synonyms you can use to describe a document.
And if I want it in more than one place? Space is cheap - I can make copies of it and put it into different places. Different copies, with the same name!
The main reason I don't like using Gmail is that I can't get used to not having a visual way of organizing my data. In my yahoo messages, I mark an email and move it to a folder. Then I have the comfortably familiar folder tree, that lets me know all of the subcategories I can choose. It's automatic, it's easy, and it does what I want it to.
Advanced search features are great, but not at the cost of useability. If it triples the amount of time it takes me to go through my inbox in order to tag every email with relevant metadata, it's not saving me any time or energy.
Folders may die, but at what cost? It certainly won't offer me any productivity increases, and people less knowledgeable than me will find it even more difficult without that metaphor to relate to.
Databases are great for compiling numbers and facts. They're not so user friendly as to become the next great interface for the masses.
Watch most users and they will not have bothered to learn hardly any of the keyboard shortcuts to doing things. They seem to like the mouse so much that I don't think this kind of thing will have any impact on how most people use their computer.
;-)
The power users will be down with this but they never really make the news.
LoB
"Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
They are directories!
Directories are dead but Spotlight didn't kill them, smart folders will. A lot of people here seem to be missing the point. No one is really going to want to search in spotlight everytime they need to find a specific file. However, uses will get used to the idea of setting up a smart folder to organize files for them. Mix that with effecient file grouping a la Windows XP and you have a great way to organize files on a whim. It's not here yet, even with tiger, bu give it a few revisions. Leopard may even bring more of this to bear but don't hold your breath.
It will be too easy for files to get lost. Say you don't label something properly, or you change the label, or you forget the name, or the name is unmemorable - what will happen to the file? Just sit there on your disk taking up space, never to be seen again?
And how about old/less useful files that are unnecessarily included in searches, forcing you to read over more file names to find what you want?
One handy feature about folders I've (automatically or intentionally) organized things in is it makes it easy to go back and figure out what I no longer need, and delete it, thus freeing up disk space and reducing clutter. Spotlight is designed to GENERATE clutter.
---If you can't trust a nerd, who can you trust?
If Apple and Microsoft and whoever else allow for searching a file by type, keywords, some new as-yet-unnamed meta information, etc. wouldn't it stand to reason that the OS itself could then manage the placement of files to optimize performance instead of having arbitrary user folders that have no particular understanding of the underlying disk layout?
It would seem that allowing the system to optimize file placement could greatly help performance and stability by reducing or even elinimating file fragmentation.
Granted, convincing Grandma that she just needs to type in a few keywords instead of opening one of 100 files on her Desktop may take some human engineering.
Mike
But how will you back up ALL of your data? I know that users don't back up but if I want to back up all my important stuff, and KNOW that it was ALL backed up, I put it in a system of sub-folders and then back up my root folder. With search, you will never know if something is missing because you typoed a file name or a query option.
Folders won't die, they're one meaningful way to deal with stored information.
Like the CLI and GUI are two interface paradigms, the Nautilus Spatial and Filesystem Browsers are two ways to navigate through folders of data, having a user decide where information is stored won't change.
The whole UI paradigm has picked up a lot from everyday office concepts: documents filed in folders in filing cabinets. That's not going to change any time soon, even with search software making it convenient to find things, because we will still need to put things in storage. Storage folders may become shortcuts-to-frequent-searches but this won't remove their existence from the interfaces we use, and will still feature hierarchical search capabilities so we can refine the bounds of what we're looking for.
Can you keep a history of your search results? e.g. If I search my hard disk for "webpages I have made, this year, that have pictures in them" and also a search for "pictures that I have used in webpages this year", is it possible to make a "folder" to store these results (maybe not the actual files, but links to them ... or is that how all file-systems work anyway ....), so I can have them avaialble easily? If not, thats rubbish. I don;t want to have to search each time for commonly used files.
b3 4phr41d 0f my 4bov3-4v3r4g3 c0mpu73r kn0wI3dg3!
MadDwarf
If Opera invented it why haven't they patented it already?
Something tells me the reason why they haven't patented it is that labelling things to make them easy to find was invented long before computers....
I'll probably be modded down for this...
Umm... tell me again how labels limit you?
Now: I work in the legal field and every attorney and paralegal in the office saves documents in case specific folders.
In the future: I work in the legal field and every attorney and paralegal in the office saves documents with case specific labels.
A folder system is equivalent to mandating that files have only a single "location" label. Labels are nothing to be afraid of.
I see moderators are being upset of renewed and well known technologies are officially finding its way to broad public under the Apple or MS sunshine.
Students and nerds often invent things that business savvy people are turning into lots of money. Nerds also experience being sued later on when business got patent granted under USPTO etc.
Do not fiddle around Linux, do something you can stand behind strong. Do not bring another timber into corporate bonfires.
File systems, starting with Apple's HFS in 1986 or so use database-type structures to store info about files. The directory a file is in is only a field in the database. So it doesn't change anything about the data structures if you store all the files in one directory or in many.
However, if you do try to iterate that directory it will take forever to do so. But in theory that isn't going to happen, as directories are no longer organizational strategies at that point.
HFS stores all file data in B-trees.
http://lkml.org/lkml/2005/8/20/95
If you look at the photo in the link, the guy has a section called Bittorrent movies in the right-hand side. Clearly that is intent to distribute!
MPAA swarm!!!!
I personally have iTunes set to not organize, and not move my music. I keep all my songs in folders organized by genre on the second partition of my HD. iTunes gladly "imports" them by just remembering where they are.
Go to: iTunes->preferences->advanced tab
uncheck "Keep iTunes music folder organized" and "Copy files to iTunes music folder when adding to library"
Create your iTunes playlists the same way your folders are aranged. Select your playlist, drag your folder to it to import those songs into it. Your songs stay where you put them, and just their location is recorded in the DB.
"That's so plausible, I can't believe it!" - Leela
For average computer users, maybe getting rid of folders makes sense. But for those of us who are organized--it would be a curse.
I've used Spotlight and for me it is an annoyance. I actually have a good idea what things are called based on what they are and the project. So, if I look up a job number, I don't want to find that there was a similar number in 5000+ emails on my drive. Spotlight needs to be dumber at times and just find files that contain "X" in the name, or "Start with X" --you know, kind of function like the previous find command. I like the new GREP-like functionality on occasion. But 90% of the time, I know kind of what I am looking for.
And spotlight needs a way to search within results. It returns too much data. Like trying to drink from a fire hose if you ask me.
I organize by drives on whether something is a Project, a Resource, Application, or Information/Personal. Then by client or project type. Then by job # and date (like a folder for 2004). The main thing I need find for is to look in my projects and tell me if I created graphics for the web that I could use in a presentation. Or, I might look in Resources and find some stone texture to map on a 3D object. Searching for me helps find redundancy and things that could be in more than one category. Spotlight actually makes redundancy, versioning and synchronizing projects MORE difficult. Hopefully Tiger, might have something to synchronize my Projects with my FireWire drive. I could also wish for a "Smart Synchronize" which could realize that File X in Folder Y is newer than File X in Folder Z and that this could be a moved or redundant file so what would I like to do about it; A; replace old with new. B; increment the newer name and move to folder Z or Y, or C; create an alias to the alternate file in Z and/or Y folder. Backing up and working on more than one computer is still a big headache (especially since most often we are behind firewalls and most of the gee-wiz goodness of Backup doesn't work for all the preferences and bookmarks). But that is another topic.
>>"ad space available -- low rates!!!"
Of course!
Hierarchies suck for large amounts of data (when was the last time you went to ODP or Yahoo Directory to find something?)
That (folder hierarchies suck, search rules!) is one of the main hypothesis behind Simpy [1], a social bookmarking service with tagging and full-text search (think of it as a better and prettier delicious), so there is even a FAQ entry about it:
http://www.simpy.com/simpy/FAQ.do#hierarchies
[1]
Simpy's demo/demo account, to see the goodness of bookmarks without hierarchies
Simpy
...that needs to be done on the data entry side of things before this can really work.
Personally, I don't see how this is going to work with things like pictures and audio files. Not to mention proprietary files that might be harder to parse and categorize. If you have digital images from a digital camera in JPEG format, EXIF data could be searched. But think about how much useless data is in the EXIF portion of a JPEG. If someone types "Candy and Me in Waikiki Last Year" as their search, is the EXIF data for the picture they want to find going to actually contain that information without them having to enter it?
What about MP3 and Ogg Vorbis files? You can store all kinds of information about the file in their tags, but again, the user has to do this at some point. Think about the number of word processor documents named "Untitled-10.doc" or whatever default the application provides.
And what about files that don't have any facility for storing useful metadata that can help assist a search like this? Standard WAV? Basic AVI?
If anything, the OS is going to have to force the user to input identity information that goes beyond a user id, password, first name and last name in order to stick signatures in files. And that still only resolves the issue of files moving from one system to another. It doesn't help with files on the same system.
Applications will have to force users to enter decent metadata for files that aren't text based. And... we're always going to have the problem of people filling in those forced entry fields with garbage and then not being able to retrieve their file later.
Once you start talking peripherals and gadgets that generate data (digital cameras, digital music players for example) of their own without any useful way to input metadata, this really becomes messy. If speech recognition were made to be much more accurate than it currently is, *maybe* these sorts of devices could force the user to name their photos verbally before uploading to their computer. Still iffy if you ask me...
-"...bad old ideas look confusingly fresh when they are packaged as technology" - Jaron Lanier (Digital Maoism on Edge.o
Should be renamed birth of meta-data querying.
d oc
Instead of what I do in windows, where my file names are like 2005-06-07-Assignment-1-Accounting-Due-end-month.
how do you get X files to someone not running this blessed OS?
Tom
Someday, I'll have a real sig.
If anything, Spotlight has encouraged me to introduce more hierarchy into my file organization scheme. Before Tiger, I tended to avoid deep hierarchies, because it takes time to dig through several folders to get to a file. It was easier to just throw everything into the same folder, open that folder and then type the first few letters of the file you're looking for.
Using Spotlight, I can do that without having everything inside the same folder.
I organize files hierarchically, but I tend to open them using Spotlight.
You are thinking that the only way to organize data is hierarchally.
/home/username/work/myproject/main.cpp
Consider a document:
Another way of thinking about it is to have a document called main.cpp with the following tags:
author: username1, username2 (maybe more than one author)
category: work
projectname: myproject
filetype: cpp
date: 20050505
version: 4
status: checked-out
reviewed: no
whathaveyou: something
As you can see you have the same amount of information attached to the file in both cases (about). The difference is that with the second case it is not hierarchical.
With the hierarchical approach you have to know higher levels in the path to narrow down the number of possible hits. With the non hierarchical approach you can narrow down the number of possible hits by picking values for the different tags in no particular order.
The tags that you assign to files can be different for different files.
All this said, I don't think the folders are going away any time soon. I think we'll have both.
The Internet is full. Go Away!!!
How can endless little searches this ever top a well organized project directory?
If there is one thing that Apple pushed, it was the folder-and-document metaphor and GUI. Now that Apple finally comes around to discovering how inconvenient it is, and replaces it with other standard retrieval techniques, they are being hailed as "innovators".
You surely could use this meta-data to make folders?
It is simply a feature that you can or may not want to use.
It would almost certanly have work that way for backward compatabilty. Consider haveing a webserver on a Mac with this file system. The URL is going to have to conform to the current spec.
You know that memory trick, where you remember a long list of items by mentally walking through your house and assigning them positions? There's a huge chunk of our brains that's devoted to remembering *what* something is based on *where* it is.
... it had a name like Exercise or Expendable, I forget ... Now I need to find it. What do I remember about it? That I saved it to the Desktop.
...
So for example: 5 or 6 days ago I downloaded a plugin for some blog package or other, written in php or perl I think
That kind of thing will always have a place in my Finder. I like metadata search too, but I'm just not with-it enough to give up my brain's best way of remembering things
Opera's M2 has been powerful enough to eliminate the need for folders for YEARS.
However, while I love the idea, it's quite different from a system-wide version of search-oriented file-finding in that M2 is dealing exclusively with text.
Making it work for a whole filesystem, with text formats that may or may not be readable by the OS, is a whole 'nother matter, and far more difficult to accomplish.
I wish Apple (and eventually Microsoft, and some day Linux; hey, Reizer's already there...) the best of luck into such a large undertaking.
btw, does anyone know of a good iTunes rip-off for Windows? I love iTunes, but I don't want all my mp3's relocated to some random place on my hard drive. Basically I've love to see something like iTunes for searching and playing, that uses links to the files exclusively.
Such a feature already exists in Outlook, you can have all your mails in the inbox/sent items and just put them in multiple categories. A mail from mom on saturday dinner plans goes to family, mom, plans category so that in each of those category views, i see the same mail
That said, I havent seen too many people using this feature. People still like to physically move the mail to 'family' folder and then to view a bunch of mails, hope that they have the same subject or date range.
People have been trained to use the folder paradigm, changing would take quite a bit of effort
I don't know much about these new database file system schemas. I was just curious how a file's uniqueness is determined? Since a flat file space means that "foo.txt" can (and will) occur several times in the same physical page, the system obviously can't use a traditional naming schema as the primary key in the database. Perhaps the system uses something like an MD5 checksum to determine uniqueness, and just stores the filename as metadata? If so, it would be cool if the hashing algorithm was standardized, as that would do wonders for P2P file system searching and data mining.
I think the article makes interesting points but also think they could have picked a better source than the people who designed the interface of wmplayer. Talk about an interface nightmare.
I'm assuming they worked on the recent versions though.
The first computer I ever used, the Apple ][, had no concept of folders or directories. Are we really going back to the late 70s / early 80s? I guess I won't need a time machine after all. W00t!
If Spotlight = Google, saying that Spotlight is the death of folders is basically the same as saying Google is the end of bookmarks. And that's just not the case. The folder structure isn't going away, Spotlight just makes it more usable than ever before.
LordBodak's journal.
If they can't handle a right-mouse button, it's no wonder they're frightened of folders.
DreamLogic
No matter what, you'll always have a visual representation. Be it saved searches like GMail, or something else more like the Finder, you will always be able to visually navigate the data.
That's important. If you can't visually navigate it, then it's far too easy to lose stuff. It's just that the bulk of your organization is going to be done by a search engine. What's nice about that is that you can retroactively organize things. Ever had a pile of downloads and wish that you had organized them more? Well, now you can!
It can also be a tool for organization, not just the end of organization itself. Extending the cluttered dowload folder above, the first thing you could do is break the downloads up into groups ordered by date, broken by weeks. You could also search for things that have never been looked at (creation date is the same as modify date).
It also means that multiple people can share the exact same filesystem, but look at it many ways. Your children may only care about the games, email and webbrowser. You probaby care about these things, but you also care about your work.
It takes some abstract thought, since no one has a system that really makes it perfect yet, but Spotlight is a huge step in the right direction, and when we get there and polish ip up, it'll be a boon for everyone, from Grandma to Larry the Bitter IT guy.
Slashdot. It's Not For Common Sense
I can make directories & sub-directories, I can create links to them however I like, I can "find" files in those directories and I can "grep" what's in them...
Gentoo Linux - another day, another USE flag.
ibook G4 owner since November. First Apple product I've ever purchased, have loved the look of OSX for ages, been wowed by the gui and all the little things that make OSX so fun.
Have always been anal about file structure with my Windows and Linux boxes. Always fussed about location of page/swap files and partitions. Constantly cleared out temp folders. Forever concious about the level of defragmentation of the file system (ntfs or ext3). Etc etc etc...
Only recently realised that I do NONE of this with my ibook! For the first time, I'm totally content living in the GUI. Don't get me wrong directory structure is still important, but all the fussing is gone. And I'm not saying the gui is perfect. I AM saying the labour of creating folder hierarchies and organising files is a non-event with the ibook.
Yeah, I for one welcome the change.
Try not. Do... or do not. There is no try. ~Yoda
I don't think there is a significant time savings. I'm almost sure it's a penalty. That doesn't really matter these days. As long as it can serve up something like Slashdot (lots of static file accesses), then it's fast enough.
Slashdot. It's Not For Common Sense
I'm not all clear on how these new file systems compare to the current crop, but what about the current problems of file quantity restrictions?
I know Windows will just about cease functioning if any single directory contains more than 15,000 files, and I've heard tales about Tiger having problems moving as few as 8000 files from one location to another. Do these new filesystems have the capacity to store that much data in a single location without crippling the machine?
I'm not tense. I'm just terribly, terribly, alert.
This isn't a dupe, but how many ideas and concepts is Apple going to "kill" before we ignore the media hype? The death of hierarchal structure? Is Apple going to put everything in a flat file or a relational database? No. I find this annoying since it's on the heels of the "death of linux" story.
I also like a well-defined folder structure, regardless of the beauty and utility of Spotlight. Maybe I'm just a belt-and-suspenders kind of person, but I really don't like all my files in one big directory. At any rate, I do agree with you, I think there should be a button that starts the search after you finish typing the search terms, or a way to activate/deactivate live search. There is a simple, if inelegant, way to accomplish this: type your search term in any text editor then copy and paste it into the Spotlight search field. This will return your result much faster than live search.
They did it. It's done. It's called Spotlight. I no longer remotely care where my files go (except for the /Users/ legacy directory.
Metadata rules my system. Files are tagged in every way I can think of. If I think of more ways, I add more data. My files can now be arbitrarily combined, sorted, diffused, grouped, ranked, browsed, and displayed.
I now think of my files as bits of light in the fog. Spotlight swirls them around and collects them into patterns that I dictate. If I find a pattern useful, then I save it for later reference. If a saved pattern is no longer needed, the pattern is removed and the dots swirl back into the ether.
Hard folders or directories now feel clunky and antique. I feel limited and constrained by them, but they are in some cases still unfortunately necessary. (I can't combine all of my web pages into a flat folder for example.) But for almost all of my files and projects, smart folders provide all the organization that I need; be it a loose collection of files or a tightly constrained selection of multiple levels of metadata.
I was trying to install the X11 package in my mac-min w/ Panther.
Inserted the CD. The install applet didn't find my package. I went to terminal. I ls into the cdrom and inside the System directory in the cdrom I saw the X11 package. So what? The Installer only worked on GUI and the GUI Finder won't see it because is designed only to access certain folders. Using the terminal I copied the damn package to my desktop and Now Finder sees it and installs it (thank God this was not OS9).
Here's a new paradigm to the team of dudes/dudettes in Cupertino California "enhancing" OS X.
KISS!!!!!!!!!!!!
- these are not the droids you are looking for -
"I brought you into this world, and I can take you out of it."
*cough*link*cough*
Call me old fashioned, but look... if tagging and labelling were the way things were done intially for decades and along came this really cool thing called folders - people would go apeshit over that. For those with short attention span, change is better than substance. It's called thoughtfully organizing your work people. It works. Use it.
FOLDERS! Yah! Soon, us Benders will rule the world any you all can kiss my shiny metal OS
rewriting history since 2109
Yeah, but you have to make the links, each time you add a file to your archives. We're talking possibly dozens of folders here. With this, you set up your smart folders once (and add one when you get a new client, or every time a new year rolls around, whatever). As long as you fill out the metadata in your files, it will categorize them for you. This is much easier than adding a folder for, say, filings made on tuesdays, and then searching for all filings made on tuesday, and manually creating links to them.
====
Crudely Drawn Games
They want to do away with directories, but say nothing about drive letters?
Can you be Even More Awesome?!
In part, I agree with you. I'm anal about my folder heiarchy as well. But the problem comes when I have a file that's buried in my d: drive amidst 50 folders, 25 sub folders, and 65 sub-sub folders. Especially when I've not opened that file in 2 years, but remember that it had something valuable in it.
I think that it truthfully depends on the application of search technology, but there's no reason why you can use both methods and win both ways.
Mercy was given to me by Christ...I must give the same to others.
It doesn't label the disks, but it does allow you to keep them all in one place and is daisy chainable. I'm planning on getting one soon to house backup discs.
Really cool automated CD storage unit
You make some interesting points, particularly about Spotlight's inability to use regular expressions. Spotlight does, however, have a simple query language that supports boolean expressions, as described by MacOSXHints, and Apple's Spotlight Developer page. You can also get some idea of what the language syntax looks like by creating a smart folder and doing a "Get Info" on it. You'll see your search terms expressed in this query language. I don't use Spotlight extensively, so I really couldn't say whether this is of any value to you. Check it out and see. I expect a 3,000 word report by Friday :-). All the best.
The Shareaza P2P client (www.shareaza.com) has had "virtual" folders which present things to browsers for quite some time. e.g., a folder which displays all MP3 files with a certain artist name in the ID3 tag, or a folder that shows all files you have manually tagged as something or other.
There is one thing that's a little tough with folders. You have a resume Word document, and you need to save it somewhere. Where do you save it? Which folder? Job? Personal? Career? Job Hunt?
What if it fits in all of them, somewhat, as it does in this case?
Oh, shortcuts/symlinks? A little hacky and hard to maintain. This is one of the problems solved by tagging systems (see below for an example). With tagging I don't need to know which folder my resume document is in, I'll just tag it with all relevant labels, and I'll find it later on in no time.
http://simpy.com/ is built around this concept, and so far its been working great and people are loving it, so I think tagging is an improvement over folders.
Simpy
So, the metaphor is like sets or bags in mathematics rather than files and folders.
"I'm not impatient. I just hate waiting." - My Dad
Chris is the bomb! I saw him in that movie where he said that thing to the guy. That was awesome!
So, someone thinks that typing in a search query every time you access new data is easier than simply double-clicking on a folder? Well that would certainly suck much ass. I doubt that anyone in their right mind would get rid of folders. They work fine as they are; queries are a great addition, not a replacement.
We apologize for the inconvenience.
For webmails, I consider GMail superior because it let's me do half of what I want (save information in "multiple places"). Other webmail clients let your save to folders, but they aren't heirarchial.
What I see as an ideal file system would be to have virtual folders. Store the data in a central location and then use a virtual directory system so the information could be "stored" in a heirarchial manner. So my invoices could be access through Client->Order->Invoice or Vendor->Invoice or Hardware Type->Invoice In each of these folders, the files would appear as if they were located there. It would appear as if the file actually existed in each folder, but would only be saved in one place. I could then move the file from one folder to another (Orders-New to Orders-Closed) or delete the file altogether. When deleting, I would have the option of deleting the file all together or just deleting it's relationship to a virtual folder. Say I save vendor invoices for 1 year, but client orders for 5. After one year, I can delete the invoice from the vendor folder, but keep it in the client folder. After 5 years, I could delete it from the client folder and choose "Delete All" and purge the invoice from my system all together.
Folder virtualization would also help with legacy applications. I have a few old games that I dual-boot to Windows 98 because they were hard-coded to look for certain system files in a particular place, but Windows XP saves them elsewhere. By having virtual folders, I could imitate the Windows 9x structure and the Windows XP structure.
We have virtual CPU's and virtual PC's, I don't see why we couldn't have virtual folders on top of all that.
Free MacMini
As long as you fill out the metadata in your files, ...
That's really the issue at hand. Will the unorganized person fill out metadata in order to make their system become organized via search? Does the organized person need a method to search for information that, odds are, he or she has already sorted into appropriate groupings?
A search tool is useless if you have to add the search terms this file should be found by. It will be the one file that you did not tag that you will need to find.
Maybe dumping everything into a single area makes sense for some folks, but I shudder to think about it. I work in the legal field and every attorney and paralegal in the office saves documents in case specific folders. This becomes especially helpful when, two years after the fact, you're asked to track down some obscure brief, correspondence, or the like.
:-/
Yeah, however the thought behind a no-folder system is that these old documents should then have had their metadata written to them by the application saving them, so when you need to look them up, you don't need to know/remember where they were stored, but just create a virtual folder (Windows Longhorn) or perform a search (OS X) on it with your criteria (say, the department name and year) and voila, the folder/search would instantly contain the documents you were looking for.
However, I think this system feels very fragile as so much depends on the metadata. Let's say it the metadata wasn't there for some reason, you didn't know how to best specify the keywords, or the metadata was somehow corrupted... what are you going to do then? Look in a computer-generated illogical folder hierarchy, or just a huge pile of files?
Beware: In C++, your friends can see your privates!
But I love my folders!
:(
The simplicity, the logic, the clean, the everything, the structure!
Folders, I love folders!
Dont take my precious folders away from me!
I will dearly miss them!
Decent question, wrong answer. I'm not the kind that has folders nested 10 deep, but for me, it goes in ./job/portfolio, along with anything else resume-ish. Job leads and tips go in ./job/leads, which doesn't exactly fit resume, you know? I might also have career as a subdir, and so forth. Most things that are fairly hierarchial.
I'm not an opponent of meta-data, I just know that it's hard to come by, and I don't want filesystem navigation to go away. If you want to know what I think can be searched for in a meaningful way like this (it's a pretty small subset of files), ask me sometime. It's currently a SQL database that has the most complex relationship diagram of anything you're ever likely to see...
For computer programs to consistently find files, there must always be a consistent way to do so. Folders will never disappear entirely.
For humans to find files, it's a bit different. Logically made folders work great most of the time. With symbolic links/aliases/shortcuts files can exist in more than one folder and/or have multiple names. Folders can even have multiple names. But what are folders? Folders are a name given to a list of files (though, files must exist inside at least one real folder to exist). Search labels are really no different (except for the disk allocation part); with a little modification, one could even save files into them.
I am getting bored drawing the connections here, but really both can exist quite happily. Doing away with folders completely would be a very difficult task indeed, and IMO foolish.
Meta data can be simply implemented on most FS's now, using symbolic links. The trick is hiding them from the directory listing when unwanted (and making modifying them extremely simple), and making sure they get copied/moved/etc with their associated file.
Search labels within search labels can create the nested folders that most of us use correctly. Again, with a bit of modification (like changing meta data so it fits a search), they can act like normal folders, maybe even a designated "Unfiled" real folder for those that have not initially been saved in a real folder.
I see it all as win-win, and even programs can take advantage of finding files inside of search labels, as well as in the "real" filesystem.
-JDS
With a database core, everybody can have their own VIEW just like a database.
This is the point of a DB based core.
If you want a traditional folder type VIEW, then so be it, if you want a VIEW based on metadata attributes A B C etc, so be it!
No, it's irrelevant. The system offers a new way of organizing, not more. If you have already organized everything in folders, that's fine, but if you're organizing new things, you might just as well use metadata and queries and have a more flexible organization.
Of course the user has to tell the system how to organize things. Whether this means creating folders or adding metadata doesn't matter.
Meta data is pretty cheap, actually. It's provided by authors of documents, web pages, blogs, etc., augmented by consumers of these resources, can be automatically extracted with software tools, etc.
... and, you got me curious about that RDBMS now, care to share?
The problem with the above example is that we are talking about a handful of files, while the limitations of folders show most when dealing with a large number of files and folders. Even you used the word "resume-ish" - clearly, things get fuzzy, and it becomes increasingly harder to find a perfect folder for your files.
Simpy
What many people don't seem to realise is that using a purely metadata filesystem can imitate the folder heiarchy by simply giving the old folder name/s as a keyword.
All in all, it's much more efficient. You get dynamic viewing of files, and saving and loading is much quicker, because you don't have to trundle through multiple folders to retrieve the file/folder you want. To make saving much quicker still, you would be able to select the 'keyword group' that you used for the previous saved file/s.
Why OpalCalc is the best Windows calc
Did some Mac programming many moons ago (pre OS 8 or so) The older file systems didn't really have directory paths like some one from DOS/Unix thought of them. Sure there was a finder and things where displayed like they had a "path" but you had to go through the toolkit and you couldn't have two files in different directories with the same name. It was bloody odd and somewhat difficult to use from a programmer stand point. (Any other crusty Mac coders want to clarify? Hypercard stack people excluded. >;) )
I say bully for anyone trying something new on file organization, but I think the battle is more on the users interface to the files and not the actual storage. Making people dig for things is rather low tech anyway. PC files should be organized by concepts and not location (again, as the user sees them) If I want to look for files, I should put in "Bob" ( a predefined concept that I put in) and get back pictures of Bob, documents I wrote (for/with/to) him, e-mail, and even files that I got from him. Add concepts, narrow the search and you have your file. Apps would of course have to be set up to tag as much data for you (Bob's e-mail was on the message that contained the file, etc) But it's a good start.
We can't hope to reach a semantic web until we get used to dealing with all data in a semantic nature. The problem is that issue is that it goes against the profit model of companies on the Net. (That only want you to get lost in *their* poorly designed site. Anyone else miss the *old* CDNow?) I can see a start with Semantic Shopping(TM) But you need a vender netural, Froogle-like set up where you can look for the concept "Rock" and "Top Forty" and get links to buy songs, CDs, and videos of the related query from all vendors. (An yes, at this rate it will be Google doing it.)
Anyway. I'm sure what ever it is it will be Insanely Great (TM), The Next Big Thing (TM) and if there is any merit to it Redmond will quietly add it at some point after Apple has ditched it... and we'll then be able to have comments about way back when Apple was ahead of its time.
It's called a Document Managment System, and I've been kicking myself for not writing one at work. For some reason, people who otherwise have good organizational skills - and understand how an office filing system works - completely lose it around PCs. Couple that with share names and drive letters, and people completely lose track of that single folder with ten years worth of documents all lumped to together, all without properly descriptive names.
I only wish I'd patented the idea, because it looks like it's going to become popular. Granted, there are DMSes that already exist, but none that I'd consider feature-complete for a work environment.
Fred
"A fool and his freedom are soon parted"
-RMS
Spotlight only complements the Finder, and will never replace it. Here's the Real World Analogy:
The Finder is a table of contents from a book. The book is your data. The drives, folders, sub-folders are chapters appearing in the table of contents. If you want to find things related to "The Partition Function", you look in Statistical Mechanics->Ideal Monatomic Gas. Those are chapter headings in a nearby book. Or are they folder names?
But what if you want to find things related to "Lennard-Jones potential"? It doesn't appear anywhere in the table of contents. What do you do? Try to remember where in the book you last saw it? NO! You look in the index. There it is, page 259.
Spotlight is an index. Finder is a table of contents. The two get along just fine together.
The hard disk has no concept of hierarchy. It's a big, flat space full of blocks. You store data in those blocks.
Hierarchy is added by the filesystem you use. For a simple example, the FAT filesystem keeps a list of the root directory somewhere in that flat space. It contains pointers to blocks of data which could be files, or could be other directories, or could even be additions to the same directory.
The concept is basically that instead of all that, you could add metadata to your files that describe what those files are, what they contain, when they were made, etc. A lot of this metadata can be automatically generated (a lot of it is already). Then you build databases to index and sort this metadata. Then, instead of a hierarchical system to organize your files, you query the database.
And hey, this doesn't necessarily have to *replace* hierarchical organization. Some people will never do that anyway. But metadata structures like this can be built alongside hierarchical organization. Who says you have to organize in only one way? That disk is just a big flat space, after all, you could have several different ways of looking at the same stuff without too much effort.
And this database of metadata doesn't have to be limited to just descriptions of the files, it can contain the content as well. All those word documents? Read the content, build indexes based on it, then you can search for keywords in them as easily as anything else.
You can save these searches as well, so as to make repeated use of them easily. Heck, you don't even have to abandon the hierarchy concept, just turn "folders" into "named searches" and you can organize it just the same. Like a folder called "Word Documents" which contains all those, and it has a subfolder called "Stuff I worked on in the last month" which is all those word docs you messed with last month... Makes finding things a bit simpler, don't you think?
Yes, this indexing takes up space, but storage space is cheap and getting cheaper, you know. Gotta use it for something.
- Give a man a fire and he's warm for a day, but set him on fire and he's warm for the rest of his life.
Yeah, however the thought behind a no-folder system is that these old documents should then have had their metadata written to them by the application saving them,
.doc files have the title "Sean O'Hara" because Word assumes the first line of text in a document is the title.
And we all know how well programs are at figuring out what the metadata should say. 90% of my
However, I think this system feels very fragile as so much depends on the metadata. Let's say it the metadata wasn't there for some reason, you didn't know how to best specify the keywords, or the metadata was somehow corrupted.
There's also the problem of remaining consistent with files created on different days, sometimes by different people -- on Monday I create a file where the metadata says "foo bar"; on Wednesday you give me one that says "bar, foo"; and on Friday I make another file that says "foo-bar".
Les Miserables Volume 1 now up with my reading of
I love spotlight. It works great so far.. But why doesn't it work in OPEN dialog boxes? When I save a file, the window needs to have a Standard/Advanced option to make it oldstyle and Spotlight friendly. Biggest thing, I can't choose File:Open and use a Spotlight search to find the file. I have to use the top-right icon, which isn't really a habit and I'm not sure will ever be a total habit. Needs File:Open:Spotlight bigtime. How about having available Volumes in a Sidebar so I can copy files from Spotlight searches onto my USB Flash Device or my iPod. Can't do that unless I use the Finder. Seems it's only partially implemented.. Why bother? Getting us used to using it before fully implementing it? Why not just implement it now. I'd use it solely if I could copy files between volumes, setup 'where' files copy TO on each volume, etc.
Everyone knows folders must crease to exist.
(Ba dump)
So now I want every brief from 1995-2000 in a folder, but don't you dare move them from their case folders.
Are you going to make individual links for each one of those?
Wouldn't it be so much better to have a smar folder that says:
Type: Brief, Dates: 1/1995 - 12/2000
and fills it all in for you?
T Money
World Domination with a plastic spoon since 1984
I'm skeptical about ditching folder concept completely. For example, the labelling in GMail is useful but you soon build up lots of labels - what's next? You want to start organizing them into folders of labels.
Labels like 'AppleApps' and 'WindowsApps' both should go into 'Project' -- not on two ends of the alphabetically-sorted label list.
If this does happen, it would mean the end of all hierarchical storage/navigation. You wouldn't even need hierarchy among labels, though it would be handy in some cases. Assuming that ALL folders are just search queries, /etc/foobar would just be files with the labels etc and foobar. It sounds like a very interesting future possibility
then, for instance, files in
that is, if the search function is really good?
If you can trust the computer to find a file, why not let the computer organize the files for you? Either the computer can look at the content of the file (or you can tell the computer about the content somehow), then the computer will put it in a folder or organize files in some other way you don't even need to know about.
This assumes the operating system software is good enough that we can trust the computer to organize everything for us and still find whatever file we ask for.
I am so smart!
I am so smart!
S-M-R-T!
I mean S-M-A-R-T!
I am honestly trying to wrap my brain around why "desktop searching" is such a big deal, and I can only think of a few possibilities.
a) MS will copy Gnome in future versions and users will be stuck with Spatial file navigation.
- rendering any file/folder deeper then 3 levels a nightmare to get to.
b) Mom&Pop Users have become such cattle that they refuse to learn how to find stuff that they download or save, and search companies see $$ with all those helpless Baby Boomers now that the geeks have moved out of the home (or they are now in nursing homes and we got kicked out).
c) Through a magical process that has eliminated the need for human intervention and good IDv3 (metadata) tags, these new searches have indexed all the lyrics to your MP3/movie collections so you can now search your collection for:
"That song with the girl who tries to act like an '03 version of Madonna"
"that tune that had "some day I'll get laid" in the chorus".
Programs like iTunes that are a Db are fine and good, but only if all the data needed to search is there in the first place.
Teaching end users proper file structure will last a lot longer then this hand-holding, because at the end of the day the folks being "helped" are still idiots.
Lets just give them a remote controll, and scroll the programs/files on the PC like the TV-Guide does for thier TV shows. "OOh, freecell is on at 7:30!"
-------------
c:\mp3\rock\ACDC\Who Made Who\ACDC - Shook me all night long.mp3
Result: music (to some)
Start -> New Search -> Music -> ACDC -> Who Made Who -> Shook Me all night long.
Result:
Would you like to buy ACDC music online?
Are you interested in an ACDC Concert?
Search Ebay for ACDC stuff?
Would you like to know who made you?
Are you searching for God?
Would you like to meet others who have bad taste in music too?
Are you stuck in the 80's and need help?
Do you want $$ to get out of your parents basement?
yeah I can't wait to see the financial models that predict Desktop Searching being profitable.
A little education sure will go a long way to avoiding pain.
"The price good men pay for indifference to public affairs is to be ruled by evil men." ~Plato (427-347 BC)
To make any filesystem searchable, you have to store metadata on the platter. The traditional folder structure is user-defined metadata, and it seems to be working pretty well so far. A database filesystem with full-blown user-defined metadata (more than just folder names) will use more space on the platter, making disk access slower, data density lower, etc. I don't think folders will be dying any time soon.
This post expresses my opinion, not that of my employer. And yes, IAAL.
Am I the only one to find Gmails system of having no folders kind of annoying? I mean, I like folders... and isn't labeling just like folders anyway?
Meh.
About what a Folder is. Do you think your hard disk is divided into little named areas where all your files are kept? No, your bits are scattered all over those platters. Your filesystem already is a database, albeit a limited flat file database. Or should I say FAT file database (File Allocation Table, and yes, extfs2 has those too).
The issue is that your FAT (database only lets you assign one location key to each instance of a file, while a nicer database will let you assign it many different pieces of metadata.
When you "save a file in a folder" you are labeling that file with a simple piece of metadata - the folder name. When you "open a folder" you are calling up every file with that label, and no, they generally won't be right next to each other on your hard disk. If you have more sophisticated metadata tagging ability, you can just "save it into multiple folders", or just the same ones you would have anyway, why is that such a confounding paradigm shift for people?
So the issue is, do you do this in userspace, with the same old FAT in the background? Or do you replace FAT with something nicer? Spotlight (unless I am mistaken) is a userspace tool on top of a standard FAT type filesystem, while Apple is proposing moving to a relational database for the actual low-level filesystem, at the kernel level. In that case, I don't see why you can't browse folders uing your favorite Finder or Explorer whether they are database, FAT, or other. Userspace tools might not have that feature, but if a database is your kernel's filesystem, then whats the difference to you? You can search and/or browse a regular filsystem, a database just makes the search part much nicer.
I'm -so- there.
Vague qualifier of lacking 'solidity' notwithstanding, Beagle, a free software desktop search program, was up and running before Apple shipped Spotlight. I'm not saying that Beagle was the first to do this job, but since it was distributed before Spotlight, I don't think that it is fair for Apple to get credit for being first here.
Digital Citizen
However, I think this system feels very fragile as so much depends on the metadata.
This is the biggest obstacle to relational file sytstems, of course. Without good metadata, everything becomes hard to find.
The solution to this is multi-layered.
1. Index everything. All the content. Not just the first few words of the document, but every word in the file. This takes storage space and time, but space is cheap and getting cheaper and computer speeds are already pretty far above what most people can reasonably use anyway; slowness in computer systems is due to sloppy coding and design more than it is anything else.
2. Keep good metadata. Just as it's hard to find something in a folder hierarchy when you don't remember where you put it, it's going to be hard to find something when you didn't put good keywords on it. This will take some effort by the user, of course.
3. Build it alongside or on top of existing filesystems. Nothing prevents you from building a relational filesystem for files already in a hierarchial system, or using both simultaneously. The disk is a big flat space already, any organizational structure is added on top of that. This will make transition simpler, or simply let people choose how they prefer to work.
- Give a man a fire and he's warm for a day, but set him on fire and he's warm for the rest of his life.
We have files that are needed in multiple places, but we don't want to duplicate them for fear of losing sync and for sake of storage efficiency. Could be MP3s that are "Jazz," "Christmas," and "Instrumental" or a common character set used in multiple LCD products. Shortcuts or common libraries can be handy in a folder heriarchy, but are unweildy when one file is a part of 50 projects.
:-/) that I can drag files into that are then tagged with a piece of metadata, probably just the name of the bucket (FooWidget or Funkadelica). Now the bucket can forever live on as an indexed search.
... unless I double the amount of metadata needed to protect against this (see "I'm lazy above").
...
What I'd like is a named bucket (aka a folder?
Why?
I often know what I want in the bucket before I do the first search, just like I know what I'm going to name my folders. I don't want to selectively re-edit every single file's metadata to add a new project's tag. I don't want to create a text-based INCLUDE file just to populate some metadata. What a waste of time. _I_ know the context of the grouping before the metadata even exists. As I create new files, or discover files that belong in this context, the process of adding the metadata MUST be easier than retyping every time because, like MOST people:
* I'm lazy and won't do it.
* I'll have typos and corrupt the metadata
* I won't remember which exact metadata tag I want to use, but I'll probably remember visually how to find that bucket (tip of the tongue memory issue)
* I'll forget half of the metatags I actually wanted to use right now, but in the future I'll remember them in dribs and drabs as I come across the files or buckets again. The easier it is to add the metadata (no typing, just dragging around) the more likely I'll actually populate it more completely.
I understand that Spotlight gleans metadata from within your creation, but the last thing I want to do is make sure I type in all the correct tags in my files everytime I start a new document.
In fact I should be able to grab multiple buckets and drop them on a single file to instantly add tags that way, too! Yeah! Yeah! That's it!!
Man, I wish I knew how to program something more complicated than a VCR clock
I guess my beef with all this searching is it is after the fact. I'd like able to create metadata before I make a my first file and have the bucket ready to put the file into.
I'm thinking about it, therefore I might be.
How would you recommend to your legal folks who "get" the folders when they ask you if there's an easy way to link cases though case files are in separate folders?
It's not just a matter of saving the file and knowing where it is (for which folders help) but how is that information useful in the overall context of everyday use (linking, connecting, drawing upon, and so forth). That is where labels and metadata come into the picture.
Folders is a fine start, but metadata is what makes data relevant and personal. I'm not talking about vague ideas here. I recently helped an 80-year old CEO move to labels and metadata for about 8000 emails that she kept as reference. Now she says "I don't know how I could've lived without it".
The way I see it, metadata is what opens up a whole new way to store and retrieve files. By all means continue to use folders; this not a mutually exclusive option. Moreover, I've seen older people exploit this concept more consistently than younger users.
I was hoping for more. You had a good joke going there, some Matrix allusion, it was funny. Then you tacked on a worn out Slashdot meme, the NO CARRIER "joke". That just doesn't fit! In the time of The Matrix, I can assure you that technology didn't regress to the times of analog modems. It just doesn't fit, and it ruined my enjoyment of the joke. I'm not trying to be a dick here, just giving you some feedback on comedy.
I'm seeing people complaining about namespace clashes, removeable media, flat file systems, mis-labeling, labeling, and 'lost files'.
People, these are issues you ALREADY deal with.
I've not upgraded to Tiger yet, but I'm dubuous that Spotlight will help me. Much of my data is stored on Samba mounts and is shared with other people.
If spotlight keeps its database up to date by tracking changes you make to files, how can it handle files that are potentially changeable by others? It would require some kind of server involvement or inter-client communication that I haven't yet seen considered.
-db
cd /usr/bin
... if you've seen WinCVS or MacCVS (http://www.cvsgui.org/) then you know what I'm talking about.
ls
Try Search type:Application
Or on MacOS take a look at all the pfiles and see what they can control and what they can't
Try Search *.plist
Or say you want to find a way to make the dock transperent and you search for Dock Transperance. While the real term that the search will find is Dock Clearness.
Try Search Dock Transparency in Prefs, and, chances are, you get Dock Clearness. It's really smart about such things. Or, try search Dock, then browse the results. A whole lot better than sorting through pref tab after pref tab trying each label on for size.
Or that file you saved way back when you don't know the date you did it or what it is about but once you see it you know that is the one you need.
Do you know anything about the file? Then search for that. Nothing? Then search for type:document and start browsing!
Sure I like spotlight but there are some cases where it just fails me mostly because I am absent minded.
That's funny, cause as much as I like Spotlight, the only places it works for me is where I'm absent minded. My current project I can get right to in Finder; I need the search to find the project dealing with vampire bunnies I worked on last year. My last 20 emails I can find something in easier by just scrolling down in the list; I need the search when someone asks me about something they sent three days ago (or, exceedingly commonly, three hours ago).
IMHO, search is never a replacement for organization. I did some support work for a girls' softball team this year and thought I'd try out using Spotlight as my key organizer. So, I created a stored folder for Softball, etc. I found it was an order of magnitude faster just finding the file and opening it rather than going to the Spotlight saved search folder, waiting for the files to appear (Spotlight's quick, but not quick like Finder!), and then finding the file amongst the smaller stack.
Regarding "labels", OS X has had them for a few revisions, although you're limited to seven different colors which can not be added to, and each file can only have ONE label, not many. However, the one label per file is of course in addition to the organization you give your folders, so it fits many but not all requirements. Right click any file and set it's label color; voila! And you can use them as metadata in Spotlight.
The logical extension here are "instant" searches. For instance, sitting in my "Documents" directory, be able to filter the direct contents by the "Red" label with 1-2 button clicks, then be able to flatten the heirarchy and see everything in Documents or below with the Red label, then be able to select a file and see where it actually lies in the file system. I'm not talking about creating a new smart folder, specifying my search, selecting to search only in , etc; I'm talking about clicking a "flatten" and a "filter by label" button in the toolbar and the display changing to only those
find and grep
This is the second comment I've seen in this discussion so far that discusses symlinks as a hack. What is "hacky" about a symlink? Methinks that you may never have had experience with real files in real folders in real filing cabinets, back when that was the industry standard. For a real file that needs to be in two real places at once in a real file cabinet (or system of file cabinets), there are two ways to do it: One, duplicate the entire contents of the file and place it in the second location: label it as a duplicate and put notes in both places (with the original, that the duplicate exists and where it lives, with the original, that there is a duplicate and where). This is a bit of a pain because then you have to keep two things up to date. Two: Do not duplicate the file. Instead, put a note in at the second location which points to the file's existence at the first location. Sounds a lot like a soft link to me. Also sounds like perhaps the most reasonable solution.
Shortcuts and symlinks are difficult because they need extra management - what happens to the symlink once you remove the target file? You get a dangling symlink. :(
This is a good site to read up about tagging vs. categorization: http://www.tagsonomy.com/
Simpy
so where do all the files get organized? and how does this help me at the command line.
"getting rid of folders" cannot mean "getting rid of directory nesting". it's obviously some marketing bullshit for those who never ever ever use a command line.
yes, the UI will suck less someday, we promise. we will unbreak the finder by removing it. or something.
blah blah blah
Because it searches as you type. It's fast for me (G5 PowerMac), and it works insanely cool. Let's say I'm looking for one of short stories, and I can't remember anything about it, cept it had to do with bugs and monkeys. When I start typing the drop down list changes the more I type, expanding and contracting while searching inside folders. So as I type, let's say I finish the word bugs and I still don't see it, so I type in monkeys and it brings it up. It's fast and slick and it's great because it searches inside and out. On windows there is a third part app that does this-> SearchOpia, and it even searches inside folders. And has been doing this for almost 10 years now.
click me
Tagging is good: there's nothing wrong with it. But that too is an old problem with many good solutions. Library card catalogs are an excellent example of metadata done right -- they even refer you to related "searches" (subjects) at the bottom. But it doesn't mean that the library shelves are removed and all 10,000 books are lying in a heap in the middle.
Sure, you can let files accumulate on your hard drive with all the organization of a pile of sawdust, but finding the files by Spotlight or a Smart Folder query is only the beginning. You will want to copy or move one or more of those files, so it is vital to have an application with a good interface to arrange the move between target and destination (e.g., hard drive -> optical). Spotlight/Smart Folders are by no means ready for that, and by the time you add that functionality, you're back to something like the Finder all over again.
I meant it even searches inside files, not folders. Bleah.
click me
Is it just me? It kinda looks like the Outlook UI. So what, Apple copies it and it's revolutionary? Besides, I doubt we'll see the end of folders anytime soon.
If you want to store something efficient, it will allways be hierarchical, even search algorithms usually store their data like this.
The death of directories would therefore mean that the system takes over the hierarchical organisation of your disk and you lose control over it. You would then only be able to access your disk with search tools.
Euhm... is this just me or does this sound ridiculous! Having an efficient search mechanism is one thing, using it to organize your stuff is another. I do think we need fast search mechanism and I think if we could do something like what 'locate' does on *NIX, but then update the db realtime and after every number of actions on disk, instead of manually (or w/a chron job) having to rebuild the entire database.
I do not think people would be happy to have no longer the control over their organisation of files, therefore the combination of regular directories with a powerfull and easy to access search is the best way to go.
Assuming folder/directory disappears, who is going make all your path needs satsified with legacy apps? How about webapps that all map to resources using path/to/file?
On another note, my biggest complaint about iTunes defaults is the "Use error correction while reading CDs" checkbox. I ruined much of my library on importing because I left this unchecked when I first started importing my collection. A lot of songs sound like crap; random distortion really loud, and there's no way to know which songs got screwed until they are playing. Why have an almost hidden preference that will ruin your library if not checked? Perhaps other people have better luck importing with this turned off than I do, but now whenever I use a computer's itunes for the first time I make damn sure that box is checked before importing CDs....
I'm pretty much a fairly sophisticated user but my main tasks are email, web surfing, etc. When I installed Tiger, I wasn't that impressed with most of the new features. Dashboard is cool but not revolutionary and I considered spotlight to be a replacement for the Find command, which I rarely use. Then the other day I wanted to open Photoshop, which is on a firewire drive and nested under a couple of folders. I decided to try to use Spotlight to pull it up. After I typed "P-H-O-T-O-S" I could see photoshop selected as my "top hit." It reduced my interaction with the computer and allowed me to quickly get to work. Personally, I think this should be the goal of all software developers...to reduce interaction with the computer and to allow the user to work. After figuring out this neat trick a few days ago, I really haven't used the finder since, I just start typing the name of a document or application and it pops right up. I described it to someone as the document comes to me....I no longer have to go to the document. I think there's something truly revolutionary about that.
and an idiot
> better eyesite
I think my blind grandmother has better eyesight than you. To confuse a "te" with a "ght," is just terrible
There is an easy way to achieve this with an ordinary file system TODAY. Just use shortcuts or symbolic links to place a file or folder in several places.
Searches make you lazy. Folders are perfect in that they make you remember. :)
I've been using IronSentry's service to archive away all of my email to their service; the handy and immediate web-based search lets me find my old email so conveniently, that I no longer bother sorting my mail into folders. I just read it when it comes in (and even delete it from my local client), knowing I can find anything instantly with a convenient search.
The power of quick searching really will change the people work. It's a bit similar to how people tend to keep booksmarks less than they used to in the old day, since they know they can re-google any site they want, if they forget its particular URL.
Love many, trust a few, do harm to none.
Using tags/metadata/whatever is an interesting idea, but so far it looks like the only implementations are for GUI. Imagine now you have this working with your favorite shell:
instead of cd to a directory, you would restrict files you can access by default by selecting tags. Could be pretty cool, and incredibly powerful.
Problem: how would you address a specific file that, under the current system, you know is in some specific directory? Specifying some tags and then a filename may not result in a unique file. Anyway, how will the OS deal with filenames? Would you still have filenames? Could you have two files with different filenames? After all, filename is just one particular piece of metadata. Could you have two files with exactly the same metadata? How would the OS distinguish them? Would it still use some sort of hierarchy, only hidden from users? Would that be really irritating for power users to have to switch between two different ways of looking at their files? Would the underlying structure lurk under the search/metadata/tag based method like some scary skeleton, poking out at unexpected places and ruining user experience?
AccountKiller
...people who RTFA before modding a post as 'Offtopic'.
You must think in Russian.
I can't help thinking one reason the nested-folders approach has become so strained is because it's been abused so much. As a tool allowing users to organize their stuff -- basically, their applications and their data files -- it was a good idea, and it's still a good idea.
The problem is, when you install Mac OS X, you are installing 30,000+ files and God only knows how many thousands of directories and sub-directories onto your hard drive. Even my backup program is nearly crushed under the strain of indexing them all, and almost none of them are of any pertinence to the user whatsoever.
It's the same story when you install any major application, or install a game, etc. It's not uncommon for them to incorporate several thousand files and a maze of nested folders, none of which are of any relevance to the person using the program.
I believe Apple took a step in the right direction when they introduced application bundles. It doesn't get rid of all those files and folders, but at least they are hidden from the user. I don't think it's enough, I think it's a baby step toward a solution.
The way I see it, the big problem with the hierarchical file system is that it's being forced to perform two different jobs. For one, it's being used by end users to organize their discrete objects -- meaning media files, documents, and applications. These are individual, self-contained items that have some kind of meaning to an ordinary person (who is not a programmer). The system as it now stands is perfectly adequate for doing this job.
Secondly, it's being used as a development tool for programmers to organize the profusion of data structures and obscure components that make their programs. This is where the bloat comes from. This is where the thousands of files and folders come from that have no meaning to a user, except that they strain the filesystem, confuse its organization, and provide a "haystack" of entries for the user's needles to get lost in.
Apple's Finder hides a lot of stuff from the users -- bundle contents and many system files -- but in the long run it should be possible to come up with a better answer to this problem.
"The organization itself derives from, and can only derive from a human mind. Thinking "the computer organizes the data" is the main reason why virtually all databases are giant Mongolian cluster fucks."
I have a large one-time pad of numbers. What should I organize it under?
"Until know about their data, what it is, what it "means" and how it is expected to used I can reorganize it a billion different ways without in any way organizing it in any useful fashion."
Organization comes from the repeatability (pattern) or sequentiality (date) of one or more characteristics. As I hinted above, you'd have trouble "organizing" a completely random collection.
I've had this in windows for a long time. AppRocket(www.candylabs.com) provides a quick search of all your files and folders, and uses past search history to prioritize the results
Google Desktop Search is good for searching content, but for finding files, folders, and programs, AppRocket is my #1 pick
Now, I'm not a big advocate for speech recognition these days, mostly because the technology, while out there, hasn't been implemented very well yet. But a labelling style file system coupled with speech recognition could be absolutely beautiful. I was thinking about the one drawback to labelling, and that is that since typing is a bit slower than speech for most people, end users would be hesitant to meta-label their documents with more than one or two descriptive words, fogetting that as they build up more documents in their computer, searches will become more complex.
Once commerically available speech recognition gets to the point that uses can simply say a word and it is printed on the screen with reasonable accuracy (especially after identifying the individual's unique phonetic cues), a labelling system could be as simple as the following.
To label a specific Finale Music Notation file, say:
"Label this file as 'Business' 'Warner Brothers' 'Composition' 'Film' 'Action' 'Red Hot Warriors' 'Parts' 'Cello' 'Second Draft'"
A transparant dialog box appears in the middle of the screen, and as each word is spoken, it is added to a list of labels in the box.
To open the file again, prompt the computer then say:
"Open music document Red Hot Warriors, cello, part, second draft."
While this is being said, a transparant box appears in the middle of your screen, after each word is spoken, it prints each label being said, to make sure that it's transcribing your voice correctly, and a list of all files that could fall under that catagory.
Since the file type is "mus", a finale music file, it is redundant to add this label, since the computer knows it's a music file to begin with (as long as Finale does this automatically, and for anyone out there who knows MakeMusic, you know they'll be the first pro company to forget to do something like that).
It isn't perfect, but most people don't mind saying a few words, while typing 8-10 labels for every single file may feel exhausting after a while, espcially after working with lots of little files. I'm not saying that speech will make the keyboard obsolete, I think it always needs to be there for more "computer oriented" things like coding, naming files, or when the speech program fails, but a good GUI maes a lot of analogies to real world applications, therefor, we should use whatever technology is available to make more rudementary computer tasks seemless with our environment.
Multiplayer Gaming (defined): Sitting around, discussing single-player games with my friends, at the bar.
This is all fine and dandy, but how would you search for code, like the source code for a program?
No existe.
On the original Newton PDA, all info was tagged and thrown into a 'soup' so when you wanted anything it was retrieved by a search. It completely removed the concept of folders and directories. But people like folders and so psuedo folders appeared, which were simply the results of searches. Apple also pushed the concept of a 'folder' icon opening a window and actually showing search results in the pre-Copland (spelling?) OS8 days.
There was an unknown error in the submission.
On the original Newton PDA, all info was tagged and thrown into a 'soup' so when you wanted anything it was retrieved by a search. It completely removed the concept of folders and directories. But people like folders and so psuedo folders appeared, which were simply the results of searches. Apple also pushed the concept of a 'folder' icon opening a window and actually showing search results in the pre-Copland (spelling?) OS8 days. I don't think that CP/M had folders or directories?
There was an unknown error in the submission.
That should be complement, as in something that makes the other thing more complete, and not compliment, as in an expression of admiration.
Sorry, I try not to respond to little mistakes like these, but sometimes the pressure just builds up and I overflow.
Warning: The intelligence of this post may be larger than it appears.
Right. But in no way does that mean a folder-based hierarchical classification scheme should not be used.
As you aptly point out, the whole purpose of a database-based file system is so that you can view files using multiple classification schemes. This does not preclude the use of a folder-based scheme. There's no reason at all to preclude users from keeping their data organized in a folder-based scheme. It may simply turn into a virtual folder-based system rather than one that reflects how the disk partition's actual directory tree is organized (this is already not true with some filesystems).
A directory structure where I can sort files by location, type, content, product type, filename length, etc. would rock. But a directory structure that can only be accessed via search queries is hardly different from a folder-based system - you're artifically limiting the user to a single means of accessing his data.
You could search for php files you edited 5 or 6 days ago.
Jobs won't rest until he has killed everything that made a Mac a Mac.
He killed OS 9 which was we just about forgave him for as there was a need for something more.
He killed PowerPC which many of us will never forgive him for.
He is now about to kill the Finder. This should come as no suprise at all seeing as Apple have left the Finder to rot for 5 years , so it was obvious he wanted to replace it with some new Stevism.
Well FUCK you Jobs. Your problem is you have overstayed your welcome back at Apple. I always knew you would never leave until you had killed the Mac and turned it into some dumbed down bimbo toy which seems to be your intention.
I may be missing something,but... How is this the death of folders? All this article talks about is a tool to search for files; not how they are stored.
That's what autocomplete is for. Or you can make a list of tags/labels/metadata.
Considering it asks during the install whether he would like to manage his music folder or let iTunes do it for him, and the default is to leave it to the user to manage it.
And thing like mail search and picture browsing are already solved.
<^>_<(ô ô)>_<^>
Losing 'Finder' or any similar filesystem UI, in favor of dynamic smart folders, queries and searches, is a bad idea. In the real world, you put stuff in a closet, in a trunk, under the table, in the attic, in your left pants pocket or your shirt pocket... you devise all these great schemes to know exactly where everything you own should be. When you get to your car you pull your keys out of your pocket. If you have them in your briefcase, you get disoriented for a second. When you leave your car, you feel the keys in your pocket and are reassured everything is as it should be. But, in the real world, you also lose stuff because sometimes you misplace it or you forget just what your logic was so many years ago. So search tools, maps, etc, are great aids to finding stuff so you can once again use them. But they are NOT themselves the stuff you seek. Feel free to throw away a used up map and be confident it's just a map you're tossing, not the destination itself. Also, you're always going to want some stuff to be far away and archived, out of the way, out of sight, out of mind so you can focus on the stuff you care about right now. You really don't ever need to see that invoice from 1989 again, do you? Well, just in case, you'll keep it in a safe place, but out of the way. Enter Spotlight and smart folders. Amazing tools that help us find the long lost stuff. Cooler still is how you can use them as reporting tools. How many different times did you write something about your pyscho ex? Spotlight knows. But the signal-to-noise ratio when using such tools is disorienting and unreliable. If I go to my kitchen to use my favorite chef's knife, I depend on it being in the place where I put it. I don't want to utter "chef's knife" to a 'smart drawer' that suddenly slides open showing me all 10 different chef knives in my household and poke through them all just to select my favorite knife. No, I want to move my hand to the exact spot where I know it always will be; right at the top right of my other 4 premium cooking knives, none of which is a chef's knife, and all in one nice wooden block, on the counter, in my kitchen. Now imagine the chaos of a shared environment or corporate setting. That's where smart folders actually shine. Because now each person in the company can organize the files for which they are responsible as they see fit, and everyone else can use smart folders to cross-reference across departments or use search tools to find specific cases. But Smart folders must remain exactly what they are: a _View_ of an existing organization; not an organization unto themselves. Users must never confuse the two because a file may be found in more than one smart folder. So it's imperative that the user understand that the file really only exists once. Back to my kitchen, while it would kick ass if I could open one magic drawer that give me access to all the chef's knives so I can take inventory, or I can decide that it's time to replace, sharpen or retire one or another, and another magic drawer that shows me all kitchen utensils of a certain brand, I don't want these dynamic slice of the current state of my kitchen to become the organization of my kitchen. Finally, think about this: databases can be searched, sliced and diced in anyway you like. But you still have to organize the data into tables, never repeating the same information twice; any database guy worth his salt will bend over backwards to keep it as normalized as possible. It's not just one big table. The filesystem is no exception. Reality is not an exception. Even your brain can't effectively perceive the world using a model that would be an exception. It can't. So why bother pursuing an organizational system without logical groups, hierarchies, and spatial cues?
Asking people to think is like asking them to buy you a new car
Emphasis on 'WORK'
When I need to back up or transfer a clients work, dragging and dropping a folder to the FTP site or the blank CD is a hell of a lot easier than tracking down all the separate files with a search engine, no matter how effecient it is. And NO, adding metadata won't work because it's a pain in the ass and it's easier to just put stuff in one 'place' If it wasn't, we'd all still by typing pathnames at the DOS prompt.
There is also the issue of encrypted folders. Don't you think it's easier to select an object and lock it and it's contents, than tracking each file down doing what I don't know.
It's so laughably obvious these soi-disant 'critics' don't actually have to produce any work for anyone, except some blah-blah junior-high-level-of-writing 'article.'
If you're working on a big multi-media project with hundreds, if not thousands of different pieces of different media. How do you propose organizing it in a meaningful way without using folders? And please don't bore me with an add-meta-data solution. I don't have time for that.
I shall always think of directories as folders, even as my brain makes a rediculous zerbert-sound.
Long live the folder.
"Love is like pi - natural, irrational, and very important." (Lisa Hoffman)
or maybe i'm starting to become old? time to write a handwritten letter to slashdot now
Question: Do you use the iTunes interface for burning CDs? If you answered "Yes", then you've already used the described type of interface.
:-)
Thank you, please drive through.
Javascript + Nintendo DSi = DSiCade
I tried adding notes to my pics in iphoto but after a while it gets tiresome.
Notes aren't meant for organization. Get Keyword Assistant and use it to make adding as many keywords as you like to each and every one of your photos.
I don't know why Apple doesn't make keywords this easy itself, but I'm glad this plugin is freeware. (I loved it so much, in fact, I sent the developer $10 just for being cool about it.)
By "kill", I mean that the OS (the one I'm designing anyway) shouldn't assume that the creator of a file will assign a particular filename to it at any particular point.
Unix requires the creator to assign a filename when the file is first opened for writing. That leads to all sorts of annoying workarounds for cases when the name isn't important, as in temporary files (witness all the security and related bugs with multiple users accessing /tmp) and emails (which have no need of names, more of "to" addresses).
Internally, filenames are metadata in much the same way as are folders (directories).
So why not let users simply "Save" a document without having to give it a name? Ditto for the underlying API to the OS.
Never mind "Save". Why not just automatically journal any work in progress, and let users easily browse such works by timeline ("whatever I was working on yesterday afternoon" becomes easy to spot, given a suitable clock or calendar presentation) or content?
Historically, much of what we refer to as "files" and "directories" came from the assumption that "save to disk" was synonymous with "make sure this survives a system crash or power outage".
Nowadays, most personal computing devices are as reliable as historical disk drives, sometimes moreso (many have solid-state hard drives), most are networked, and many are constantly "on line" to that network.
So, instead of "Save" and requiring a user to name a file, replace that paradigm with one that allows users to more easily and naturally "Commit" changes to a document, "Publish" the document to a given audience, "Secure" the document such that it never is recorded on certain devices or networks, and so on.
Eliminating underlying OS dependencies on filenames and file folders can free the OS, API, and even the user to focus on characterizing and searching for the data in the documents being composed, not on trying to figure out just what is the right name and/or "location" for a document as it is being composed.
Practice random senselessness and act kind of beautiful.
You trying to get me divorced or something.
I've been talking about this for a while, and wrote a blog entry about the need for such a filesystem months ago...e -is-needed.html
http://obsessiveatbest.blogspot.com/2005/02/chang
Searches don't kills folders, people kills folders!
Omry.
when the CLI dies - never. Spotlight is a fantastic tool, just as Folders are a fantastic tool. They compliment each other. Lets liken them to firearms:
Spotlight is a shotgun - no training required, you'll hit anything thats close to where your aiming. Its up to the user to decide if there was an appropriate 'hit'.
Folders are like sniper rifles - years of training and disciplin required. One shot, one kill.
The Joe Public militia will always use shotguns above sniper rifles - but you'll rarely find a professional army of developers withouts its marksmen.
Scared of flying, pointy things snce 1979!
Ok lets prove you right. I got OS X Up and runing with tiger and Indexed! /usr/bin files I know I have ls on my computer somewhere.
.plist here we go. A huge list of files. What!!!! There are about 20 different info.plist that is not going to help me. Without folders I don't have any context on what they are used for.
type: applications... HEY where are all my
Ok *.plist no results found Ok...
Dock Transparency in Prefs... nope not a thing Um Just dock well it is only giveing me files with dock in it still not much help, for someone who doesn't really know what it is called that they are looking for.
type: document... Holy Crap I never though I made that many documents. Oh wait I didn't a lot of them are OS X help files. Umm Ill guess ill give you a point there. But I would probably be more efficient with finder though.
Well I guess I am still right. with perhaps about the document but still the finder is better.
Don't get me wrong I like spotlight and it works well but I do want to have the finder or a folder/directory structor for those cases.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
The largest and most sophisticated user of system resources is usually the system itself.
There will be folders (directories) in the file system for a long time because they are an efficient and simple way to organise the data objects needed by the base operating system.
Anybody can tack some cross-indexing tool on top of application data. Better let the system backup/restore/delete functions have a clear and unambiguous view of it all.
The second is obvious: the OS automatically indexes everything written to the disk, no matter how (i.e right down at the FS level). Because the index is always up-to-date, searches are immediate and quick.
But the first isn't so obvious. To find what to index on, it's not just searching the data for text strings. Instead, it understands about different file types, and extracts metadata from the file contents using that knowledge. It knows about text files, HTML, Word docs, PDFs, emails, JPEGs, MP3s, MPEGs, and most other common types of file. And app writers can tell it about their own filetypes too, via an API.
This differs fundamentally from the BeOS/etc. model you describe because the metadata is part of the file datastream, not separate from it. That may seem non-ideal, but it has several major advantages today: you can move or copy files with any tool, across all existing filesystems and OSs, without losing that metadata. It doesn't need to be stored in a special filesystem. And most file creators and editors, on all OSs, already understand and write it, so you can share files with non-Spotlight-aware apps on Macs and other machines, and have them all using the same metadata.
It means I can download an MP3, and as soon as I write it to disk, Spotlight reads the ID3 tags and indexes them. Ditto practically any type of file. And all that's here working right now, without the major changes (filesystems, apps, OSs, transfer protocols, &c) that external metadata would need.
At present, the UI for this is very simplistic. But you can bet that more powerful searching will soon be available, if not from the Finder then via other apps.
Ceterum censeo subscriptionem esse delendam.
But that doesn't scale directly to an entire filesystem.
For one thing, iTunes is handling just one sort of file: music. It may play AIFFs, WAVs, MP3s, AACs, and ALEs, but they're fundamentally holding the same sort of data, and -- more importantly -- have the same kind of metadata. They all have lengths, names, last played dates, &c. And most have artist, track number, &c. So keeping track of that metadata is easy -- you store the same stuff for each file.
But on a filesystem you have umpteen types of file, with umpteen types of metadata, so you need to cope with each file having a different type of metadata. Which isn't easy. (That's why it's really impressive that Apple seem to have managed it efficiently.)
OTOH, there's one great idea we could pinch. We've talked about Smart Folders, which are the filesystem equivalent of Smart Playlists. But iTunes also has (dumb) Playlists, which are the equivalent of filesystem folders. iTunes dumb playlists are different, though: a track can be in more than one playlist, or it can be in none (other than your 'whole library' one). And I think that would work wonderfully on a filesystem, solving most of the problems people are mentioning here.
You could use dumb folders just like at present, with each file in exactly one. Or you could put a file into two different folders at once, which would remove the need for aliases or symbolic links. You might even be able to get rid of the Trash, too -- removing a file from a folder wouldn't delete it, and if you wanted to you could tell it to physically delete all files that aren't in any folders.
The one thing that the iTunes-model folders wouldn't have is nesting, but I can't see any problems adding that. Then you'd have a model where smart folders and dumb folders could coexist neatly on a filesystem, giving all the benefits of both an iTunes-like repository and a traditional filesystem, and lots more!
Quick, wait here while I dash off to the Patent Office...
Ceterum censeo subscriptionem esse delendam.
That alone justifies the moderation, here, being so unusually "insightful" for a /. posting. Nobody else seems to recognize when they're unqualified. Or maybe we all just take that as a given.
Spotlight does essentially offer both the scope of searches you describe and significantly more "depth" in terms of the metadata than anything else I've seen. I've got "Smart Folders" set up using date windows for my kids' IDs, and the idea doesn't throw the 11-year-old demographic in my house anyway. They're notoriously sloppy about where they save stuff.
"Fundamentalism" isn't about divine morality. It's about human authority.
No, I don't. I only back up MP3's on DVD's and I don't use iTunes to do it.
CD Audio is so 20th century. *YAWN*
type: applications... HEY where are all my /usr/bin files I know I have ls on my computer somewhere.
... you can't use them from the Spotlight/Finder/Gui interface!
.plist here we go. A huge list of files. What!!!! There are about 20 different info.plist that is not going to help me. Without folders I don't have any context on what they are used for.
.. which is both kinda useful to show up right then and very unlikely to have ever come to my mind had I instead rooted around in Finder looking for com.apple.Automator.plist, but YMMV).
/Library/Application Support/ (which, of course, is done by going to that folder and doing a Spotlight search under there) or just the ones in /Library/Preferences (in which case, obviously, use the Finder's folder interface alone!)
Okay, I misread your original problem as looking for applications as opposed to command-line utilities. Yeah, if you're looking for command line utilities, you should use the command line (well, yeah, you can do it in Finder too, but 'ls' is much easier and sitting on the command line you can try each of them out). Spotlight excludes command line utilities because
Ok *.plist no results found Ok...
You do have folder context, you just have to click on each and every file to get it (the "info" button). You asked for all "pfiles", that's what you got. You want the plist for a specific application, include that application name in the search. For instance "Automator plist" puts the Automator plist right there at the top (along with a mail message from an XCode mailing list which talks about problems with an Automator plist setting
Or are you complaining that you wanted only the plists under
My point is, and was, if you want "all" your potential configuration options, a Spotlight search will get you a lot faster than poking around in the Finder trying to find them. As I thought was obvious, but apparently needs to be stated, if you are looking for the contents of a particular FOLDER, then Spotlight is not the right tool to use (except perhaps to find the folder).
Dock Transparency in Prefs... nope not a thing Um Just dock well it is only giveing me files with dock in it still not much help, for someone who doesn't really know what it is called that they are looking for.
Um, okay. My bad. I didn't look this one up and took your word for it that "dock clearness" led somewhere.
So where, exactly, do you see an option to set dock transparency? Anywhere at all outside of a third party application? And just how do you expect Spotlight to show you to a preference page which doesn't exist?
The Spotlight results seem like they'd be pretty clear to the average user: no, you can't set dock transparency with what's on your computer. Unsaid: search the web and find the haxie to do it if you really need to.
type: document... Holy Crap I never though I made that many documents. Oh wait I didn't a lot of them are OS X help files. Umm Ill guess ill give you a point there. But I would probably be more efficient with finder though.
If you want only documents you created, add "Authors" equal to your name. Again, i'm not sure how a scan of your entire hard drive can possibly be more efficient than narrowing down the list by known criteria (which would include the folder you saved it in, if you know that, but you didn't indicate that originally).
Spotlight narrows the possibilities down. The more bits you know about the document you are searching for (that it had a footnote regarding Othello, perhaps, or that it dealt with evaporative coolers, or that you saved it somewhere in your Documents folder) the more Spotlight can narrow your search.
In the end, it always comes down to picking what you want from a list. The beauty of Spotlight is that it makes that list flat (instead of having to descend your folder tree and look in each fo
Well, a lot can be filled out for you. Date. Filetype. Resolution. File size. The account which created the file (which could come in handy on corporate networks). Length of the song, as well as artist, album, writers, people who were sampled, producer, record label, year of album release. Heck, assuming the smart folders can be set up to do full text or keyword search, you could have a folder of every file on your PC that mentions "taxes" or "student loans".
====
Crudely Drawn Games