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;)
It's already here. It says right in the article that "[...] Tiger's Smart Folders feature, which lets the user save the results of a Spotlight search as a virtual folder that automatically updates as new items matching the search are added to the system." This sounds quite similar to the smart playlists in iTunes eh? I use the smart playlists in iTunes quite a lot, and I'll definitely be using this smart folder feature once I get Tiger.
What's a sig?
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
This advice is probably too late for you, but you can actually tell iTunes not to reorganize your music folder in the preferences.
I agree this seems like a stupid thing to have turned on by default. I also find the behavior where it copies mp3s that you play to the music folder automatically strange. But I guess some people would get confused that deleting a file from their desktop makes it not playable in itunes anymore. *shrug*
FiGZ.COM - A waste of perfectly good web space
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
I'm an avid and early Gmail user but I think it's only fair to point out that Gmail borrowed the folderless labelling system that it uses from Opera's M2 mail client.
As far as email is concerned, labels are an Opera innovation (unless, of course, someone can provide an earlier example), not a Gmail one.
"Accept that some days you are the pigeon, and some days you are the statue." - David Brent, Wernham Hogg
This is kinda how Quicksilver does this with pre-10.4 (Not having 10.4 I can't talk about how quicksilver interacts with spotlight et al). It didn't matter where anything was, particularly, you just type in the name of it, and hit enter. Voila, it opens up in the approptiate program. This idea does take some getting used to, you're quite right. But I think after people try it for a couple of days, then they'll realize that the database (or catalog as far as QS goes) model is vastly more efficient. Rather than having to remember where you put that file you use twice a quarter, you can just type it in, and there you have it. No hunting, no guessing, it's there when you need it and that's what's important.
Game Overdrive - Gaming News
We had this problem at an office I worked at a while back. We were a manufacutring borker broker, and we would get an invoice from a manufacturer that was to go a client in turn. Physically, we would put the original in the manufacturer's file, and put a photocopy in the client's folder. When we were computerizing, my manager thought that we should have copies of the scanned invoice in both the manufacturer's *and* client's folder.
I explained how much extra space this would take, and there were other documents that belonged in *several* folders. This was easily going to chew up all of our available disk space and backup in a few months. I tried to get them on a **relational database**, which stores the invoice *only once*, and cross-lists it under both the client and the manufacturer. When you do a query, either for client or manufacturer, you get the files that apply to the query arguments.
Anyway, my ideas never got traction, and AFAIK, the office is still using paper.
Computers are useless. They can only give you answers.
-- Pablo Picasso
Now, let's take this into the work place, where you have 300 users and a central server. How do the users know they're working off the "official" version of the file from the server?
They should show up as different devices. i.e. I have a repository over here on the server, and over here on the desktop. This is pretty straightforward on OS X. More complex solutions in the future (which would pose some issues) could allow administrators to force all user files onto the network, then keep a "cache" of the files on the local machine for travelling laptops. (Windows has a similar feature now.)
What happens when a user makes a typo when entering meta data for associating files with a project?
Not a GMail user I assume? Look, Labels work just like folders. You create it, it looks like a folder, walks like a folder, talks like a folder, and quacks like a folder. Therefore it's a duck^H^H I mean... Folder. The primary difference between the Label and Folder is that you can have documents under multiple Labels. Links, as we use them today, are just a workaround to make up for this missing metaphor.
This is just more of Apple introducing ideas that will make actual work more difficult in the interest of letting increasingly stupid users write letters, pirate MP3s and surf for porn.
I can only assume that you've never actually USED an Apple. Because I get far more work done on my Apple than I ever do on my PC. Or perhaps you'd like to explain how ubiquitous spell checkers, applications as a file, built-in Java, files getting saved properly even if moved, automatic file associations, Expose, and a billion other USEFUL features for getting work done are only targetted at Pirates and Pr0n lovers?
Javascript + Nintendo DSi = DSiCade
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
Sigh. Go read the Opera website. Opera is heavily opposed to software patents and in favour of competing on merit rather than through the courts.
If they weren't - if instead they were patent-happy and litigous in nature - then Firefox would have been stripped of several of its features, as a great many of them were borrowed from Opera.
And, I didn't say Opera invented labelling, only that they introduced labelling rather than foldering to email way before Gmail did. Had they wanted to, Opera could have easily patented labelling in emails, especially with the way that the USPTO gives out patents to everyone who so much as looks in its direction.
All clear now?
"Accept that some days you are the pigeon, and some days you are the statue." - David Brent, Wernham Hogg
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
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
I'm spending a lot more time replying to these posts than I should. Still, I can't let them slip. :)
:)
A study was published just last year about how the desktop paradigm breaks down when a lot of files are trying to be stored. There's nothing wrong with the folder system from a technical standpoint. The problem comes when you have hundreds or thousands of files that need to be sorted and then found. Your capacity to remember such things is finite. If you know even vaguely what you're looking for ("Hmmm, it was about 2 weeks ago, I think it mentioned nintendo, and James may have written it..."), it's probably easier to find by searching than by trying to figure out if you filed it under James, Nintendo, or the documents that you got 2 weeks ago.
If you'd like to read the study, try and get your hands on the ACM Transactions on Human-Computer Interfaces, June 2004, Volume 11, Number 2. It's quite interesting; a lot less dry than most papers.
Note that every one of those keys is searchable (not directly from Spotlight, but from the search function or the command line).
How are GMail's labels not metadata again?
Smart folders go off of arbitary metadata. I use spotlight comments in much the same way I use gmail labels. Some file belongs to a particular group? Add a keyword (label) to it to indicate that.
I am going folderless as we speak. Look back in my comment history for a long post about how spotlight and smart folders have changed my computer use for the better.
I use spotlight as an app launcher and it works very well. Type in a few characters of the app name and then hit command+enter. Command pops the selection down to the top hit, which is always the app for me.
I trimmed my dock down to almost nothing thanks to this.
No, I never liked QS. I don't know why either.
Have you actually used spotlight? or are you just another it's not windows so it sucks person?
Spotlight searches within files , not just names.
You forgot to label a file?, You can remember the title? type it in to spotlight, one of the results will be your file.
3 seconds you just narrowed down your list of 3,000 files to search to just 10.
I don't use label's a lot in gmail, but I did import all my gmail into my Apple mail folders. Why so i can do an offline search of all 30mb of my email in folders.
Spotlight is close to being a local version of google.
i thought once I was found, but it was only a dream.
Spotlight also remembers the last time a file was used even if it wasn't altered in addition to the last modified date.
mbbac
Mathematically, it's like sets, yes. Sets are core to database theory, and file systems are a form of database. The catch is that current filesystems are missing a lot of relational theory that could improve their usability.
Metaphorically, it's like using a label gun to apply those little stickers that say "Property of XYZ". If you were to lump together the labelled objects from my wife, kids, and myself, some stuff would be labelled mine, some stuff would be labelled my wife's, some stuff would be labelled with a kid's name, and some stuff would have a mixture of labels. e.g. Wife/Husband, Kid1/Kid2, Wife/Husband/Kid1/Kid2 ~= Family Property (Family Property could be an alternative label for everyone.), etc.
Javascript + Nintendo DSi = DSiCade
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.
Someone else mentioned comments, and they do sound like they fit the bill. The only catch to watch out for is that the comments are actually linked to by the file, and not just applied as free text. The reason for this, is that if I change the text of the comment, it should get changed across all files. If it doesn't, then we have a problem. :-)
Javascript + Nintendo DSi = DSiCade
OS X Tiger breaks down metadata in several types.
1. ones inherit to all files (creation data, last modified data... etc)
2. Common files types with metadata. Spotlight knows all about how these types of files organize their metadata.
3. specialized file types. Apple leaves it to the file type owner to write a metadata extractor for it. The API on how to write a spotlight plug in is well published.
You can use the mdls command to see what bits of metadata spotlight has extracted from a file.
#2 covers lots of file types. MS Word/Excel/Powerpoint, pdfs, mp3, pdfs, jpgs.. etc.
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.