Slashdot Mirror
'Type Manager' The File Manager of Tomorrow?
IceFox writes "In the past few years many of us have been introduced to a new type of application, the Type Manager. Most of us are familiar with iTunes, but there are many other Type Managers out there that are gaining market share and a rabid fan base of users such as digiKam and amaroK. Type Managers seem to have that magic combinations of features that makes users love them. I have been taken a closer look at the Type Manager, what makes them so usefull, what they really provide for the user and came to some surprising results. After creating a list of all the traits of a Type Manager I was able to define exactly what a file manager should be and discovered that there are in fact many partial Type Managers out there now that implemented only half of what makes up a full Type Manager."
its a file attribute manager. not a type manager. adobe type manager is a type manager.
who the fuck gave this guy a license to make up new technical definitions on the fly ?
...should learn to Type Type Manager Less.
Move along, nothing to see here. This is nothing but shameless self-promotion from a guy who can't even spell "useful" correctly.
I hope I'm not the only one that had to wonder what iTunes and amaroK had to do with Adobe Type Manager and Suitcase.
I know a lot of people bitch that iTunes "does too much... I use LAME to rip mp3s I don't need my mp3 player to do it.. wah wah wah..." but iTunes was one of the things that made me switch to Mac. I was very impressed with how it did so much yet was very simple, especially the part where it keeps music files names and arranged according to tags.
before iTunes I used Musicmatch on Windows which I liked a lot because of its library management, though it started getting bloated towards the end (bloat doesn't mean adding lots of features, it means adding features at the cost of ease of use).
Adobe Type Manager Light
Next time, check prior art before appropriating a phrase and giving it whatever meaning you feel like.
Not to mention, "Type Manager" is a terrible name for "application that manages files of some type".
I've been using Adobe's version for years...
Be a real patriot: Question authority. Think for yourself. Formulate your own conclusions.
Email is a "type manager" and it has been around forever. Please don't add to fuel to stupid buzzwords...
there are many other Type Managers out there that are gaining market share and a rabid fan base of users such as digiKam and amaroK
...and then the KDE-ers complain KDE gets too little traction in the market compared to Gnome. Feh.
(especially KDE developers) For the love of God, it's not cute to insert arbitrary uppercase Ks into app names anymore. Yes, it's called KDE. Yes, there's that big K where the start button ought to be. You really love K. We get the idea. Now name your apps sanely instead of making them sound like they were named by 13-year olds trying to be cute.
<grumble>
</grumble>
Go somewhere random
A couple of pages of rambling is far from "news". This might be an interesting read for someone who has never thought of content or contextual organization before, but it's really old hat.
Now, if this goober had coded up a new manager which integrated all the functions he talked about, or had an extensble base manager to replace the native file system, with a defined api for plugins that would allow you to customize the environment, that would be news.
Is it just my observation, or are there way too many stupid people in the world?
So you put 'wikipedia "type manager"' into your address box too, yeah?
New must-have! metadata!
Coming soon! The macintosh.
Excuse the Unicode crap in my posts. That's an apostrophe, and slashdot is busted.
Could he come up with a more generic and confusion-prone buzzword than 'Type Manager'?!
- For the complete works of Shakespeare: cat
That's the first thing I thought of. I really wish he would explain what at Type Manager really is in the first sentence or two or at least give a few more examples of one. I didn't feel like RTFA because the blurb wasn't compelling enough.
--mike
A "type manager" is already a piece of software that manages fonts and such... shouldn't the category that software like iTunes, etc. (including traditional font-managing type managers) be called something like a "File Type Manager" to avoid confusion?
sig.
Yeah I totally agree. Apple tried that with an "i" instead and it did not work - or did it ? :D
I think a better name would be MIME Manager
What is wrong with being KDifferent?
/. is good for you.
.. but I really dislike all the "managers", picasa, nero, hell, I _stopped_ using ACDSee when it became to cluttered (in favor of irfanview ofcourse).
Frankly I just dont see the advantage of having one heavyloading utility for each aspect of your work. Explorer does it's work, if I wanted more power on my workstations I'd be slapping Linux on them where I have amazing powers at my tooltip with some help by perl and bash.
And for the shameless plugging of his own article I can only say: tsk tsk.
This is article is idiotic and totally misses the reason why these types of applications are a success. It's not about the type of data being managed it's about ease of which you can share that data with other people who have the same interests. It's about building a community of simiar interests.Microsoft Word is the "type manager" of doc files but I don't know that many people who sit around trading doc files and discussing the differences between how Word 6 rendered text versus Word 95.
The author should dig a little deeper...it's not about the data stupid.
People who bite the hand that feeds them usually lick the boot that kicks them
"Not to mention, "Type Manager" is a terrible name for "application that manages files of some type""
Object Manager.
My friend had basically created a Type Manager-like approach. I thought it was crazy because the engineering projects that we did used multiple files of multiple types. On his system the files of any given project were scattered across all these type-based of folders.
My point is that Type Managers can be very useful if a given activity only uses one application or type of file (e.g., rip/mix/burn/listen with music). But when the activity spans multiple types it drives the user back to using a general file manager. In such situations, existing Type Managers fragment the user's access to files and become a hinderance if the project's files are scattered across an email client, a photo manager, a sound file manager, etc.
Two wrongs don't make a right, but three lefts do.
He's got all the qualifications:
1. Can't spell
2. Horrible site layout
3. Rambling article trying to make himself look smarter than he is.
It's hard to tell if he's got the self-delusion and inattention to detail fully down, but we could be looking at a Malda clone here.
...another name for a keyboard designer.
Adobe Type Manager 3.0 Easter Egg:
Open Help/About, double right-click on it and will see the designer's photograph. FUN!!!
He who knows best knows how little he knows. - Thomas Jefferson
From the website :
Check out KimDaBa Demonstration Videos for details
I could swear I used Adobe Type Manager years and years ago. But hasn't the world progressed beyond that now?
I for one really hate iTunes for various reasons. I can manage my own mp3 collection in a sensible manner, and i don't want to have to navigate your braindead library. Call me old & grumpy again, but sheesh. Not to mention iTunes is an evil kludge gui-wise on both OSX and Windows.
Will wank off Linus Torvalds for fame.
Gou grefer ghe galternatives?
(Heh, reading that out makes me sound like scooby doo.)
"Proudly Posting Without Reading The Article"
The best (and only) way to deal with a mime is with a gun.
Be a real patriot: Question authority. Think for yourself. Formulate your own conclusions.
The biggest strength of ITunes is being able to clean up multiple MP3 file tags in one sweep. Highlight several MP3s, and you can set them all to the same Artist, year, Album, etc. This is quite helpful when I have numerous tunes by the same artist show up as several artists on my "artists" list. (ie bands with "and" in the title sometimes have &, and it's considered 2 different artists).
I also like the "play next in party shuffle" when you have a hankering for a certain tune.
on the downside, iTunes has created file name pollution on p2p networks with the "03 - Ronnie's song.mp3", ie mp3 files missing the artist name, but with a vague track number. I re-name them as soon as I get them.
What's needed is a class library which speeds the creation of Type Managers. It should have a Document base class which applications could extend to contain document info, and a View base class which would abstract the user interface. Both would have base methods for all the common stuff, and you'd extend them with the specifics of what you're trying to do. There'd be Views derived classes based on common widgets, like dialogs and lists.
Additionally, there'd be a way for software components to register as viewers of file types in some global database, so that they could integrate with the default shell and display previews. They should also be able to open the type manager or print, perhaps integrating into shell's context menus.
Yup, welcome to Windows 95 with a bunch of MFC applications, COM components and the registry.
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
Maybe I'm the only one who sees some irony in people using the Comic Book Guy style of response to mock and belittle an interesting work. The small minds living in Mom's basement can only denigrate a reasonably well organized treatise on an interesting subject. Commenters have pointed out that "there's nothing new here", "this guy chose a bad name", "this is only novel to someone who hasn't thought of this before."
So what, people? A refinement is a refinement. It's stepwise by nature. This is news because someone's aggregated their perceptions of the world and the ideas they sparked into one place. One of you complained about, "why didn't he publish an actual piece of code with an api for plugins?", and I suggest that maybe someone who reads this, who hadn't thought of all this before, might take this as a launching point and actually write something useful.
Let's enjoy the journey. If we happen to visit a few points along the road more than once, it's no big deal. Seeing the same vista from a different viewpoint can be refreshing.
Not A Sig
Why do I get the feeling that this is secretly a test to see how many times a summary can have a meaningless buzzword repeated before people start complaining?
I mod down anyone who says "I will be modded down for this", regardless of the rest of their comment
This was solved with a 2-byte code in the 70s.
Anyone got any news?
I have been taken a closer look at the Type Manager, what makes them so usefull, what they really provide for the user and came to some surprising results.
When it manages basic spelling and grammar, count me in.
To all the people out there complaining about this being the same as a font type manager, first off, relax. It's kind of scary that so many people are having problems by getting them confused, and apparently the problems are big enough that you feel the need to add a post about it. Secondly, next time you might want to take a look at the line: "from the no-not-for-fonts-silly dept." I mean... it even comes before the text, so there is really no need to get confused, and if somehow you still manage to get confused, please stop bragging about the fact.
It doesn't mean anything.
Like iYpe Manager...? WTF?
are full of use!
All you have to do is create a blog, ramble on for a couple pages about "K", never actually SAY anything except describing the letter and KDE's love for it. And then promote it on the front page of Slashdot.
Obviously that appears to be all too common these days.
And where are all these "partial type managers" (bad vocabulary choice aside) that the author claimed? Steam? Mame? That's ALL? How's about something useful - like documents, bookmarks, etc.
you posted in the article, you can't moderate it. RTFM, n00b. It's right here http://slashdot.org/faq/com-mod.shtml#cm1800. Geez. n00b.
Oh, wait, I just dissed a moderator didn't I? Braces for a few comments to get bitchslapped.
I have a file system / organization already -- not that it's perfect, or that I'm not constantly shifting it around in a confusing attempt to clarify my thought pattern, but I have an organizational system such that things are clasified the way I want them to be, roughly. And when something is not, I messed up, so can in theory trace the error. These "type managers" tend to stash files places where I don't know that they are (and rely on files being in those places), etc.
... whatever it is I want to manipulate. I might open a text file in any of a dozen programs, might open a photo with The GIMP or with a dedicated viewer program, or a slideshow program, or a browser.
It's annoying to be forced into someone else's paradigm -- what if I want to view my files in a different way than someone else's "type manager"? This is not to say smart people aren't making intelligent choices, for them and even for other people, but that doesn't mean any *given* person is going to like their choices.
I don't use Mac OS X much any more, and one reason is that (while again, I know a lot of people *do* like them) I don't like iTunes or iPhoto. When using a modern UNIXy desktop, I'd rather navigate to files (whether GUI or CLI) and then choose an app with which to manipulate
Tim
jrnl: http://tinyurl.com/c2l8yr / foes: http://tinyurl.com/ckjno5
My tea almost hit my monitor after reading the poster, the title, the contents, and the .sig. Thanks for the laugh. Although I feel somewhat dirty for it afterwards. But I guess that's the way cheap thrills go...
-- The Genesis project? What's that?
Do you have to use a silencer on the gun?!?
--- I wish I could hear the soundtrack to my life. That way I'd know when to duck.
I think I'll swipe it! ;-)
I bet 80% of the plugin API could be based around an XML ontology explaining to the manager what the 'types' and 'properties' of interest are. There would need to be some custom code for content display and editing but everything else can be pushed down to the filesystem/db.
Maybe it's because I'm just weird, but many of my files have implied metadata based on how I organise the filesystem they're in.
:(
MP3s are in directories of the form Artist - Album, file names are TrackNumber - Title. I've been doing this ever since an early version of iTunes for windows screwed my id3 tags, but since my MP3s are all tagged as a matter of course when I rip them, it means there's a level of redundancy in there. However, should something else wipe the metadata again, I still have the filesystem-level organisation to fall back on. I even have a tool which can strip this information out and refill the id3 tags with it, so it'd take me less time.
I'd be concerned that letting a manager program handle all of this might result in a hodge-podge of files outwith my control, then if something should happen to the organisational data, I'd have a pile of files with little, no or maybe even unintellgible organisation...
Game dev and music blog
Just kidding.
I agree....instead of criticizing the guy, maybe people could add some insight on to how to quickly and easily manage their own files besides using iTunes?
C:/DOS/RUN
http://www.actionfig.com/simpsons/cbg_dos.jpg
He who knows best knows how little he knows. - Thomas Jefferson
Now this stuff is comprised of: Photographs - .jpg's, .psd's, etc
Essays, film reviews, stories - .doc, .txt, e-mail messages
Day job stuff - all the above file types
Now, I think a previous post pointed out that in the real world, the activities in which we engage generally are grouped together as "projects" - a heap o' photos, some line art, copy, maybe movie files.
Finding by 'type' does very little to improve my organisational abilities and really doesn't help me find things easily and neatly. Good ol nested files and folders does the trick nicely.
But when I assign the proper metadata to my photos, iPhoto (looking forward to Apterture) is a killer way to visually organise and interact with (and bloody find) my photos.
Same with iTunes - great to batch toggle the metadata that goes along with the .mp3's - although I do remember being hacked off about SoundJam's demise as (if I remember rightly) it used the Finder's file and folder info to organise the .mp3's.
Seems to me that something like Apple's Spotlight - which is one of these Type Manager apps - is just another, very simple way to display data that really isn't all that useful in the real world.
How do you organise data and find things? If by project, do you use files/folders - shove everything into some sort of database - ?
We currently have applications like iTunes and MusicMatch that enable us to hide the details of where our files are in the file system, and present them in a means other than sorting by name/date/size. And that in the future we will have more. These applications no longer require the user to be intimate with the hierarchical structure of a file system;
... thanks ... really appreciate the heads up.
Wow
It's funny that these same apps usually use a hierarchical approach to display the data back. The two music ripping apps I use store the data under directories that have chosen because I have specific drives where I store 'data', and then stores them by in artists and album directories.
So all that is really done is provided a means to resort data by tags and edit those tags. I think the author should have gone the next step and suggested that maybe the standard file browser should provide for the ability to add plug ins to display certain directory and file types.
Now THAT would be revolutionary!!! Opps, forgot that Firefox (and other browsers and graphic tools and music players) already support those also.
Ok, evolutionary. Microsoft, are you listening???
I rarely read replies, it's my opinion and if you thought about your opinion a little more, I'm OK with that.
People might be more accepting if the idea had not been repeated a million times in the past twenty years.
The real problem is that to many, it looks nothing like refinement and instead rather like a reshuffling of old ideas.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
How's it offtopic when it refers to something brought up in the submission itself? Dumbasses.
Hi all,
No doubt Slashdot geeks will scoff at this article. Geeks want to discover for themselves the best way to do anything and everything on their computer. They shun having all related functionality served up to them on a silver platter as a coherent piece of software, which geeks been trained to distrust because slimy corporations usually make them (and contain slimy commercial intrerests over user interests).
The author is suggesting "hey geeks, why don't you be the ones to make the pan-functionally coherent software"? Then there will actually be alternatives (from a novice user, non-geek perspective) to, say, Windows Media Player, which does not expose your ripped files to the filesystem at all (a slimy corporate tactic)!
The author is suggesting that all the little tools laying around like "grep" or "awk" (that novice users will never learn) be combined into larger software that is easy to use by novice users. A few nice Open Source programs are pioneering this effort, like K3B, and the author is suggesting, "hey, now do that everywhere, for everything, and Open Source will win the day." Which I agree with.
Yes, it is far more fun to nitpick his choice of the term "Type Manager", but there is a big lesson here for geeks, who often times have a hard time understanding what novice users want. Novice users (ie. the other 98% of computer users who are not geeks) want software that beautifully presents them all the best choices in a coherent application for a given activity. Open Source Geeks have the opportunity to do this and win, by doing it and leaving out the corporate slime that nobody wants.
I use the "Angel" from Arcamax to organize my information. :D
Sorry, but to me, a program that will scan my HDD's for pirated MP3's and DivX's, then lists them on my screen to "better manage them", means "U GONNA GET RAPED IN JAIL"...
And have people ever heard of FOLDERS?
"Oooh, lookit this, it's fresh, it's hot, it's cool, it's iTunes/WMP/amaroK/latest disc-scanning PoS!"
"Oh, I have a folder called MP3. And I can even arrange them in SUBFOLDERS!"
"Whoa... That looks... Complicated. How do you create a folder, anyway? This sucks."
*sigh*
Wether we're browsing HTML webpages, PDF documents, etc.. the needs are the same and IMHO, the browser for readonly should appear mostly identical. ...
To browse you need to:
- navigate the document (back, forward), activate link to open a new document
- zoom in, out.
- register specific location.
Unfortunately some of these feature are not universal: you cannot bookmark a location in a PDF usually, which is annoying..
And those feature are usually incoherent: you do not use the same method to zoom in/out for a PDF or a for a webpage, too bad.
The OS should have registered actions for transports and datatypes. So each "scheme" (protocol) in a URL, like "http:", "ftp:", "mailto:", "rtsp:" (omitting the ":", to be exact) has a registered app or process for transporting in that protocol. And each datatype in a URL, like ".html", ".mp3", ".xml" (omitting the "."), or its overriding Content-Type header from the server in protocols like HTTP has a registered app or process for rendering the transported data according to its MIME type. Disambiguation among registered alternates is done in the application receiving the URL request, with a default. So a desktop context menu can offer a prioritized "Open with...". Apps can handle URLs for which they have transport and/or rendering facilities, or hand off to whichever app is registered. The only complexity is that the renderer might differ whether the data is to be "read", "edited" or "executed". The app ought to be able to differentiate the mode from the context in which the URL is requested, but the OS would have to register those modes. The key is that the facility resides in the OS (or its execution environment) so every app always has it available - it's IPC.
This approach is already part of the FreeDesktop.org model. GNOME has already implemented it, with some bugs, while KDE has committed to implementing it. The key to the implementations is interop among all of them according to the same rules. Which was the key to the success of URLs themselves, which knitted together the Internet into the critical mass that made it the environment we know today. Then we can build whatever apps we want on that simple facility. Getting rid of the "file system" (replaced with RDBMS and views, for example) will be just one early victory built on that strategy.
--
make install -not war
Ok, I don't really care about the article...
.mp3 collection management tool favorites. Anyone got some that they wanna tell me about?
But I was hoping to see people talk about
Here, here. Mod that post up. The fact that we're still using the folder metaphor on computers is insane. Having to organize your stuff by putting each thing in only one container is a limitation of the real world. It's like a word processor that makes you paint your screen with Wite-Out to undo a typing error.
Insert witty sig here.
The only difference between a file path and, say, keywords is that the former is thinking in terms of the computer (sort of like C), whereas the latter is thinking in terms of the data (sort of like Java).
That's wrong. Paths are not just metadata, they have specific semantics associated with them that, say, tags don't. Furthermore, paths have semantics that users grasp easily and that they rely on.
Now, people have been attempting tag-based, non-hierarchical, database-based and other file management and navigation strategies since the 1960's. UNIX itself used to be graph-based, not path-based. All such attempts have been failures. Paths seem to combine power, usability, and correct semantics in a way that no other system has managed to do to date. There are specific applications (like MP3 jukeboxes) where other approaches are better, but for organizing all the stuff on a computer as a whole, sooner or later, you end up with paths and path semantics again.
well I do have a lot to manage!
pr0n: now ive got your attention click here
This is the whole problem with computers and the internet....
more words does not equal more productivity!
What is this? Advertising day?
Yes, and many of us have commented that it may *NOT* be such a good idea to use this ``everywhere, for everything" as you suggest.
These interfaces are only useful if you want to work the way the designer wants you to. If you don't, then these interfaces get in the way of what you might want to do.
It also does very little to help the author's case that he is sufficiently ignorant of the field that he has chosen a name that (i) is taken (ii) by a very different sort of software. Even if the author has a clever idea (and granted, I don't think he does), he should have done just a bit of research to not sound like an over-enthusiastic n00b.
"Every decent man is ashamed of the government he lives under." - H.L. Mencken
Type Managers can include basic manipulation, but should not be a full blown editor.
A type manager is a browser interface optimised for a specific type of object.
There's no reason it can't conceptually allow arbitrarily complex operations on those objects, just because the basic interface is a browser. That's just a silly restriction.
The article probably misses the point by looking at files of omogeneous type, but the problem with current... erm... paradigm (sorry for the swearword, couldn't think of a better one) for managing files is real. The problem lies not in the manager, but in the storage mechanism itself: the hierarchical filesystem. We all hit the point, sooner or later, where the organization of data in our mind can't be matched with a rigid directory hierarchy. What is really needed is a graph organization. Links and metadata help up to a point, but in order to make things work smoothly it will probably be necessary to abandon current ideas about directories as "containers", slash-separated file paths, filenames as unique identifiers (and attributes as mere extensions of the filenames), probably even the graphical interface used to access and manipulate the filesystem. The listed applications do their best, but they still work at a level too high to replace a file manager. I don't have a solution. Maybe Reiser4 can be of some help.
Oh, I guess this will be modded down because I said "paradigm", anyway...
Nuffsaid
________
Don't know about his cat, but Schroedinger is definitely dead.
Thankfully, the best part of it (Live queries) are being reborn in OSX's Sherlock.
It's like iTunes for all your files.
I think everyone who has filed stuff in a hierarchy has lost a file at least once.
The problem with a hierarchy is that only one "attribute" can be assigned to the file: that is the file path. Any other attributes the file may have are within the file, not the file system, and result in a click on the "find files" button which iteratively reads all files in the selected path looking for matches.
If on the other hand, at file-save- / -creation-time, multiple tags could be associated with it, and this stored in a (relational) database, then finding like-files would be a database search...much quicker.
So the question remains: Will it ever really work? not sure. The key reason it works for MP3s is the existing database of songs + tags, with legions of people updating it with new data ... no such database can exist for custom documents I create, or my organisation creates. These tags must be created either by the person authoring the file, or by me when I receive the file, and this is much more time-consuming than clicking Save-As and dumping it a folder...
If you think imaginary property and real property are the same, when does your house become public domain?
The author's "Type Manager" is nothing more than a manager utilizing more metadata than normally.
Your classic file manager *IS* a type manager because the file name is a metadatum and the parent directory is a metadatum: neither are direct data (such as what I'm typing now). So organizing, say, a code base on a directory hierarchy that may include module names or library names or file types (docs go here, man files there, source files over there, etc.) *IS* feeding metadata to your filesystem to organize your files.
The "Type Manager" has existed from Day 1 when files were given names. (Punch cards are before my time but I suspect the punch cards that represented a program were stored together and each program was stored separately. At this point, *you* are the metadata organizer.) Since then, it has only progressed from a flat file system (the likes of Apple IIc) to a one-level deep filesystem to a multi-level filesystem (no linking) to a multi-level graph filesystem (includes linking). Now apps are taking it to the next step by merely using more metadata. That's it, nothing new.
In the end, the bits that represent your actual data is a long string of bits (losely stated) and your filesystem is just a type manager organizing your bits by file names and parent directories. bash, Windows Explorer, Finder, etc. are all just wrapping your metadata organizer (your fs) and some (previously and now) are using file-specific metadata for further organization.
Big whoop.
From the article:
Type Manager applications are not new, in fact you probably have been using one since you got an internet connection.
It appears the author doesn't even fully understand the concept of metadata (*ahem* "Type") and it's usage has long existed before your email client and long before your internet.
Seriously, nothing to see here! In fact, I want my time back for reading it...
:wq
After reading the article, I still don't see what he wants that Windows Explorer,
Nautilus, etc. don't already do. You can already search volumes, see thumbnails of images, move and rename files, and so on in any of these. With file association, you can double click on the icon or file name and bring up an editor or viewer. So what do these new "type managers" add?
"Type manager"?!? /., this is stoopidt even for you.
C'mon
Hitler used to manage types... of people! D'oh! Godwin!
The problem is that the article makes a distinction beween a file and type manager and doesn't actually draw a line between the two.
Type Managers need to remove the user from the details of the file system. They need to present the files in a hierarchy that best suits the files
"A hierarchy that best suits the files"... sounds like a file system to me. What "details of the file system" is the user removed from again?
Type Managers must provide the means for viewing and editing that file meta information
File managers do the same thing. Right click on a music file, click properties and then click Summary. The functionality varies a bit and is broken, but it's there.
If on the other hand, at file-save- / -creation-time, multiple tags could be associated with it, and this stored in a (relational) database, then finding like-files would be a database search...much quicker.
How is this incompatible with having file paths? There are similar systems that do exactly this: beagle, spotlight, Copernic Desktop Search.
IME these full-text tagging and searching schemes are less useful than is claimed. The primary problem is that it frequently takes multiple searches to find a combination of search keywords that deliver a reasonable number of hits including the file I want.
(And why a relational database?)
These tags must be created either by the person authoring the file, or by me when I receive the file, and this is much more time-consuming than clicking Save-As and dumping it a folder...
I have a project which consists of some mp3s, PowerPoint files, Word files and some psds. How exactly would 'type managers' sort out that problem? I sure as hell don't want to sort through files using various differnt programs.
What I think he really means is that he hasn't thought the problem through fully. File Explorer in Windows does most of what it needs to. For a better look at attribute or tag-enabled filing, have a look at what Windows Vista is trying to do.
Overall, nice try, but seems to be lacking in knowledge of what's already out there and what's being done.
...don't want a type manager. Multiple references to the same file (by type, by project, by status (draft/final), by age)? Yes. If you can do it well and maintain integrity so I don't end up with a number of dead shortcuts/symlinks, that is something I'd like. It's all in the implementation though...
Live today, because you never know what tomorrow brings
Use MP3Tag for building id3 tags. Free, fast, and easy to use.
http://www.mp3tag.de/en/
Can do mass changes very easily, file renaming based on tags, reads almost any audio format, freedb support, and lots more.
Jfilerunner
I've been using it for a few months, I like it.
No, I'm a graphic designer; my google box was fed by the terms "Adobe Type Manager" and "Suitcase manager". I'd used Adobe Type Manager pretty frequently before Windows 2000 introduced integrated ATM font support. But thanks for asking.
OS/2's folders were highly configurable and extendable far beyond anything you see in Windows or even the Mac OS X 10.4 finder.
One extreme example of exactly what this article is talking about was RexxMail. From what I understand, instead of having a mail program with a dedicated custom interface, the developers of RexxMail simply extended the standard folder to list files of type email so that you can see the To: From: Subject and so-on in the view. When double clicking the file, it would open it in an appropriate editor and provide different options. This way you could use all the power of the Operating System's file system and folders to manage your email without having to learn some completely different interface that insisted that your email go in some specific place. Very cool.
Signatures are a waste of bandwi (buffering...)
That's the beginning of a "type manager" interface. There's presumably some way to register your own application with File Manager so that it is automatically invoked when the right type of file comes along, but I'm not currently doing Win32 programming, so I don't know which API to look at.
- Location has been used for decades (flat filesystems are not that common today)
- Content type is used to associate actions but BeOS live queries shows the way to spread its use. Nautilus specific views is another form
- Properties is the ability to query based on some properties included in the data (Exif, MP3 tags,
...) Already available but only for certain data. There is not yet a general way to deal with properties.
- Tags have prove themselves usefull for links (del.icio.us) but could be used for any file
- Computed values are special values that are maintained from the data (thumbnails,
...). They are similar to properties but are computed instead
- Full indexing (Google desktop, Beagle,
...). Also include the transformation of data to text (Google images)
- Recommandations is the way to ask other people their opinion about a specific chunk of data
There is no single way. All of them seem interesting and I try to implement them in JDistro. However, support from the filestystems (inotify and al.) or from the databases (triggers) are required to make them viable. In summary: I don't think Type Managers are the future, Data Managers are.Million Dollar Screenshot
I like how the formal definition of a type manager calls for specific features like a "vew last import" kind of playlist/album. That's not an inherent thing in a "type manager"--it's a feature. This article is just a piss poor attempt at saying "I like iTunes"
Programs that deal with groups of similar types of data? That's fucking inventive that is. This is nothing new, programs like this already exist, and have done for a long while. And with only a few exceptions, they've all managed to piss me off having crappy limited interfaces. I much prefer to use a standard file manager to organise my data, grouping them by folders lets me get to the right things just as quickly, and without having to start a number of different applications for each type of data I'm looking for.
My Documents, My Pictures, My Music, My ebooks, My oh my, make it stop!
Haven't we already seen this many times before?
Comment removed based on user account deletion
when you have to "go through a dozen different smaller tasks before I can achieve my goal" you should *really* think about learning what pipes are .
I'm afraid this is entirely typical of so many recent articles referenced to on slashdot; people writing about a *nix based systems obviously dont know *nix
When the seagulls follow the trawler, it's because they think sardines will be thrown in to the sea
I think he was born in 1991 so to him, having an internet connect is the start of "computers".
A file having a path is not incompatible with attributes, but perhaps becomes redundant, or more specifically, the file-path can be an attribute.
While it may have taken you multiple searches and time refining searches to find specific documents, I can only assume that if you were forced to search for it, you didn't already know a hard-path to it - Once you know that Ford's cars can be found at www.ford.com, you don't search for it again. "www.ford.com" can be viewed as merely an attribute that you give your browser, but instead of resulting in multiple hits, only a single 'file' is returned. In a database, this is commonly called the primary, or unique key...
Why a relational database? why to relate the data of course! :)
I guess my point was only that just because people use the current hierarchy of folders to store their files doesn't mean it is in fact the best possible way to organise and retrieve them. Folders and Files are a virtual implementation of a physical operation: Filing cabinets filled with documents.... since when does a physical-world operation translate directly to the best virtual-world implementation?
If you think imaginary property and real property are the same, when does your house become public domain?
One of the classic problems with file systems is "where do I put stuff?" In an even more ironic twist is that the desktop represents some part of the file system which doesn't correlate directly to the file system (the desktop is the first thing you see and yet is not "the first file system"). Ultimately one of the more frustrating support issues is finding lost stuff. People may claim "it is the user's fault for not having better organizational skills" but wait...isn't the computer there to assist with this in the first place?
A simple use case is "I want to save 'file type a'". A simple question relatating to this is "where should this be?" for which any number of people will have any number of answers. And the OS isn't going to help because its answer is "anywhere you have permission". One user will place it in a directory "My FileTypeAs". Another will place it on the desktop (where ever that ends up being). Another will bury it 4 deep in a structure they only understand. So on and so forth. It creates a situation where there are unwritten (and unenforcable) rules on what to place where. If the rules change, then all of the old data runs the risk of being misplaced or lost while it is refiled.
Enter the idea of 'type managers'. This is a view on files of a given type that is customized to everything in for that data type. It knows how to "view" the data type. It knows how to store that data type. It knows how to search on that data type. Currently in the "explorer style" file manager, you have one frame that is trying to interpt all the possible types of files on the file system trying to deal with any number of metadata types which it might not be able process. In the type manager system, if you want to find images one would use the Image Manager and search much like Google/GMail. No more searching the entire file system in a haphazard manner or other mnemonics to try organize in order to leave a trail of bread crumbs back to the files. MP3s don't exist in the world of the Image Manager so it never has to deal with them. If you used the Image Manager to save the file then you know it is in the system and it will be located.
Most of us already use one popular type manager: your favorite Email program. To read or write email you aren't bothered with most storage details (imagine if you were forced to "save to the send box"). In fact any time you pull it out of the Email program to the file system or something else (like a printer), you create more work for yourself. Your email program is more than able to track and handle Email than any "explorer" can. It knows about Junk Mail, friends, threads, and all of the other Email centric paterns that an "explorer" just couldn't handle. And specifically GMail gets it right by avoiding the "folder file system" idea.
The best system would be to have both a "file manager" and many numbers of "type managers". The best UI systems allow for power users to play with toys without crippling novice users. Power Users would rather use the file manager and track things themselves but the cost of this shouldn't be that novice users end up losing things and being frustrated.
PalmOS has 100s of files in one dir (in fact, you can't create directories but rather apply categories to files), and it's hell to manage that. But if you start some app (e.g. text processing, image viewing etc.), the app itself shows only the files it supports and usually offers some kind of classification for easy navigation. It's great if everything works as intended, but usually apps leave their files behind after being erased or create temp files and don't clean up.
The concept is good when you work with simple stuff, but it fails when you need to do hacking.
In my opinion, the best choice would be separate apps for music/images (because these are special) and a desktop search app like Spotlight or Google Desktop for everything else. And certainly a traditional filemanager. For example, I still use mc even in KDE (although I like Konqueror more) simply because it works better at hacking stuff. But it fails in simple tasks like sorting documents in folders.
My personal assistant manages all my typing thanks.
useless sig advice - Read Nabokov.
With multidimensional views on the intersections between tags.
Anyone know of similar?
Deleted
The 'K's and 'Gn's and 'X's in desktop application names are "Consonant Metabranding".
That's at least as good a meme as "Type Manager".
I guess I'm seeing a problem that has not been adequately resolved - people do lose files on a semi-regular basis. A problem that has not been perfectly resolved is a prime place to look for and implement some new innovation... i like to flex the creative/problem-solving part of my brain once in a while, to keep it from getting too rusty.
:)
Certainly. My suggestion is that giving up on heirarchal organization schemes is likely to fix some problems and create other problems. The predicted demise of the file heirachy is greatly exaggerated because heirarchies offer quite a bit of utility, and work well with the ways in which human beings tend to think about the world. I also don't think that seaching entirely solves the problem of "I lost that file."
Why a relational database? why to relate the data of course!
Which doesn't answer the question. "Relational" in "relational database" indirectly refers to data. More precisely it refers to flat tables linked by relations to other flat tables. Why the assmption that the best type of database would be built around that kind of structure? (Of course you can convert any kind of database structure into a normalized relational database, but not without some sacrifices along the way.)
I guess my point was only that just because people use the current hierarchy of folders to store their files doesn't mean it is in fact the best possible way to organise and retrieve them. Folders and Files are a virtual implementation of a physical operation: Filing cabinets filled with documents.... since when does a physical-world operation translate directly to the best virtual-world implementation?
Well, to me this is begging the question. Folders and Files are a virtual implementation of a physical operation. But filing cabinets filled with documents are a physical implementation of a cognitive operation: sorting items into types and subtypes. Other examples of this include Dante's Inferno (sorting types of sin into categories), the Dewey Decimal System (which combines hierachies with indexes), and the Linnaean Taxonomy (sorting animals and plants into multiple levels of classification.) The indexed database of "attributes" or "keywords" is also a virtual implementation of a cognitive process that was performed as a physical operation. (For examples, card catalogs, book indexes, and a thesaurus.)
So I would argue that Folders and Files are a good implementation because they are strongly analogous to the ways in which human beings think about their world. At any rate, a "Folder" is simply a collection of objects that can be manipulated (viewed, copied, transferred, archived, deleted, etc.,) as a group. That folder could be created as a heirarchal list of attributes (such as $HOME/Projects/Mushrooms/Yellowstone/) or generated as a search through an index. The point is that file paths are just one way of assigning attributes to files.
It was just a joke in the Ubuntu thread when I said we needed a Web 2.0 tagging type manager! Jeeze luise.. ;-]
With the amount of files we are dealing with these days we need a new way to find them. Apple is getting close but it still needs work. Windows has nowhere to go but up from their terrible search engine in XP.
I wish Apples Hot Sauce project would have panned out into something useful. Their 3d based browsers with relationships was quite cool and with todays processors and graphic cards would have been a nice alternative view...more so with spotlight metadata.
More obviously, how about simply making files smarter rather than gluing ever more solutions on to guessingly tease out ever more intelligence from our files?
Everything is not a file, everything is an object. A file is simply a collcetion of data. An object has associate things you can do with it.
Or hopefully never will have to :)
He who knows best knows how little he knows. - Thomas Jefferson
The concept of a "Type Manager" is essentially what Microsoft is going for with the Vista + WinFS combination.
WinFS will allow for developers to store easily accessible meta-data about any kind of file they want. They can publish these schemas (all XML), and allow other developers to utilize this meta-data in their own applications.
Vista's explorer allows developers to extend the display of a file or folder to include this meta-data, as well as implement their own UI components for displaying non-textual meta-data, such as album art or video icons or whatever.
The key thing that Microsoft is doing that others don't seem to be (perhaps because they aren't in a position to mandate it) is that the extensibility should not come at the cost of UI consistantcy. Yes, it's awesome to be able to open a folder full of music on your computer and have that inteface be taylored to dealing with music, but it shouldn't be so different that common tasks that apply to all file types aren't possbile or are convoluted.
While these videos are somewhat outdated, they show the direction Microsoft is trying to take things and it's quite innovative:
Vista Concept Videos
In particular, the Higher Education video shows the potential of WinFS.
Yes, it's so much easier to knock down other people's ideas than to come up with your own.
Sure the article presents ideas that aren't new or ground-breaking. Yes, the term Type Manager is aweful [it's more of a Domain Expert or Task Facilitator]. Still, the article is thought-provoking. What it is really about is answering questions like "why is iTunes so appealing?" In fact, it is really talking about is Application Patterns (like Design Patterns). Could be an interesting book subject...
Wanted: witty unique signature. Must be willing to relocate.
(In mid 90's) back then there weren't many files to manage either.
Huh?
No application was complete until it integrated e-mail somehow.
HUH?
Does Slashdot hand out "This Year's Worst 10 Articles Award?" If it does, I would like to nominate this one.
There are two types of people in this world...
One type is like my mom (or my manager) who no matter how organized they can be with the closet and a European vacation itinerary, don't have the first clue about organizing stuff on the computer. Every email they've ever received is sitting in their inbox; every file they've ever downloaded sits in their "My Documents" folder; they have two hundred icons on the destkop; and they buy new memory cards for their camera when their old card fills up. These people HATE the file manager. For these people, the type manager is the solution to all their problems.
Then there's the other type of person, like me. We have a clear grasp of "file" and "directory" concepts. We are able to create subdirectories, and thus organizational hierarchies. We don't need a type manager to find our songs, because we know that "Money" is located at "share/music/ogg/Pink Floyd/The Dark Side of the Moon". There's no need for us to type phrases into a googlesque searchbar to get back 43,000 results which we need to narrow down by further typing in the searchbar. We don't hate the type manager, we just find it annoying and useless. For us, the file manager works.
Unfortunately, the first type of people whine the loudest, so they're most likely to get their whims catered to. It doesn't matter that the Unix desktop is wholly unsuitable for their needs, we're going to give them preference over Unix using type two people anyway.
The correct solution is NOT to dump one type of manager for another. Instead it is to understand that the file manager and type manager solve two DIFFERENT kinds of problems, that we need both. And by "both" I don't mean one is the default and the other an undocumented configuration setting hidden somewhere in the depths of gconf. There's no reason both functions can't be incorporated into the very same manager.
A Government Is a Body of People, Usually Notably Ungoverned
This is true, however, the way that statement is worded gives a hint to part of the problem. Hierarchies (plural) do offer quite a bit of ulitity. The problem is that, at the moment, in most systems a given file can exist in only one hierarchy.
Shortcuts exist, but are hard to use, as you have to find the file and drag links to it elsewhere.
Desktop search is great, but often needs metadata as all too often we lack context. And how many people take the extra step to apply metadata and/or keywords to documents?
Any sect, cult, or religion will legislate its creed into law if it acquires the political power to do so.
The OS should have registered actions for transports and datatypes. So each "scheme" (protocol) in a URL, like "http:", "ftp:", "mailto:", "rtsp:" (omitting the ":", to be exact) has a registered app or process for transporting in that protocol.
That sounds suspicously like the ages-old KDE KIO system (introduced back in KDE 2.0).
So a desktop context menu can offer a prioritized "Open with...". Apps can handle URLs for which they have transport and/or rendering facilities, or hand off to whichever app is registered. The only complexity is that the renderer might differ whether the data is to be "read", "edited" or "executed". The app ought to be able to differentiate the mode from the context in which the URL is requested, but the OS would have to register those modes. The key is that the facility resides in the OS (or its execution environment) so every app always has it available - it's IPC.
You are aware that you are aware that you are basically describing what KDE has been doing for ages?
KDE has committed to implementing
KDE has had the functionality for a long time. I don't know about implementing the standard, but the functionality has been there since at least the 3.0 days.
Sorry, but anything like Itunes, the 10 different camera file managers and all that freak me out completely..
.mov on a camera, mp3 or wma on my mp3 player, and just don't want to deal with 20 different producst, that are on top of all bad are mostly windows only.
/images or /videos /audio /mp3, and maybe /gallery or /playlists ...
.... I listen to 2-3 genres of music so the ipod menus are completely whacko for me.....
.....
.... could not they just give me one that works? Or use one that works with every single other app on my system ?
I prefer a manager with 2 tabs, and shortcuts instead dragging stuff with a mouse...
I know there is jpg or tiff or
I want to see a filesystem with :
I even considered getting rid of my ipod, because iTunes gives me the creeps, and would prefer just seeing a filesystem, copying files and then "find songs" or open playlist or similar on the player
still in the search of something usable, especially because that crap yesterday forced me to upgrade, then had to wait 35 minutes until it reexported circa 9gigs of "library" over various network drives
On the other hand half of the managers force a bunch of dependencies on you, e.g. my nokia manager only runs on a certain bluetooth stack, and as opposed to the one I had it costs like $50 bucks
grrr
It's not incompatible, it's an extra complication. Perhaps it will be useful when someone figures out how to know what tags you're looking for. So far we can't automagically generate them and people aren't willing to code them in (and we haven't figured out a core set that specifies all the files on my computer better than paths).
Why relational? Because relational databases don't give priority to any particular kind of access discriminator and are flexible enough to represent any organized data structure. The first databases were hierarchical, like file paths. If we implement alternative hierarchies, we're looking at the second evolution in database managers. But they have the same problem they were trying to solve - each access pattern has to be designed-in or it's dead slow. Relational databases were invented to give good (not best) performance for any access pattern, because we found out you can't predict how you'll want to access your data tomorrow. So let's skip that wheel and move on.
Oh, and object databases are just hierarchical databases with variant records -- it's been done before by those old mainframers you replaced. The relational model can represent any organized system of data that's been thrown at it as efficiently as or more efficiently than any other model. No-one has found a counterexample that I'm aware of. So why bother with something different? (if you have a good, provable answer, you can probably cash it in for a PhD)
How did such a stupid "article" (actually just a blog) make it on to \.?
He is late to the game. The train left the station. Where was he when the term Digitial Asset Management was coined?
HELLO!!!
You're missing the point a bit, it's easier to locate files if you can search by type sometime, rather than navigating through a heirarchy of folders. For example, if you want to see all the video files in your home + subdirectories, you could click on a folder called "My Videos" or something and all the videos would show up in that folder, even if they aren't really located there. There are advantages.
Suppose that I replace the fine filesystem you usually use with one single lookup table, and a relational database that maps the old file path to the entry in the lookup table. Now, the file path is just metadata.
Yes, and it doesn't behave like a pathname anymore.
From the user's perspective, nothing has changed in how they use the system at all, but that path is now very clearly metadata.
From the user's perspective, a lot has changed: the path isn't associated with location on a device, their standard GUIs don't work on it anymore. In order to make all of that work, you need a lot of additional code. That's why pathnames are not just metadata.
Oh come on - You obviously don't have enough vision here. In fact I think this should apply to UFS as well. Let's do away with the idea of folders, that's soooo old. Instead everything will have types. For example all the users files will be under the type 'home' then they'll also have a type for which user they belong too. Then the users can put more types on them as they please. System configuration files will all be of the type 'etc'. And system files, programs, and such will be under the type 'usr'. We'll have a 'tmp' type for temporary files and a 'var' for variable files. You might also want an 'opt' or 'pub' type. Also we can make types that start with a . special and normally hidden. Then we can make our next generation file managers. You can have then 'browse by type'. For example by default it will show you what's under the type 'home' and $USER (whatever that is). Under there it will show you a list of types that are under the type you are viewing. If you click there you will browse under that type as well.
Nonsense! Everyone knows the best way to kill a mime is with a rocket launcher.
"Destroy science and religion. Science would re-emerge exactly the same; but not religion." - Penn Jillette, paraphrased
locate and grep are my type managers. I name my files very long filenames (thank you tab completion) based on keywords.
Then, I just run locate and grep out or in the stuff I need. It gives me command line power similar to OS X's awesome Spotlight.
Don't think that a small group of dedicated individuals can't change the world. It's the only thing that ever has.
I just figured out where all this, "metadata is easier than directories", comes from.
My mother in law wanted to burn some images to a CD, but could not find the images trough the Nero file manager. The problem was that she has one old "My Documents" (MD) directory in C:\ and one in the default place. Now when she wants to get to her documents should she click MD on the desktop, My Computer, C:\ or C:\Documents and Settings\...\MD ? I think this is the problem, Windows does not tel the user where the files really are but try to "make it easy". The Unix way where the user must know the real path to the files does not give this problem. If you have shortcuts to the directory or file, then let the user know!
J. River Media Center has an iTunes-like interface, buand has support for not only audio, but images, video, and documents as well. You can create custom view schemes, and custom database fields. http://www.jrmediacenter.com/features.html
"I'm not a cool person in real life, but I play one on the Internet". Galley
If on the other hand, at file-save- / -creation-time, multiple tags could be associated with it
You can: office documents, images, sound, video, and other major file formats give you that option already.
and this stored in a (relational) database, then finding like-files would be a database search...much quicker.
Quicker than what? The problem isn't that metadata is slow to search through, the problem is that people don't bother to add it.
Eh? I have a file whose path is:
(Forgive the slashcode-generated space in the path)Lots of "attributes" there: the fact that the file is "media", the fact that it's music, the name of the artist, album, and track. Also the track number...
Also, the same file might appear under another name (a symlink). If Primus had a greatest hits album, I might have:
...as a symlink to the original file. So that's two sets of "metadata" stored in paths.BTW, sorry about posting to such an old article, I forgot to hit submit before leaving work :)