Overall agree with the article: use individual apps that do well what they're designed to do, and use the file system to organize your stuff. Unfortunately, a tremendous number of users cannot make heads or tales out of the file system.
I believe the chief causes of the confusion are (1) different apps save newly created/edited files to different places (2) many users are clueless about navigating the file system within the Save As dialog (3) many users rely on an app's Recent list (4) many users don't know how to search their own file systems.
(1) Some apps always want to Save As to the Documents folder. Some always want to Save As to wherever you finished off your previous Save As. Some always want to Save As to the folder of the file you most recently opened in that app. Etc. No one can remember each app's convention.
(2) People don't realize they can navigate within Save As to someplace other than wherever the app feels like plopping the file by default at that moment. True even when shown over and over and over how to do this.
(3) Once you move a file, its entry in an app's Recent listing is probably borked. This sends people who rely on Recent for finding and opening files into a panic.
(4) They don't know how to click on the desktop and hit cmd-F or to click on Start->Find. And once they do, the Find window looks nothing like Google or a web browser. They don't understand what the various search criteria are or how to enter them. They don't know how to read the results.
The file system is a lie. It's all a blob of memory and your data is like trash, thrown about all over the place. How it appears in your little Windows Explorer has little to do with how it's stored, so why can't we have more flexible systems for structuring our data?
As soon as you toss out the idea of separate folders, then your points become moot and are solved (except for maybe point 4, though I think a new OS would want to heavily promote that).
(5) Different apps want to open files in different places. One app's Open dialog wants to look in the directory you finished off your last Open. Another app wants Open to look in the directory of the file you most recently opened by double clicking on its icon. A third app wants Open to look in the directory where you completed your most recent Save. Etc. As in the Save As problem, no one can remember each app's Open convention.
I know several older people (in their 70s) who sorely miss the days of yore when they worked with their user files on floppies. Each project lived on one or two floppies. Even if the OS allowed subdirectories, they didn't bother creating them. So, on DOS or early Windows, say, the users could always find their work at A:\ (unless there was no hard drive; then their work was at B:\). No getting lost in Open or Save As, no Recent nonsense, no digging down through folder icons, and hardly a need for built-in search.
We've been living with hierarchical file systems for several years now. And the upshot is, the Finder (or whatever it's called) and desktop apps plus newfangled browser apps don't present a consistent interface to the file system.
I don't use any of the programs he mentioned, but I do use VoodooPad; it has an "Everything Bucket" feel to it, but I characterize it more as a personal wiki.
The strawman argument here is that if you just dump all your crap somewhere you won't be able to make much focused use of it later. Well, no shit. But let me trot out my own scarecrow:
Every time you add a program your repertoire you're replacing productive time with administrative time. Of course a great program is worth the extra time you spend alt-tabbing to it, filling it with data, and then exporting and munging it to work with the other 100 programs you have.
You know what? After 20 years of being a hard-core computer geek, I've ultimately found that huge percentage of electronic tools end up being less useful than simple pen and paper. A personal wiki bridges that gap and allow me to keep todo lists, code snippets, quotes, ideas, and whatever else I need to be easily accessible in a single easy-to-browse window. I can put any one-off content in there I want, and re-organize and hyperlink at will. Alex's suggestion that a directory structure and a bunch of text files is more useful is laughable at best. His advice is ill-informed and unimaginative. I'd like to see what he does with his random bits of information that he need regularly.
I use a wiki as a personal wiki, personally. I've become much more organized since I started using it - things that were previously on scraps of paper and miscellaneous text files are now tagged, cross-linked and searchable.
I'm not sure what else you would do.
A few types of structured data, like addresses/other personal info, does go in a specific program - but mostly because I need to export/input that data between applications. Generally phone numbers get input in my cell, exported/integrated to Outlook, and then exported to Skype (and to my wiki).
These tools are useless because they struggle to play well with all the other tools on your computer. There is no common interface that allows them to share functions or data easily. We're working around this with various hacks but it doesn't solve the underlying problem which is your modern operating system. If the OS allowed easy sharing of code and every program was forced to share its functions/classes, then it would be trivial to move data around between applications.
He's right in that these apps are a horrible filesystem replacement. That's why I cringe when I get emails from users trying to use it as one. I think there's a lot developers in this category could do to position and market their apps better.
But what's really happening here is that people are grasping for a better Finder. Using one of these apps, for some people, makes managing certain types of data more concrete and understandable. But using it instead of a folder, for heaven's sake, is folly.
When will programmers stop blaming users for the deficiencies of software?
Many users do not understand hierarchical filesystems. Few users care that "Computers work best with structured data." Why in the world would one argue against making a computer do work by searching for information?
Use Google to search for your home address. Try a UPS tracking number. Enter the first few characters of a word and wait a moment. Enter your zipcode followed by the word weather. These are the types of solutions that we should be working on.
Nice straw man. I advocate better software, not more software.
find, locate, dir /s and the horrendous file search in Windows XP (I haven't used Vista) are not solutions for most users. Under what circumstances are Spotlight, Google Desktop Search and Windows Desktop Search context aware? If I'm in MS Word will any of these tools show Word documents before other types of documents? If I'm using a browser will they perform a web search?
It wasn't meant as a straw man. Better software is only going to happen in addition to software that exists already - i.e. more software, leading to (I suggest) more complexity.
Filesystems work. At least, Windows simple filesystems work, Linux's mess is less pleasant. People who want to can learn to put files in My Documents and find them later - it's not massively complex, there's a place for shared documents, a place for your documents, a place for Programs and a place for Windows. That's about all you need to care about.
At some point I object to the "Why should people learn anything? It's all the failings of the computer!!!!" attitude of your comment - there's always a trade off - any alternative will come with alternative problems. A more complex system (e.g. context awareness) will have more complex problems with it.
I often wish for more context awareness in software, but when I stop and try to visualise it in detail, it seems much less desirable.
This is basically like hypertext and Ted Nelson's ideas outlined in Computer Lib/Dream Machines. He says that our data shouldn't have a strict hierarchy since that hierarchy is subjective and is typically inflexible. What we should have instead is data that can be reordered/represented in different ways. The Everything Bucket apps that Payne is talking about are totally flexible and you can make up whatever hierarchy you like for your data. This is powerful and I haven't used any of these tools, but I assume they aren't powerful enough to allow multiple data representations.
At one point, I created an Everything Bucket program of my own. It was called Nodepad, and was simply a tree on the left and an HTML-rendered-display with editing controls on the right. However, it didn't keep anything in a proprietary database; it literally was just a guzzied-up view of the filesystem.
You opened a "document" by pointing the program at a directory. It then spidered downward and found all directories whose names ended in .node, and added them to the tree. When you selected one, what was displayed on the right was the content of the file "[name].node/[name].txt". (The files themselves weren't rich-text; each file was actually stored in Markdown, and styled by a "document.css" file at the root of the tree. This was to discourage playing with styles while working in favor of getting in, doing your work, and getting out.) It was autosaved when you stopped editing for a few moments or navigated away from the node.
It had no other functionality beyond that. For example, to search for something in the document, you'd just press Ctrl+L to get a "Run command on current subtree" dialog and type "grep foo". I've since stopped maintaining the program, and thus it no longer even runs on my computer, but I still use the "documents" I created with it, simply browsing them in Explorer or in a shell and editing files the old-fashioned way.
If anyone's interested, I might put out a new version. Apparently I'd have to change the name if I wanted to find it on Google, though :)
I want native tagging. When you save a file it should give you the option to add a searchable tag. This is much easier and more efficient than creating a folder hierarchy for everything and then constantly navigating through it. And i want the tags to be independent of the filename.
You want Mac OS 9 or earlier. :-) Well, maybe not quite. You could attach text comments to a file, but I can't remember if the comments were searchable. And before OS 8 or so, the comments disappeared whenever you decided to "rebuild the Desktop." You could assign a searchable, color-coded category to a file, but you were limited to one category per file and something like six or eight categories altogether. Still, though, the earlier Mac OSs were very handy (except for the crashing...)
Leopard is almost there, with Spotlight comments you can tag really easily but it's not [yet] part of the save routine. I don't use comments too often though since the fulltext search is good enough for my uses.
Well those are virtual notebooks. Notebooks usually don't have a fine structure, they are not good for particular data but they're enough good for almost anything that you would like to record. Another point is that those are usually synched, filesystems usually are not, except Dropbox etc.
Usually I use Evernote to collect some ideas or articles that I want to read or think again later. I wont use bookmarks(or delicious) since it's hard to recall from title/tags. I won't print them since I have no printer and don't want anymore paper. I won't save them to my disk since its not synched across devices and it's still PITA to open all the files separately. Also I want to record some automatic metadata and not to input all myself.
I don't think the author's claim that the filesystem works is entirely valid. For a lot of user data, having to pick a single folder in to which a file should go is a poor fit at best. At worst, the filesystem confuses users to the point that they actually don't know where things are going.
I haven't used an "everything bucket", but it actually sounds like a very good idea. I want a program to let me throw whatever is on my mind in to a single place, attempt to make some sense out of it and let me find it later without having to do much manual organization.
The article suggests these programs are poorly implemented. Even if that's the case, Good and Wrong is a start.
I rarely comment on here, but this article is crap in so many ways. DEVONThink is not even close to what he makes it out to be. It's a research tool-- an incredibly valuable one that does /way/ more than he gives credit for. Using these apps as a filesystem replacement is of course stupid-- that doesn't mean there aren't a staggering number of reasons why someone might choose to use one of the apps he mentions.
This is a "look at me" article that doesn't amount to much more than a lot of whining IMHO. Waste of time.
This is true for now. It becomes interesting when we develop context-dependent AI that can crawl through the 'everything bucket' and create structured data from it. An Evernote or Shovebox that could automatically add events to my calendar, automatically filter and categorize incoming e-mails and bookmarks, etc. would be a game-changer.
Disagree somewhat. If every app has to standardize, how does one innovate and differentiate? How does one improve things and at the same time keep everything the same?
How about evolving schemas? Allow for users/developers to collaborate on improving the different data types.
Even though a schema may have a fixed set of core attributes, there's no reason why some app couldn't extend that in any way it sees fit. This way, any "file" generated with this app would be readable by everybody else while at the same time include the application specific extensions. Once the extensions become used by a large enough number of apps, they can be rolled into the standard.
I believe the chief causes of the confusion are (1) different apps save newly created/edited files to different places (2) many users are clueless about navigating the file system within the Save As dialog (3) many users rely on an app's Recent list (4) many users don't know how to search their own file systems.
(1) Some apps always want to Save As to the Documents folder. Some always want to Save As to wherever you finished off your previous Save As. Some always want to Save As to the folder of the file you most recently opened in that app. Etc. No one can remember each app's convention.
(2) People don't realize they can navigate within Save As to someplace other than wherever the app feels like plopping the file by default at that moment. True even when shown over and over and over how to do this.
(3) Once you move a file, its entry in an app's Recent listing is probably borked. This sends people who rely on Recent for finding and opening files into a panic.
(4) They don't know how to click on the desktop and hit cmd-F or to click on Start->Find. And once they do, the Find window looks nothing like Google or a web browser. They don't understand what the various search criteria are or how to enter them. They don't know how to read the results.