Slashdot Mirror
How Do You Organize Your Data?
kpellegr asks: "After returning from a well deserved holiday, I was faced with an exploding inbox. While organizing and deleting my mail, I realised I was having trouble classifying each mail into one specific folder. I had the feeling I should be able to link to one email from several folders (e.g. product information should be linked to from the 'vendor' folder, as well as from a specific project folder where this product is used). The more I thought about this, the more I realised that trees (such as the Windows filesystems) are not really ideally suited for organizing data. On UNIX-like filesystems, symbolic links allow the creation of simple graphs for organising data, but I have the feeling data could be organized more efficiently. How does the Slashdot crowd organize their data? How do you manage files, email, contacts, meetings and all the relationships that might exist between them?"
This is exactly the concept behind virtual folders. The idea is that folders, whether they be in the context of an email program or a filesystem, are actively updated searches. For example, all of your emails could be in one pool, invisible to you. Then each folder would be associated with a rule similar to email filter rules we use now. If an email matches, it shows up, maybe in multiple folders. Bayesian rules allow for even better classifications, if an email is similar enough to several catagories, it can show up in all of them.
Spencer Ogden
A dash of arbitrary directory trees and a pinch of grep.
But seriously, this subject is kind of lacking. The problem I have with organized storage is keeping it organized. I don't have the time nor the will. I need some sort of automagic organization.
What I do is I realize "This is all a mess" as I see thousands of files in my home directory. So I created a bunch of subdirectories for various things. (Some were logical, some were just by file type -- /png, /txt, etc.) Then, I found that making such an organized structure was too complicated, and stopped before I acually moved anything into any of the subdirectories.
On Windows, it's slightly different. I save everything to my desktop, then when it gets 'full,' I delete just about everything, realizing I no longer need it.
Not that I RECOMMEND these strategies, but it works.
________________________________________________
suwain_2
The good news is, that while the Window's file system may not support this, if you wait until 2005 (2006, 2007?), this highly demanded feature will be in the next release of Windows -- yes, everyone's favourite Longhorn will turn everything into a database.
Frankly, I don't think turning an OS into a DBMS is the right thing to do, but for certain applications, having this functionality omnipresent will be useful. Well, OK, for this one application, I'm still waiting to see examples of others.
A friend of mine used to use what he termed an archaeological filing system.
It was based on the simple principal that the older something was the further down in the pile it would be.
Your all-in-one-folder technique and "ls -t" would work equally well.
Boffoonery - downloadable Comedy Benefit for Bletchley Park
Hmm...this isn't how I do it, but what about something like this:
Disclaimer: This works in theory; practically it would require a hell of a lot of resources.
The basic idea would be a relational database. You've got say the files in one table, and categories in another. The categories can have a parent, so you get something of a tree view going. Then, when you select something from a tree view, it comes up with all items from that category.
Creating this would be easy; optimizing it wouldn't be.
I have my home directory (on a redhat box). the root level of my homedir is my crap directory. dangerously, the organized data lives in there in directories. for example, i have an Organized directory. in there, i have sub directories of different types. i wont go into how that is because i'll change my mind about it once i write it up, and i dont feel like redoing it right now. so, everything just ends up in my home directory until i sort it.
/, and then have a /system/libs/stuff, /system/configs, and all that good stuff. either that or just not see it period. speaking of period, i guess it'd be ~/.system instead of ~/system. I hate organizing my stuff, too. It's arguable how much easier it is to find once I organize it, if my mind one day decides that it should be "schoolwork" and another day "development", et cetera. i guess organizing your piles of junk is like your fingerprint... everyone's is different.
i wish my view of the system was more abstracted. i'd rather have my homedir as
another thing i wish, though, is that the filesystem were more... i dont know what to call it. but i wish i could store more meta data about my files. i wish my filesystem had a comments field, and i wish that doing a directory listing would spit out file attributes like dimensions, content length, number of words, and whatever other stuff i could glean by hand. i just want it to all show up. hell, i wish i could do a recursive directory listing based on file type, not file name. and not based on the extension... cause who says i use extensions? (of course i do, what are you, daft?!) unless its a text file. unix spoiled me and i dont put extensions on those.
heck. i wish there was a way to just export my entire home directory with everything i said into one giant 22 gigabyte compressed file that i can save somewhere, drop into a new computer, and just be up and running again just like that.
slashdot: where everyone yells sarcastic metaphors to themselves to understand the issue
Seriously, I try to keep different partitions set for specific things, this helps in case something gets borked on one drive, it won't mess up other partitions, of course there are backups made to ensure not much is lost.
Try doing something like this (if on *nix)
- /dev/hda3
/home/$USERNAME/pers (personal stuff like diaries or so)
- /dev/hda4
/home/$USERNAME/codes (if you're a programmer)
- /dev/hda5
/home/$USERNAME/music (take a guess)
Get the picture? The good thing about this setup is, one could always umount in case someone gets physical access to the machine, heck it could be scripted to mount and unmount on login and logout. Or you could encrypt the partitions for added security.At first it looks bulky, but in the end it's very easy to maintain since everything tends to fall in place. e.g. If you're scripting you could just cd /home/$USERNAME/code and not have to wonder where to save this. Unless you're really odd (like me) and begin everything with test.c or test.py or something.
MoFscker
All mail are kept into one place (say, a MySQL database). You, however, setup filters (that is, SQL queries) that show your e-mails in virtual folders.
That is, messages can be in as many folders as they meet the selection criterion of.
In addition to the obvious "from", "date", "subject", you could assign an arbitrary number of categories which could constitute more selection criteria.
David Gelertner, the comp sci professor author and unabomber victim, has created software he calls Scopeware. It basically organizes information in a series of related chains. These can be date based or otherwise. I haven't used it, but I've read that he is responding to some of the same concerns you mention.
On a less lofty, but free, note, Evolution has "virtual folders" in which you can place anything a filter expression can select. I use them to sort my email by sender address. I still have my main inbox, and all the categorized subfolders, but the virtual folders select particular people out of the massive mail database. So I can recall that Joe said something three weeks ago that relates to a current problem, and look in the "Joe" virtual folder to find it. There's still no easy way to add arbitrary messages to a virtual folder, other than adding a filter rule that selects just that one message. At least I haven't found a way. But it seems to address part of your concern, for email at least.
"Even if you are on the right track, you'll get run over if you just sit there" - Will Rogers
Or is that KAOS (as in "Get Smart") ?
I'm currently playing around with putting all my mail messages, bookmarks, web pages loaded, file accesses (on a day to day basis) into a database. Maybe not all the actual data, but the stuff that might help me find it when I need it. I'm hoping to eventually scan everything that changes on my computer or that I do for keywords and so on and then organize them so I can browse them by some kind of visual graph/map metaphor on any of several axes (type of file, date/time, keywords, directory ....).
I want to be able to go in with a query like "sometime in july I did something having to do with a picnic and watermelon" and get a list of possibilities, then be able to rate those in the hopes of finding the exact info I'm looking for.
OK, so far I only have some pieces of it. But I'm getting closer to a database schema for the information and that will help me figure out better what info I need to collect.
As many people will probably point here, you should check out Evolution's "virtual folders".
JWZ once proposed a more sophisticated approach to store mail without the hierarchical folder structure limits. You can read about it here: Intertwingle
I don't what came out of that. I think it is a good idea still waiting to be implemented.
I know other people have mentioned Evolution's vFolders, but here a little more.
My goal is to never have an email that has value to me land in my inbox. Every time I get an email of "value" which stays in Evolution's inbox, I right click, and "Create Filter from Message". (I'm paraphrasing.)
Every good message should have at least one filter putting it into at least one folder. Some emails have more than one rule, but the whole right click -> create filter thing makes this quick and easy.
-Pete
Soccer Goal Plans
Anti SCO T-Shirts. Donates to the Open Source Now Fund with each purchase.
Well, I hate to say it, but Groupwise (while being awful at a lot of things) I can organize my emails rather quickly and rather well.
.02
The Link-To feature allows me to store it in multiple folders at once.
I especially appreciate the Shared-Folder. It makes it easier for me to make emails, documents, etc, available to those that haven't a clue. The IT department is busy working on too many other things and the Novell iFolder is unacceptable for my use (my other option).
I wish I could make subfolders under Search Folders but that's for another version maybe.
Just my worthless
Keeping email organized is a lot harder than it should be. There is no good way to deal with things like a seminar announcement that I need to keep for two weeks but is junk after that, or stuff that I need to remember to read or reply to but don't want to read right now (or stuff I keep because I should read it but don't want to actually read ever).
It is also hard to remember that, when someone emails me some document, the place to store it is not in an email folder, but a directory dedicated to that project or subject. Like if someone sends a reference for a paper I am writing, it should go in ~/papers/journalname/papername/references or something, not just stay as an attachment in my inbox.
And once in a while, you have to waste a day or two reorganizing your crap and deleting old email. This is especially hard when I have copies of documents or programs on different computers, because I have to figure out which ones are the most recent and are the authoritative copy. CVS and rsync help here; CVS makes it obvious which copy is the best one (the one in CVS), and rsync makes it easy to keep things identical on different machines so you don't have the problem to begin with.
What was the question? Oh yeah. Let google index your entire file tree and use it to find stuff.
I think MIT has a project called Haystack designed just for this
Back in the day, BeOS dumped all the messages in one folder with each message as a separate file (like Maildirs) and used file attributes to add any label (or set of labels) you'd like.
1.msg [ classification=Spam ]
2.msg [ classification=Inbox,classification=Spam ]
The desktop interface let you sort out files based on their attributes. Better e-mail clients also understood some of the common file attributes.
Linux now has attribute based filesystems that are getting mature - it should be possible to do something similar.
A DYI solution, but what on Linux isn't? (:
www.thebrain.com
It has an interesting way of organizing data - and you can link any data item from multiple places. It is a very interesting idea and I have played with it some. It can link to Lotus notes messages as well.
One disadvantage is that (besides the Windows platform) that it is the entire environment you'll "live in", with your data. I bought it at work and tried it out, and the first impressions are good. If the company has a broadbased support and widespread adoption, I'd probably use it, because at work we seem to be keeping the M$ platform for at least a few more years.
You can check out the personal version free of charge - download from the website.
We got a call last week by a customer that was having problems with mail.app (OS X email app) getting "poor performance". Come to find out she, and most everyone at her company, had upwards of 2000 emails in EACH mail folder, and they had many mail folders.
Somewhat at a loss for good ideas, I suggested she try Enterage. That's apparently what they used to use, until they broke its limit of a 2mb index, at which point Enterage crashes.
Sheeeeeeesh. Some people just don't know what it means to keep a clean email inbox. But in her case, their business revolves around receiving customer email, and they're already keeping their mailboxes trimmed down.
Is there any email app for OS X that can handle "industrial" needs?
I work for the Department of Redundancy Department.
Ted Nelson's ZigZag system is a new way to store related data without resorting to a relational database. At first glance, it seems really goofy. This is usually an indication (to me, anyway) that it either really is completely goofy, or brilliant beyond my comprehension. Given Nelson's record, I'm inclined toward the latter.
Schwab
Editor, A1-AAA AmeriCaptions
Ugh... hate to say it... Outlook client using exchange.
There I said it. Ok, to be fair, I use it because that is what is available and that is what everyone is use, all 800 or so of us... and that is in our org, which is a child org to a much larger org... so a total userbase of about 6000 users...
Here's why it works. I use partially Bayesian based InBoxer to kill spam. Our exchange server also runs Norton anti-virus (which has saved us from SoBig all that crap)... and then the exchange also has a spam filter which adds "spam:" to the subject of all incoming know spam e-mails (which does me not much good).
Ok, that takes care of spam. All list-serves I belong to get put into their own folders... Emails for friends get put into a specific folder. This leaves my inbox. My inbox is shared with all my 'trusted' co workers. When I am gone, they check it on a regular basis for me while I am gone. If I am expecting a high priority e-mail from a certain person, I set it up so an alert e-mail is sent to the right person then that comes in.
For my tasks, that is also shared. When I am gone, I forward my tasks that are due during that period to the right person.
My calander is also shared. On my calender, I mark when I will be gone, and then setup a special list of those who should be alert when they send me an e-mail or task during that period (this stops an e-mail alert being sent to those list-serves I am on when I am gone).
As for files: I manage the share on our central server that we all use. We just went through a major undertaking to get it up to par. ALL files are saved on the server. Everyone has a private drive, and then each 'task' or 'subject' or 'project' has its own folder on the server. Some folders are public, or 'all on our domain'... a majority are 'departmental access' (every one in our small org)... the rest are specific, generally with 3-4 people.
It takes work. But I have access to the files I need and so do the other people in my org. It takes a lot of user education, training, and hand holding.
Couple all this with decent VPN (cisco based) and most users get what they need when they need it.
Oh, and this is at a college. Most departments are as well off as we are. And, yes, slammer has been a bitch to deal with as students move in... but many dedicated staff have solved that problem (not to mention some ingenious network guys... hats off!).
I tend to organise my mail by who it is from or who I am having the conversation with. If I am having the conversation with several people, I put the email into the folder with the person who started the conversation.
If there is product specific stuff that I want to put in more than one place, I tend to copy it to text or word or whatever format docs and save it into folders.
Now I am entirely dependant on filters to store stuff into the right folder. Usually all that is left in the inbox is spam or new contacts.
There are things for sales or support staff called "contact managers" or "customer resource managers" (CRM), which let you link up documents and mail and even records of phone conversations and reminders in a more intuitive fashion. I've yet to decide which one is best even though I've spent months trying to figure it out. I guess it is too far away from how I work as a programmer (mostly). There are these ones for example: Le Grand
ACT!
Microsoft have one that they got from Great Plains software
And there is one unix based one that I know of in Finland! Nemein Hmm, having trouble getting it to load but it was there last January. Try looking for Nemein.Net Sales just to prove I'm not imagining it review
Anyway I think some of those things are completely over the top but if your email systems are out of control they may help.
-- it must be true, it's on the internet.
There's always The Brain (thebrain.com) which has a pretty high geek factor but works on a fairly simple premise that data can be organized many different ways in ones brain and provides paths to information based on those associations.
"Do not be swept up in the momentum of mediocrity." - anon
While I don't use it, I beleive Opera 7's mail client is designed to allow you to do just that. You can create as many categories as you like, and any given email may belong to any combination of them.
First, email:
.Net, ASP, PHP, My SQL, Delphi, Outlook, MS Word, etc. all rolled into one. It can link a single email into multiple folders since Notes 3 in 1995.
Get Lotus Notes. I am a Domino evagelist. It is my god! I will spare you the details, but the Domino server and Notes client is the same as
Next Filesystem:
Linux boxes are highly organized already so for Win/Dos/OS2 systems:
Create an OS partition. Only install the OS and patches and device drivers here.
Create an Application partition. Install all applications in a heirarchy starting with Apps and Games and branch apps into Office, Development, Graphics, Utils, MediaPlayback, etc.
Create a Data partition. Create a \Data and a \Temp. Under TEMP put a Downloads, CDBurnoff, Ripping, Testing, and Receipts & Status. Under Data put Documents, Media, Source, Settings, and Pictures.
That should get you started. Create a catalog.txt file and put it on your desktop with notes about where everything should go. Good luck scaling up to 200 GB and keeping your head from exploding!!
SCO: 800-726-8649
Verisign: 800-361-8319, 888-642-9675
Diebold: 800-433-VOTE (8683)
It helps to have a filesystem designed with Database features in mind (ie. just like the BeOS file system). Emails are stored as normal files, with attributes like To, From, Title etc stored in the database. The same concept can be used for media files (MP3 attributes are stored into the database). When you wish to search your data, you can write queries, which are live on the BeOS, and have the results displayed in a directory window.
:-) have been using this feature for years now...
It's rather awkward to explain, but it works amazingly well in practice. Once you've tried it, you realise that there is no need to store data in directories, just make sure that the attributes are up-to-date, and finding any file is a query away. Rumour has it that Windows will adopt a similar system in Longhorn. Yeah, we BeOS users (all 20 of us
Revolution = Evolution
I think this is a step in the right direction. I have been using it for a while now - check it out.
"The goal here is to do for email (starting with your personal mailbox) what Google did for the web... The Google principle: It doesn't matter where information is because I can get to it with a keystroke. So what is Zoe? Think about it as a sort of librarian, tirelessly, continuously, processing, slicing, indexing, organizing, your messages. The end result is this intertwingled web of information. Messages put in context. Your very own knowledge base accessible at your fingertip. No more "attending to" your messages. The messages organization is done automatically for you so as to not have the need to "manage" your email. Because once information is available at a keystroke, it doesn't matter in which folder you happened to file it two years ago. There is no folder. The information is always there. Accessible when you need it. In context." ZOE
I put all e-mails in one folder and if I need to find anything I use grepmail:
http://grepmail.sourceforge.net/
If you have OS X check out Zoe:
http://guests.evectors.it/zoe/
I used to beta this thing by this company called Autonomy which would sort and sift all your (and everyone elses) cruft to assemble a list of relevant links (to your stuff and others) in response to your activities.
;-)
IMO it did this in real-time, must have made for some impressive indices.
Maybe this is the answer, open-source Autonomy. I am a mere perlmonks acolyte so I will leave it up to the real brains to figure it out
thebrain (www.thebrain.com) could be used, or at least something like it, to handle this. Basically, everything is treated as a thought and can have multiple parents, siblins, childern.
I've always classified thebrain as 'really neat', but not very usable. Once you put a lot of information into it, the interface becomes difficult to use. But the concpet is still sound.
I have a large storage. I found that if you want to organize anything without symbolic links you must stick with a strict model. In my root I organize any platform specific applications and others into platform named directories. (eg: Win32, Linux, MAC, Palm, PPC, etc...) I also have directories called Incoming, ISOs, Media, and Projects I also have a seperate share for my, and my wifes user profiles.
Under the Win32 folder I have a scheme of Applications, Gaming, Drivers, Servers, etc... Under Linux I have a scheme similar to Source, Binaries, Modules, etc...
Once I decided on that organization model it was really easy to just keep expanding the same principle to subdirectories.
You talk better than you fool!
Not sure if this has been mentioned (probably has), but the new Longhorn release of Windows is supposed to be shipped with a new file system (WinFS) which does exactly what you need. It (again, all just theory right now) will work by using a SQL database instead of a FAT table. This means you can now classify files.
So you'll access a "folder" which basically has a list of properties, and all files with those properties will be show. So if I want all my pictures from my vacation to hawaii, as well as my monthly financial reports, I'd create a folder that "contains" all files on those subjects, and whenever I accessed that folder it'd show me all files that fit those catagories. But on the same hand I can have another "folder" which shows me just my vacation pictures.
I have pondered the same thing. Being a relational fan, I of course lean toward sets instead of (or in addition to) trees. Here is my webpage describing various post-tree approaches and interfaces:
http://www.geocities.com/tablizer/sets1.htm (I know, geocities sucks, but there are too many links to it already to switch.)
Table-ized A.I.
I know other users have already pointed out how well Evolution works for sorting mail, but I just wanted to attest to how well it works even for large amounts of email.
I used to create new folders for specific types of email, but I found it very difficult to manage and search all the folders after a while, so I ended up moving all of my email to a single folder, Inbox. I currently have 24,949 messages in my Inbox and Evolution is still extremely fast when it comes to sorting and searching through them all.
I also make use of the excellent VFolders feature of Evolution, to save frequent searches into their own folders. I've been using Evolution now for several years, and it just keeps getting better and better.
--It's Pimptastic!--
For file systems I use symbolic links in a column viewed filesytem. I really like what a company formerly known as NeXT has done with some of their products. Their software for pictures and music both have a "Library". From there you can drag songs or pictures into "Playlists" (music) or "Albums" (photos).
Very cool.
As for software, I use OmniGraffle and OmniOutliner from OmniGroup. OmniOutliner is especially simple, yet unique. I wonder why no one else has an idea organizer that is so incredible? I couldn't do my job without it. Well, I could, but I'd use a lot of paper or spend a lot of time in OpenOffice messing around with things.
We use lotus notes at our organisation and I find it a great email client (for corporate mail). It has great folder support, there's only one document in the database, but it can be displayed in any number of folders. There's also the views which are defined by the admin which show what they're told. This means that you might remove a document from all folders, but still be able to find it in All Documents.
The only down side to Notes is that it's quite expensive, but it does run on Linux *grins*
Z
Most schemes for organizing information just result in more folders that need organizing. I think we can learn a thing or two from meatworld systems... like libraries.
I use faceted classification to name my folders. Specifically, I use a variant of Ranganathan's Colon Classification. Each folder has four facets applied to it: Matter, Energy, Space, and Time. Ranganathan also specified a Personality facet but I can never figure it out. Instead, I assign a proper noun. The Time facet is also redundant given the time stamp on the files. Instead, I replace Time with Stage.
So, here's a typical folder name for a sales project I'm working on:
Personality- Smith (name of company)
Energy- Sales (activity related to contents)
Matter- Presentation
Space- Montreal
Time- Stage 3
The resulting folder name is: Smith_presentation_sales_montreal_3
It looks like ass, but it works pretty well since the facets are semantically orthogonal. The syntax makes searching and sorting pretty simple.
Good luck!
For me, I find that the ideal way to organize things would be a combination of folders, piles, and stored searches (or virtual folders). While folders don't always meet my need, I still find that they are a perfect starting or ending point for information. Say if I'm starting to research on "Small World Networks", I almost always start by creating a folder, to place collected data. But while researching, I often find the need for saved states--its one of the primary reasons why I enjoy tabbed browsing. Being able to quickly saved 5 or 20 open sites, and then go back to them is just great.
The concept of piles, I love as well, the ability to just quickly dump things into piles of interest, that can be later organized better, is a great concept. Typically when something is active it need to be related to lots of other things, and for me the piles concept allows this. Of course this brings me to stored searches.
I find I used the concept of stored searches like data mining. I typically do it for past projects, or to gain new insight on past experiences. I remember Sherlock use to allow you to save search criteria as a clickable link that would then rerun and update the results when selected. I don't think the new version does, but I am hoping that Apple will introduce the opportunity to do so, when they start making improvements on the new finder to be introduced in Panther.
porn1
/home kills my /home disk freeocity, so I move some stuff to /usr/local, and set up a symlink. Then /usr/local runs out of diskspace, so I set up another symlink to /var, etc. Eventually, it all comes crashing down when I can't make a symlink in /dos, because of stupid lacking features of a dos fs.
.avi files though, that I didn't know I had :)
porn2
New Folder
New Folder(1)
unsorted_porn
mp3s
I made the mistake of making too many partitions on my drive. So my porn on my
I'm sure I've got all this porn stashed away somewhere on some random partition on my drive that has no links pointing to it, so I'll never find it. I love it when I do find a 1GB stash of
Have a look at http://guests.evectors.it/zoe/
It is wonderfully easy to use, and does everything you want. Oh, it does take a bit of getting used to.
Evolution has addressed this with its email "VFolders" for some time; these act much like a saved search across all of your folders.
A simple example is creating a VFolder that will show all items flagged "Important", allowing you to immediately view and modify any such email. Any changes you make to messages within a VFolder applies to that actual message, wherever it resides, kind of like a hard-link.
After a lot of vareity of putting stuff in folders I realised that my mind does not think
the same way while organising, as while looking for something. I never seem to find
the right folder when I want it.
So now I am using evolution, put all mails (except SPAM, CVS, Bug reports etc) in INBOX
folder and create virtual folders based on keywords. But most of my successful hits are
when I filter for keywords over this INBOX folder as I need info. Its works 90% of the time.
Infact another rule in conjunction to this: Never delete anything
I am next going to break my INBOX into separate folder for each 3-months and try doing the
filters over the whole set of inboxes.
DO NOT PANIC
The most USEFUL directories you can create are three personal tmp directories, like so:
- tmp
- tmpSep
- tmp2003
When you do that, you instantly get rid of clutter: small files that you don't really care about, but want to keep for some finite amount of time.There are many more techniques. I'd like to write about them some time, but now is not that time.
The base directory describes the block.
Take for example software. There are two possible ways of sorting this: stuff from vendors and stuff by structure. I use both, but the majority of stuff gets stored in the vendor tree, and the minority under the opsys tree. So if i want a non-descript OS/2 utility for file management, i would look under opsys/os2/fileman/ while something from say file commander/2 [which i use a lot] is /vendor/fileman/harvard/os2/.
Personal stuff gets stored under the tdisk tree. These are grouped under broad catergories, eg 'maths', and then a date directory. eg: /tdisk/maths/nbfk/
The whole idea is if ye take a bucket-load of backup cdroms, ye should get a single tree that is easy to sort through.
OS/2 - because choice is a terrible thing to waste.
It sounds very much like the concept of WFS in Longhorn. Virtual folders should be done easily with a SQL-based filesystem. However we still have to wait at least two years for Longhorn to arrive.
The idea is to "attach" qualifiers to data, so that the data doesn't have to be ordered in a hierarchical way. The data is looked for dynamically, as the system creates a tree structure on the fly, based on the qualifiers the user has attached to the data.
Example: A novel on the history of mathematics would by one user be stored in the folder "History" and another user would look for it in "Science". The ICMS solution lets the user attach "History", "Science" and "Novel" to the book, so that he himself (and other users) can find the book by looking in the folder "History" AND/OR "Novel" AND/OR "Science".
Neat eh?
(If you're interested in buying anything from them, contact me at mathieu.dhondt at quatris.be - I'll give you a discount).
"Those innocent fun games of the hallucination generation"
I organize my emails by putting everything in a single folder. No need to agonise over classification or get grumpy at myself for misfiling. Then I use ISYS to find whatever I need to find, using a plain english description of my need. Works a treat. ISYS is a swiss army knife search tool, but best of all, there's a stripped-down, email-only version coming out in a couple of weeks.
You understand the problem; do you understand the solution?
Computers have advanced well into what were just dreams even in the 1980s. But we are still stuck with many of the paradigms that were created to:
- Make computers useful when they were very slow.
- Make computers understandable to people who did not grow up with them.
Today, many of the concepts should be obsolete, but I have not heard of any real advancements other than finding uses for the better bandwidth. So I started a company to develop what will hopefully be the next step.
I believe you need to understand both the old fast relational database paradigm, and the newer but slower document-based database paradigm to be able to see what comes next. Luckily, most techies hate Lotus Notes. The former secretaries and managers who do work with it have little concept of how revolutionary it is; they just like that they can understand it. I hope this leaves the field open for me and my company to become in the next 10 years what Oracle was back around 1990. Wish me luck, or try to compete. Either way technology will advance.
I spend my life entertaining my brain.
I'm very spoiled. We use Lotus Notes at work for mail and applications. Notes has a construct called multivalue fields that can be used to create multiple categories (similar to folders) for grouping and sorting of data. I can file mail in as many categories as I like via use of categories. Simple idea that works great
The Good, Easy system worked for me back when I used it. The premise is to get everything into plain text, and use simple tools to manage it. There's a Wired article on it, and the source documents to the Good, Easy Desktop and Good, Easy email are at Winterspeak.
i had a similar problem a while back and i started beating on it with postgresql and perl. it's actually getting to a nice polished state. currently it's webbased, but i think my roommate wrote something for the console also.
/mnt/fsroot/set1/a/2/3/a23f55abab...
here's the skinny. i store files in a tree based on their checksum. so the file with the checksum a23f55abab... would be stored like this:
then i store pointers to this file in a database along with different metadata (mimetype, original file name, keywords, mount point (set1), etc). then i define lists based on queries to the database.
so i could have a list like:
images::vacation::italy 2001
the images would have keywords like
italy, vacation, 2001
and the query defining that list would look something like:
keywords:vacation and
keywords:italy and
keywords:2001 and
mime_type:image
i need to commit this stuff to the CVS, but it seems to work.
-- john
well, for security reasons you may want to make some areas noexec, and/or nosuid (ie /tmp).
/home as a partition makes OS upgrades easier.
That's a good reason to make that a separate partition.
And having
And having /home as a partition makes OS upgrades easier.
/usr/local, although you could put /usr/local on the same partition as /home, and do a mount --bind from one to the other. Too late now for my drives, but given the read-errors I have been experiencing on my laptop the last few days, perhaps it is time to backup, low-level format, and reinstall?
You're too right about that one!
For tha very reason, I have the same justification for
Apart from the reasons stated by some other guy, I think it's nice having stuff that gets changed a lot (/tmp, /var) on different partitions. That way, I can make sure those changes don't interfer with and fragment more important file systems. And why not make / read only for security reasons? You definately don't want a read only /var. And probably not /usr if you run FreeBSD (updating ports frequently, but having base locked down until you need a security update or upgrade).
So, for email, I keep folders to a minimum of about 6. But because time is so important, I tier those, so that anything older than a week goes into a mirror folder structure under OLD. Then anything older than a month is moved into ARCHIVE. And the archive stuff is compressed, so I have to really want to look at it.
Loose information is another problem, with a simple index-card like solution. A lot of the information we need is small, like "joe's phone number", and doesn't warrant a whole file. For that I actually throw all the information into a single big file, where each datum is one line (grep-able). The information has no structure. I often cut and paste random stuff. Then I have a search that just pulls out entries which match all terms:For filesystems, I found the reiser guys have some very pertinent ideas, albeit in need of further development. http://www.namesys.com/whitepaper.htmlFor shared stuff, I really like having an unofficial document system, and my favorite is The Moin Moin Wiki because it's fairly simple to use and install.
/charles
I also have used the brain and found it interesting.
:
Big downsides however are
1. It is not free software.
2. It uses a proprietary data format, which remains unpublished. This means that your data is locked up and only accesible through the proprietary software it was created under.
These two points constitute a contract I am not willing to accept and thus I do not use nor would I recommend this product.