These are really good tips about how to manage your files and keep them in order. I think I'll keep them into account to be sure that I'll not lose any data in the future.
I Organize my files into Directories,
and find them based on your filing structure
For me, Folders is a M$ name for something that
already had a perfect name.
They are even moving it to naming it a L..... jugg
Most 'documents'- are in a web site with access security built in and its organised hierarchically and easily searchable too.
I've taken to putting other things in a db with a free form field for descriptors or tags.
eg Things like my (2500 and rising) music scores can be searched for on title, author, instrument, playing grade, key or note sequences etc and quickly retrieved via a web page.
If I think of a better way of searching/indexing the stuff its a simple job to update.
If you and your computer work together you can do amazing things - dont expect it to do the real work for you though.
Being human I do like things to be organized however am rarely so organized myself ;-)
Add to this, as age and early head injuries take their toll, am find the organizing in my mind increasingly difficult to understand :-O
Maybe the problem is that you stop and think about the consequences of what you do instead of just typing by rote with some token charting. This is usually lumped with having a "brain clog", "brain fart", or "senior moment". 45+ programmers aren't as fast as many of those half their age but the work is much more reliable because of experience as a programmer, and experience with long-term consequences in life.
Dave Keays, freelance webmaster
Haiku has user defined metadata built right into the file-system. Copy or move files an the metadata follows the files. The OS's default search program can search by the usual things like filename, date, size etc, but more importantly, by the user defined metadata too.
The search result is live too. You do a search and say, see 5 results. Use another window and add metadata to a file that will match that previous search. Boom! Instantly the search result window is refreshed (without human intervention), now showing 6 results.
This is just the tip of the iceberg of what Haiku's file-system can do. I once watched a youtube demo showing many more neat features. It's a pity Linux's various file-systems don't support any of this. :-(
Alternatively, use something like Google Desktop. That way you can search file content as well - that's if Google Desktop knowns those file types.
I have 42,000 files in my fiction collection alone. My non-fiction collection is 175,000 files.
None of the methods mentioned so far will work given the magnitude of my library. I needed a way to not only find the files but rapidly check to see if I already had the file. Reading the directory contents was useless since your brain seems to go to sleep after a few hundred titles go by. I could not find anything that I liked or that could do what I needed so I decided to build my own system.
First I created several daily cron scripts that use the command "find" to pipe all the file names in specified directories through sed--to filter out things I knew I was not of interest, format the rest so they are easier to read and store the contents in a file. Each script "indexes" specific categories (directories) and creates an individual "index" file for each category. Also I made a script that creates an index of every file in my home directory for those times when I was really frustrated...
I then I wrote a grep wrapper (or engine) to search these different files individually. The grep wrapper allows me to use AND/OR/NOT logic on up to 5 keywords--which also work on the path. This allows you to search the author's name, if you store your files in a directory under the authors name. The search output includes the name of the file and the path.
This is the "basic idea" of how I find files. It is very fast since you are only searching a single file. The index file of ALL files in my home directory, which has 500,000 entries, can be completely searched in 4 seconds.
All the new OSs and windowing systems are oriented towards content consumption instead of content production.
--Steve Daniels 2013
Your brain uses hard-wired links to connect everything together.
Good search engines do good full content, distributed indexing
I am strong in favour of a hierarchical structure, aided with links to related branches, and tags to show affiliation.
I am currently using a hierarchical structure and add symbolic links where I think they should be to affiliate stuff.
To find I using this cascading system:
1. Browse through my hierarchical interlinked structure.
3. finds and greps combined for full content search.
After 2 or 3 I add if needed a new symbolic link.
What most (all current common) filesystems lack is a facility to easily tag and do content indexing. yes a full filesystem searchengine should be in the filesystem.
I collect digital stuff I find along the way, kind of like a hoarder. I may never make that papercraft model of a train station, but I have the PDFs :)
I keep a folder structure aptly named for contents. ebooks, CBTs, videos etc. I further subdivide those with more detailed folder names based on content or subject.
I find the more I look through them the easier it is for me to remember what is in them.
Then I occasionally run Recoll to index them for those times I'm looking for something and want results across the spectrum of my files. Also it's fun to occasionally pick a word and search my file system for it to see what results.
The trouble with using 'search' is making sure you remember what the filename is, instead of what the file is about. Alternately if you aren't searching filenames but file contents, then you have that very expensive disk i/o operation of grepping every single file, each entire file.
The trouble mentioned above is very apparent in Ubuntu's new Unity interface, where one cannot look for "the audio programs I have installed," one must instead remember the name of which audio program you wish you use, and type part of that name in a search field.
I'd like to use tags, but tags don't move when files do. If I transfer a file from my home workstation to my business server, the tags don't go with it, and I'd have to re-tag it on the other side. Too much hassle when I can just put it in the heap where other like files are.
I like my structured and disciplined organization of files - they go in the right place at the outset, and I can find them every time. No matter what software, operating system or fad appears, my filing structure is what guides me to my files. I have never lost (or failed to find) anything that I filed. I also regularly review and occasionally revise the filing system structure.
I detest the current fad of throwing treasured photographs, comments and documents into some mysterious (and often unknowable) filing bin, and praying that the search system will locate things, and praying that you will remember sufficient unique details for a search. And praying that the next software update, operating system or entirely new genre of computing will maintain compatibility.
There is a point you can reach where you're being too specific with directory names and I think the last post has reached it. That isn't a practical approach and would, personally, never use it because it's too exotic.
A solid taxon is somewhat abstract and flexible, and certainly doesn't need to be searchable if you have less than 1000 files. If you don't know where you put something on the file system, it's not worth finding. It generally takes me 2-5 seconds to navigate to files. It's pretty serious if I have to search for something.
I believe that it is better to use standard locations in user-land directories or something similar on attached storage. If you can't replicate it quickly, it's not worth doing. For example, when you're on someone else's dime: if creating a fancy file naming system isn't on the list of deliverable items, the customer will not 1.) pay you for it and 2.) will not understand or give a F why you're doing it, and because they're paying for your time they're right. Most of my customers didn't care about any extras I was offering if affected the project time line. One hour of time spent ordering a file store that isn't an action item is a waste of effort and focus.
Meh - to each his own.
Finally the first to actuall call them directories!!!
Yes, I certainly agree with that.
I do not like the percolating effect of alphabetic or date order of things. For years tried many different schemes and not happy with anything until I came up with a type of grouping of files together using "numbers and categories".
All folders and files have a "number" and a "category" name such as
that would be the folder or directory name where a file would belong with the name familtree or in my system would be named "0002.family.tree" in the folder by the name of "0002.family".
See below how "0002" can be use to cover more folders/files in the "family" category with anything that has to do with the "family".
Heres a sampling of how the "GROUP NO." and CATEGORY" is used to setup my system. I placed the actual names of some folder and files right under the group and category so you can see what the actual combined name looks like:
GROUP NO. CATEGORY AND FILE NAME
0000 QUICKACTION HOT TO DO STUFF GOES HERE
actual folder/file name: 0000.hot.stuff.todo
000 Projects, naming: _pj_name
actual folder/file name: 000.pj_mylinuxbook
00 Directory Organization
actual folder/file name: 00.dir.organized
not in use yet Directory Information
not in use yet Directory Location Logs
0001 God Faith - Devotion - Church
0002 Family Family.time.together - Vacation -
Outings - Relatives - Birthdays -
Anniversaries - Recreation -
Movies - Picnics - Contacts
actual folder/file name: 0002.marriage
0002 Family Family & Friends Correspondence,
(But Not Email) Letters Postcards -
Notes - Messages - Phone Calls
0003 Health Exercise - Insurance - Medication -
actual folder/file name: 0003.food
actual folder/file name: 0003.health.exercise.insurance.meds
0004 Country Duty - Voting
0005 Finance Checking - Savings - Banking -
Stocks/Bonds - Stock-Market - All
Insurance - Bank.Deposit.Box - Will
actual folder/file name: 0005.finance.banking.checking.saving
0006 Work Contacts
0007 Education School - Courses
0007 Self-Study Self Improvement
0008 House ToDo's, Grocery List - Maintenance -
Remodeling Purchases -
0009 Autos Maintinace
0010 not in use
0100 not in use
0300 not in use
0400 Holidays Christmas - Vetrains Day - 4th July -
0500 not in use
0600 not in use
0700 Family.Friends anything but email
0701 Photos Family_Photos
0702 Photos beads
0703 Photos computer
0704 Photos keepers
0706 Photos misc
0707 Photos Linux
0708 Photos not in use
0709 Photos not in use
0710 Photos not in use
0711 Photos not in use
0713 not in use
0714 not in use
0800 not in use
0900 not in use
0990 not in use
1000 Compositions, Editorials, Writings, Letters,
Poems, Song Writing, Written Plays, Written
Stories "Other Than For Genealogy"
1000 Forms, Copies Of Forms, Temple's, Copies Of
Letters, Copies Of Charts, Copies Of Tables
1800 eOrders and other transactions
1800 product review
actual folder/file name: 1800.product.review
1900 Internet News
1900 Internet Radio
Reading Books -
Book Database -
Genealogy -Genealogy stories
actual folder/file name: 2000.history
knot tying -
map reading -
thing to do and places to go
collecting: of Quotes.and.Sayings
3000 not in use
3010 Document Pool - For Projects. not in use yet
3200 My Desktop Reference
actual folder/file name: 3200.tdr
4000 Computer General, Computer Records In General
i.e., Errors-All> Folders For What Caused The
Error actual folder/file name: 4000.errors.all
4100 General bookmarks
actual folder/file name: 4100.mybookmarks
4109 Political bookmarks(9 the number for political
4500 not in use
4600 Linux Distribution Information
actual folder/file name: 4600.debian.guide
actual folder/file name: 4600.linux.dist.info
4700 Computers Hardware
4800 Computer Multimedia
4900 not in use
5000 Computers Software - "Free-ware"
5800 Searches Internet
5900 Sound files
6000 Computer Software - "Paid-ware"
6800 Place To Log Any System Settings, System
Information That Needs To Be Kept Handy Or For
6900 Copies Of Original System Files
7000 Bash Scripts Misc Programming
8000 Computer System How To'S, Man, Info, Tips, Help
9000 Political contacts
9100 Political correspondence - Regular
correspondence - Blog post
9200 Political Information - News - Stats
9900 Political Pictures
9999 No Preset Category Names Here. Could Use Such As:
Default Or Normal Standard Directories Not To Be Changed:
My computer is setup around my life and work. Thats what it should reflect not what a can worms looks like! I know this will look to involved and complicated to most but it occurred over time not over night, and that another plus for it's flexibility.
What you see above is a tiny sampling, I really didn't get into breaking down the categories but to give an idea of what that could look like for example take "9000 Political contacts", it could look like this:
This would be for my political contacts:
9000.polit.contacts (main folder name)
9000.polit.contacts.sens-reps.federal (file name)
9000.polit.contacts.sens-reps.state (file name)
9000.polit.contacts.city (file name)
This would be for another state say New York:
9000.01.polit.contacts.ny (main folder name)
Everything that had to do with Politics would be in the "9000 Political" category, no matter what it is. and no brain strain trying to remember or searching inside files!
I don't have to remember the name of any file I just have to know what the file pertains to. Categories tell me where to start looking even if the file name dose not tell me everything thats inside of the file. Should I forget a number or category, low and behold I click on my "00.dir.organized" folder/file and get an instant review of where everything is located.
I set what percolates up not the system. Just change the number and a file moves up or down where I want it to be!
Nothing is set in concrete, changes are very easy to make. Using numbers allows me to split a category up to make a new category or many different new categories. Rarely is there ever just one file on any subject thats why the grouping.
Nothing is 100%, now and this is more like 75% production. But, when I was using search 95% of the time, that was a killer.
Sorry for the length, it just wont fit in a single or two sentences!
It seems my post raised a few neck hairs. I did leave off a few things. I have not spaces in filenames, for space I use the "." peroid. If it's a space in a persons name I use a "-" dash to show they belong together. Another item is that I date every file at the end of the actual file name. How I use "tags" in a file name. When a file has info in it that is not spelled out in the file name I make that info a "tag" of sort, so I can do a search on it.
Heres an example, I do a lot of political digging on the Internet and this file is a good example of a typical file name with "category number", "date" and the word "tags_" with all the tags to the right of the word tags:
I don't have any large amount of files on my desktop. I keep everything on an older, small form factor pc with a hardware RAID1 for OS (2 x 80GB Seagates) and DATA (2 x 250GB Seagates) 50GBs at /home for some additional space for my wife, myself, and my 3 kids, and 200GBs at /usr/cache for squid, 2GBs RAM, 1 x GigE NIC. It's an Ubuntu 10.04 LTS i386 net install with just what I need running on it.
-SSHD for remote admin
-SAMBA for file sharing - basically where we save files when downloading for local storage
-APACHE to host the samba files in a BLOB so squid will cache its content
-Dans-Guardian for filtering endless amounts of ads that would otherwise get cached too, and logs all traffic so I can keep an eye on my kids content fetches
-Squid Proxy in reverse-mode with 30 day retention (great when ISP has outage)
If I were a real pedant, cause I'm the furthest thing from, I could get something really sweet and searchable using MySQL and content tagging. For now, I'm fine just using APACHE and the metadata in file properties. Besides, you can preserve file properties from their original internet sources - simplifying history searches for content. I let apps take care of content histories.
My network hums along quite well, and everything is streamed from 1 central location - cleaned by ClamAV - so the less techy types in the house who use windows aren't subjected to internet-malefactors.
And God is 0001!?
I find your way extremely helpful, I sometimes implement it on small portions of my files. It would be a good idea to implement this system-wide. As of the rest of the time, I combine many methods: search+tree-structuring, while always trying to keep it related to the theme not the type of the documents.
Having a "music", "videos" and Documents folders is just plain retarded. I have different types of files all over the place!
The first option lets you clean your data base (files and folders) very easily and fast.
A combination of tagging and folders. Folders are GTD-task based (action, projects, reference, etc). This is my setup for Apple-specific devices only as Spotlight does a brilliant job with the above.
It takes about 10 seconds to open the file manager and double click on usually no more than two folders. A search needs a name (something I might not know the entirety of) and then I have to scroll through. Search never sounded appealing.