[joefru]

This has been beaten into the ground on every photo forum known to man. Regardless, let's do it.

I've been 'serious' about photography since 2003. Right now, I have about 25,000 photos. F-spot for Linux automatically puts images into folders by date: Photos/2008/04/30/IMGP6025.JPG. Before I was using F-spot, I would manually do things like Photos/2006/02/Feb-15-Molly's_birthday_party/P1010568.JPG.

The latter option seems much better, especially since I have almost no photos tagged within F-spot, and it won't write EXIF data to RAW files (only JPEG). The problem I have with naming folders by event is, what do you do when you have ONE photo of something? Like I'll have 2008/04/Apr-29-Kids_at_the_park, but in the middle, there's a nice bee macro. I don't want to leave it in that folder, because it's not a photo of a kid. But on the other hand, I don't want to create a folder for one picture. Similarly, what if I love taking flower pictures? I might end up with:

2008/04/Apr-15-Flowers
2008/04/Apr-16-Flowers
2008/04/Apr-16-Photos_at_the_pool
2008/04/Apr-17-Flowers

In a case like this, it would seem that creating one big "April Flowers" folder would make sense. Maybe you could subdivide it by daisies, roses, irises, etc. But that seems to go against the date organization thing. In fact, why limit it to April flowers? Why not Spring Flowers? Or 2008 Flowers? Or, hell, just "Flowers"?

Another major nagging issue is that of quantity and manageability. There is no reason I need 26000 photos. None. On the other hand, I can't bear to throw any away. That screams ARCHIVE! But how? If I archive EVERYTHING and keep only a couple hundred really good pics a year, the folder thing seems to only be necessary for the archives. Then how do I organize the keepers?

I know everyone has this same conundrum. I'd love to hear what everyone else does.

Joe

[BBear]

This is what I do with my photos, and keep in mind that for all these years I've been using ACDSee Classic (no EXIF capabilities) and I have over 30,000 pictures taken since 2001 after my 1st digcam.


I just have FEW folders, organized yearly / country as follow:

Folders: named either USA YYYY or CAYMAN ISLANDS YYYY where YYYY is the YEAR.


Inside each folder, I have the following naming for files:

Files: MM - TITLE/EVENT - XXX where MM is the month (04 for april), TITLE/EVENT is just the description (Halloween or Xmas or DC Trip for example) and XXX is just the numbering with at least 3 digits.


Well, trying to explain its not easy, but actually i can find any photos pretty easily IF i can remember the YEAR it was taken and even better if I can remember the month.


Maybe I'll do a screenshot later when I get back home. I like this system because it does not have TONS of folders, which are a PIA to browse around.



*********************

On the other hand, i've seen new Media Management Programs that organize all your photos into a library.. like Winamp Media Library does to your music media. If I didnt have a system working for my needs I'd considered switching but i dont think it is worth for me at this stage.



BB

[BBear]

I don't want to create a folder for one picture. Similarly, what if I love taking flower pictures? I might end up with:

2008/04/Apr-15-Flowers
2008/04/Apr-16-Flowers
2008/04/Apr-16-Photos_at_the_pool
2008/04/Apr-17-Flowers

that is the problem in my opinion with multiple folders My friend simply keeps the Canon folders as they come from his camera (I think it is simply the dates) and it is a total PAIN in the A$$ to find pictures in his PC... because every single day it is a folder. He has a HUGE list of folders already and some of them have only a few pictures while others have tons.

[c13_]

I make a group of base folders eg. Macros, Landscapes, people, Events, etc. Within those folders for large groupings of photos in these I make subfolders but leave the odd single shots in the main folder. It looks kinda like this:

c:\Pictures\Abstract
c:\Pictures\Landscapes\Trip1-xx-xxxx\IMGPxxxx.PEF - IMGPXXXX.PEF
c:\Pictures\Landscapes\Trip2-xx-xxxx\IMGPxxxx.PEF - IMGPXXXX.PEF
c:\Pictures\Landscapes\IMGPXXXX.PEF - IMGPXXXX.PEF
c:\Pictures\Landscapes\Wildlife\TYPE\LOCATION\IMGPxxxx.PEF - IMGPXXXX.PEF

and so on. I find this keeps it pretty logical and allows you to place those odd 1-2 shots where you can find them.

[joefru]

At one time, I did a subject-based organization scheme, but I sometimes like to reminisce about a period of time by looking at the photos I took, chronologically. Separating them like that makes it impossible.

[BBear]

I make a group of base folders eg. Macros, Landscapes, people, Events, etc. Within those folders for large groupings of photos in these I make subfolders but leave the odd single shots in the main folder. It looks kinda like this:

c:\Pictures\Abstract
c:\Pictures\Landscapes\Trip1-xx-xxxx\IMGPxxxx.PEF - IMGPXXXX.PEF
c:\Pictures\Landscapes\Trip2-xx-xxxx\IMGPxxxx.PEF - IMGPXXXX.PEF
c:\Pictures\Landscapes\IMGPXXXX.PEF - IMGPXXXX.PEF
c:\Pictures\Landscapes\Wildlife\TYPE\LOCATION\IMGPxxxx.PEF - IMGPXXXX.PEF

and so on. I find this keeps it pretty logical and allows you to place those odd 1-2 shots where you can find them.

So let's say you have a July 4th Picnic and you take 400 photos... some family member portraits, some landscape, some of the kids playing, and whatever.


how would you keep them? would you split them into the correct folders (abstract, landscape, family, portraits, etc)?

[monochrome]

This is an issue that database designers encounter all the time. What you are trying to do is very similar to what they are trying to do - i.e. design the fastest, most logical indexing protocol to retrieve records (your photos) ad hoc. First, remember that the problem is not efficient record SAVING - it is efficient record RETRIEVAL. I think you understand this problem implicitly, at least.

The problem is, in order to do what you want to do you have to give a fully-descriptive name to EVERY photo file before you save it, and you have to have a common naming convention.

Each record in a database has a unique identifier, and each may also have "group" identifiers that logically link records together. To retrieve information from a database one creates a "query" (or uses a template) that tells the program to retrieve all records that have certain identifiers. Those identifiers would be in your file naming convention.

Databases are designed to suit the needs of the users. While there are certain common structural characteristics, there is no one structure that is universally suitable.

To structure your folders for efficient photo retrieval:

• Decide how you perform searches for specific files or "records" (photos)
  • There will be more than one - rank them by importance or frequency
    • Date? Subject? Equipment? Category?
    • If you are a professional, by job or client?
• Understand how your brain is structured (this is a real consideration)
  • Do you principally group your experiences and thoughts by time or subject, to keep it simple
  • In many ways your photos are records of experiences - how do your categorize your experiences?
• Structure your folders from the broadest category to the most granular
  • Year/Month/Day/Subject is obvious
  • Client/Job/Subject is common for professionals
  • Subject/Year/Date~time is another - I use this one
    • Vacations
    • Flowers
    • Landscapes
      • Vacations
        • Cayman
        • Smokies
      • Missouri
        • Ozarks
          • Irish Wilderness
            • 2004
            • 2007
              • (I made a new sub-folder and moved the files after the 2nd trip)
    • Cityscapes
      • Kirkwood
        • Station
        • Greentree Parade
          • (years)
    • Friends
    • Events
    • etc.

The problem comes in when you want to search thousands of records for all photos of flowers, for instance, regardless of how you originally structured your retrieval convention. You may have taken a photo of an interesting flower on a vacation in 2004. You can only really do this if each image you save has "tags" that a search tool could use to call every record that has that tag and present them to you in an array. Flickr uses this technique to let you search their database of records. It is analogous to structuring the records in a database with identifiers. You can only get them out if you put them in with all the tags.

You COULD just save every file with its image number as assigned by the camera. Assuming you never restart the numbers, each record then has a unique numeric identifier. You would use a database manager that knows a particular record identifier has "tags" assigned to it - and you would constantly tell the database manager all the tags that are assigned to each file you save as you save them. You would then ask the database manager to go retrieve every file that has a tag assigned to it - or a combination of tags. I'd bet that none of us does this - but it is the right way to save our files for efficient retrieval.

Since you likely do not have a database manager that can associate the tags with the file name, assuming you have given each file a fully-descriptive name, you would do a File Search for all files that have the target in the file name.

For instance, in a folder structure Vacations/2005/April/Cayman, you might have a file named flower.Hibiscus001.PEF. Clearly, you can look at ALL the pictures you took on the Cayman Trip by just going to the folder. But you can only call the flower pictures you took on that trip if you remember when you took them.

Or you might have a Flowers folder, and a file named flower.hibiscus001.vacation.2005.cayman.PEF. You can see all your flower pictures easily, but you might have divided up the other vacation pictures into other subject folders - you can't easily recall a complete record of that trip.

You would then do a File Search *.*.*.*.cayman.* and cayman would always have to be the fifth word in the string. If you did that, you could retrieve every picture you took on the cayman trip.

Unfortunately, to make this work, you have to add the tags to all your images as they go into your folders - and you already have 20,000 records saved.

I organized my folders so I could get what I wanted MOST of the time. I usually want to see everything that happened on a vacation that has my family in it, so they stay in the vacations folder, or find a photo of a friend. Landscapes are not an issue, because they are not spread out across lots of folders.

I chose to have a multi-layered folder tree and shorter file names, because it is easier to rename the files. You could go the other way, as you desire.

A few times I have just had to spend an afternoon browsing folders in Adobe Photoshop Album

[BBear]

Very "pro" (and maybe even complicated) method monochrome. The quick-search sounds very appealing though.

I keep coming back to MP3s as a reference because I had to re-think how to organize my collection and in the MP3 scene it is reasonably easy to keep the ID TAGS updated if you have the correct file names and the correct freeware (I use TagRename). Back in the days I had only filenames and some TAGS updated... whenever I wanted to find something I had to Windows Explorer > Search > keywords. It was painfully slow... But now with a Media Library, I start to type chars and the quick search work wonders... (I use the Winamp Media Library for my MP3s)


My MP3s work very similar to what you described, but there is a big difference: I automatically get the TAGS from the albums, CDDB or even from the filenames... only a few times I have to manually type the TAGS.


But in the photo world, I'd have to manually TAG all photos so that they would work with the quicksearch. I dont see this being practical... even on the new photos to come.



So for instance, if you go to Central Park and take 500 pics... you'd get back home and visualize EACH picture and tag it accordingly? One by one? I know you could possibly select a range and TAG them at once, but still you'd have to open each one and select them in a group...

I dont see how this works in an efficient way...

[monochrome]

Very "pro" (and maybe even complicated) method monochrome. The quick-search sounds very appealing though.

I keep coming back to MP3s as a reference because I had to re-think how to organize my collection and in the MP3 scene it is reasonably easy to keep the ID TAGS updated if you have the correct file names and the correct freeware (I use TagRename). Back in the days I had only filenames and some TAGS updated... whenever I wanted to find something I had to Windows Explorer > Search > keywords. It was painfully slow... But now with a Media Library, I start to type chars and the quick search work wonders... (I use the Winamp Media Library for my MP3s)


My MP3s work very similar to what you described, but there is a big difference: I automatically get the TAGS from the albums, CDDB or even from the filenames... only a few times I have to manually type the TAGS.


But in the photo world, I'd have to manually TAG all photos so that they would work with the quicksearch. I dont see this being practical... even on the new photos to come.



So for instance, if you go to Central Park and take 500 pics... you'd get back home and visualize EACH picture and tag it accordingly? One by one? I know you could possibly select a range and TAG them at once, but still you'd have to open each one and select them in a group...

I dont see how this works in an efficient way...

500 images from one outing? That's 21,000 images in 42 outings (about a year)!

You are absolutely right - it isn't practical at all to manually rename hundreds of files, but my original point was that the problem is in file naming for efficient retrieval. That's the nature of the beast.

As a practical matter we would at least view all 500 Central park images and select the best for further processing. Some batch renaming converter such as CK Rename could be used to give a general name to all of them, a specific name to selected files, a "counter" name - just about anything (pretty flexible gift-ware) - or the selected files could be manually renamed after processing. The rest could be archived in the same complex folders.

And that's why I have a complex folder tree, and thought about what I would most (and least) likely need to retrieve before I set up my convention. I don't take so many pictures that renaming selected files is an issue for me.

[Robert Barnett]

I do something like this.

Pentax K20D
2008
01 - January
01-22-2008
Finely Park
Original Images
Work Images

I choose to organize my images by shoot. So if I go out to the coast to say Bodega Bay and I drive up the cost to Jenner if there are more than just a few Jenner shots I would name the folder Bodega Bay and Jenner. If there were only a few then it would just be Bodega Bay.

There really is not perfect way to do this. You could go to extreme and create hundreds of directories and then try to make that work. But, if you have a flower macro with a bee do you file that under Macro, flower, flower macro, bug, bug macro, flower with bug, flower with bug macro and do on.

The other thing I have discovered in the case of Bodega Bay with Jenner if I geotag my images I have even less of an issue because I will always be able to see where I took them. With that in mind I just bought a really good GPS and RoboGeo (I shoot DNG and it will tag them).

Finally under the Working Images folder I will have some sub folders like Black and White, Edited, Collage, Composite, Infrared Uncoverted, etc.

I would love to find a way that was perfect, but there isn't. Where I think you can organize out the rear without a lot of trouble is in something like Photoshop Lightroom and you do it was collections or keywords, etc. I look at the directory structure as a very basic organization with the in-depth in the software.

Robert

[KjetilH]

Heh, after taking a couple of pictures, I found that the easiest and quickest way of dealing with the mess were to create a folder named "unsorted", put all new pictures there, and when convinient, run some automatic organizer software (AmoK Exif Sorter), and get them in folders.

If I want to, it isn't much hassle to give the folders a little more descriptive names, which I (eventually) will do.

Soo, it is sorted like:
\\2008\04\23\2008_04_23_15_14-001.JPG

[Mechan1k]

I am still trying to figure out how to "cleanly" look after my images.
I run mine in a similar fashion to monochrome ... not exact ... but it is close.
I seem to find everything when i go searching ... sometimes it takes me longer.
Thank goodness I don't have much spare time though to take a LOT of shots ... otherwise I'd be sponding more time searching.

[Gimbal]

I'm currently trying out a program called Imatch, and so far it seems to be a keeper. I download my photos a usual in to any folder, and the program makes a database with thumbnails (so the originals can be stored on an extern HD/DVD or whatever and I can still browse the thumbnails).

Each picture can be taged with several tags and subtags, and the pictures can then be browsed through folders (as usual) or by tags (very neat) such as nature/flowers and all thumbnails of flowers shows up.

see photools.com - Solutions for Image Management and Digital Photography. for more info.

[baw]

I started out with folders and the like also, but luckily I stumbled upon an image management program when i only had 5000 or so images to categorise.
Do yourself a favour and use some management program. It increases the value of your images enormously imo.
Just being able to view ALL images of your kid(s), pet(s) etc together is worth the little time it takes to tag your images.
I use filename YYYYMMDD_HHMM_####.ext. I dump all new images in one dir, untill it reaches DVD size. Then I create a new dir. (Main01, Main02 etc)
Edits go into Sub01, Sub02 etc. Makes for easy backup on DVD's.

My program of choice is IMatch, with it's unique hierarchical category structure.

[JMS]

I just wrote a program in GAMBAS that copies photos to the correct folder and tags them with the event. My structure looks like this:

/Home/jesse/Photos/2008/April/22-coin macros/

That's where my RAW files go, and then JPEGs and others go in the "Output" subfolder.

If anyone wants me to customize my program for them (it runs on Ubuntu, maybe others), just PM me with your folder structure, and I'll customize a copy for you. You'll have to install GAMBAS for it to run, though.