How to get started with categories (beyond the 5 W's)

[ubacher]

How to think about Categories

Most people will first think of the 5 W when organizing photos. (Who, Where, When, What, Why)

I suggest you think of it as forming groups.
Note that you use categories to find/select images but also, equally important,
to exclude images. Imatch also lets you select files by combinations AND OR or NOT.
You want to be able to reduce the number of images through which you have to manually
scan to a manageable few hundred.

1. Grouped by what is to be seen IN the photo. I call this CONTENT.

2. Group by time
  Event
  Fotoshoot
    (these photos are usually all in one folder per event)
 
3. Group by Place
  Location
  ( I also have a location called HouseAndGarden. And you could have something like STUDIO or At Customer)
 
4. Group by Activity
  - Sport
  - Theater
  - Hiking
  - Hobby

5. Cataloging administrative features
Where does the photo come from (if not from your camera)
- scan
- photographer
- Internet

6. Why was the photo taken
- customer
- photo project

7.  Type of photo
- Constituent (i.e. used to make a final image)
   - Part of Panorama
   - Part of Sequence
   - HDR Series
   - Focus Stacking Series
- Assembled Photo
   - Panorama
   - HDR
   - Stacked Focus
   
- Stereo (Her you also distinguish sources from assembled (viewable) images)

- Example
- Documentation (photos or text files...)
- Associated Administrative Docs
- Working Material (Backgrounds and such, files used in graphic work )

8. Production Info (for files derived from originals)
- Finished Book
- Book Page
- Collage
- Poster
- Matted or Framed
- Print ready
 
9. Where was the photo used
  You may want to know that a photo was already used somewhere.
  This could be merged with Production Info or with Manual Grouping
 
10. Manual Grouping
  another way of marking photos which you manually selected/used for a purpose:
- Slide show
- Photobook
- posted on ...
- submitted to ..
- sold to ...
11. A category tree for more or less temporary use - I call it WORKING.
      Keep temporarily chosen categories in this tree. I find that I tend to use
      EVENTS now a lot for this purpose ( keeping a set of files together under a title while
      I am working on them)
     

Should every file have at least one category? I say yes - else such a file will never show up
in category based search.

What categories should at least be present or - in other words - have I forgotten to assign an important
category?

For this we use a filter preset which shows us files without needed categories.
Example:
You want the category specifying the Source-or-photographer for each file.
You set up a category tree "Source-or-photographer" containing
- Me
- Scan
- Internet
- other photographer
- unknown
- Irrelevant

You can then simply filter the files for the absence of any category of this tree. (NOT Source-or-photographer)

Each of the groups you use should use a hierarchical category tree.

Final thoughts:

1. Think when selecting categories: "How would I search for this photo/file"

2. Start simple: I suggest to always have a category OTHER in your trees. Later, when you
find what category would be useful, you create that category and reassign from OTHER.

3. Don't know how to categorize a file? I have a category "DONT KNOW" for that case.
As you learn you can review this category and assign the "correct" category.

4. Files which contain GPS location data need not have a LOCATION category - I find, however, that I still
assign a broad LOCATION category to such files.

5. There is no need to use categories for selecting by dates. Imatch lets you select by date.
(There is however a problem with scanned files unless you manually assign the correct (but probably unknown) date.)

6. You may want to assign a category not for selecting but for excluding. Constituent photos are an example:
You will want to see assembled panoramas and not the individual shots. Using STACKING can circumvent the need
for such categories: Just put the constituent files on a stack with the assembled file visible on top.

7. You can use categories like BOOKMARKS to temporarily mark some files. Locating the category on
the FAVORITES panel is useful for this. Setting a color for the category is then appropriate.
Or you can save BOOKMARKED files permanently by assigning a category.

8. Calculated categories are a special topic I ignore here for sake of simplicity. Examples are the
imatch workflow and standard categories.
You may need some when a filter preset is too complicated to set up or use.

9. EVENTS and PEOPLE are new additions to Imatch which can replace using categories for these purposes.

Please note that Imatch is very powerful and getting started with it can be overwhelming. Start simple!

[Mario]

Thanks for sharing your category organization 

[sinus]

ubacher,
very interesting, thanks!

Should give specialy new users good hints.

I have my workflow and system since years, but if I see a good idea, I think about include it in my workflow.

E.g. categories using as bookmarks, well, in combination with a favourites a very good idea, I do this since some years, because (for me) 1 bookmark is simply not enough for me and the other IMatch-collections like pins, dots .. are used elsewhere.
I have always some temporarly fies to bookmark, hence I created bookmarks with categories (you can there even give names like "eyes retousched" or "send for Angie for birthday"...

Additionally I use also some Apps like dashboar or pinboard for this. 

[Jingo]

This is really great information... and thx for sharing it!  I find photo organization to be such a fascinating subject... I have relied upon simple keywords for over 20 years... but really do think I should investigate other things like categories more.  With a simple keyword search, I can locate pretty much anything... but more options are always interesting!

[Stefanjan]

As a newbie, this is a really interesting post.  Up to switching to imatch I have been using a MESS of flat keywords in a very undisciplined way.

I have begun trying to build a hierachial thesaurus in imatch and move existing keywords into this structure but it's hard work.

Do you still use hierachial keywords in conjunction with categories.

I'm wondering whether I need hierachial keywords and instead use categories along the lines you suggest alongside data driven when and where and title / description.

I also have a file structure:
YYYY
  YYYY-MM-DD :  description

[Mario]

The reason to use hierarchical keywords: they are stored in your file, together with the other metadata. Which makes them accessible to other applications.

This is why I recommend hierarchical keywords as the basis.
And regular IMatch categories for automatic organization using data-driven categories and formula-based categories. Where needed.

The powerful @Keywords Category category mirrors the keywords in your files, and makes things like swapping keywords or moving files along your keyword hierarchy very easy and fast. You probably use it for your task.

Categories can do many things simple keywords cannot. From category formulas to data-driven categories.
Categories are super-fast, and don't require write-backs when you change the categories of files (except for @Keywords, of course).

IMatch gives you the choice, and your workflow and requirements decide what works best for you.

[Stefanjan]

This is why I recommend hierarchical keywords as the basis.
And regular IMatch categories for automatic organization using data-driven categories and formula-based categories. Where needed.

So where Ubacher suggests category for "Type of photo", I envisaged as an example a tab under favourites "Type of photo" with sub categories RAW, Final JPEg Photolab, Final JPEG ON1, Silver Efex, Comptition entry. Print A3 etc. Then I could easily click on the correct category.

Are you saying that it would be better to assign hierarchical keywords doing the same thing and have a data driven category?

I can see that it would be just as easy to drag the hierarchical keywords on to a tab on the favourites panel and have a data driven category

[JohnZeman]

I can see that it would be just as easy to drag the hierarchical keywords on to a tab on the favourites panel and have a data driven category

I'm not aware of a way to drag hierarchal keywords onto a favorites panel however you can drag @Keywords categories onto a favorites panel which gives you the same thing.

[Stefanjan]

I'm not aware of a way to drag hierarchal keywords onto a favorites panel however you can drag @Keywords categories onto a favorites panel which gives you the same thing.

Sorry that's what I meant

[ubacher]

I am not using keywords, only categories.
The reason for hierarchical categories (or keywords) is organizational. It is easy to find a category
among all of them this way.

About keywords: I discourage the use of keywords, at least for beginners. When a customer demands keywords
you will have to use them and you will probably be given a catalog of valid keywords.
Switching categories is easy and fast, changing keywords results in having to re-write all associated files - a slow operation.

Using keywords because they are stored in each file - and thus independent of Imatch seems a good argument.
If I ever have to change to a different photo management program I will just convert my categories to keywords
first.

NOTE: You can also use both - keywords and categories!!!

[Jingo]

Using keywords because they are stored in each file - and thus independent of Imatch seems a good argument.
If I ever have to change to a different photo management program I will just convert my categories to keywords
first.

The main reason I use Keywords AND store them in my files is that I use a variety of software that utilize this information which makes it easy to always have it available even directly from the OS.  For example, just the other day I need to locate an image... sure - I could have fired up Imatch and done a keyword search - but I knew the date range for the image so a quick search in XYPlorer (which I always have open on a 2nd monitor) and I found the image by keyword/date range in 20 seconds.

The flexibility that IMatch offers is the feature I promote most often.... have it your way... or many other ways!

[Stefanjan]

So many choices, it's confusing for a beginner but I shall persevere and read and reread the help and follow threads in this community.

How people use imatch for their workflow is especially interesting as it gives me new ideas.

[sinus]

So many choices, it's confusing for a beginner but I shall persevere and read and reread the help and follow threads in this community.

How people use imatch for their workflow is especially interesting as it gives me new ideas.

I now have over 300'000 files in my IMatch database (and 20'000 in a second database for 3D files).
If I were to start again, I would take 200-300 files that are typical for me.

And then with these 300 files (of course all copies, not yet the originals) I would build up a workflow.

I would really "play around" for at least a month, much better two or three months, trying this and that. I would not succumb to the temptation of wanting to start "for real" quickly.
What are 3 months for all the time I will then spend working with real files and probably for a very long time?!

It starts with the folder structure, the file names, do I make versions, if so, which endings, how to name them?
Am I likely to work with other programmes?
Then I would take a close look at what is the effective difference between categories and keywords? Do I really need keywords? Which ones?

And it is very important to stay consistent. For example, if I always name the files with the same structure, then I can certainly change things again later, with IMatch board tools or an app or whatever. (Changes will almost certainly come, because it is impossible to foresee everything and one's own ideas will also change sometimes).

The categories also need to be well thought out, what makes sense, what doesn't.
And I also think it's important to always keep the workflow the same. That also helps later when changes are needed.

And start simple, not too complicated, that's also a good way.
Either way, with IMatch you will have a very, very good basis for managing your files optimally, I am sure.

Your idea to read everything well first, to inform yourself, to think about it, is very good in any case. I wish you much success!
And if you have problems or questions, you have many nice people in this forum who can and want to help :-)

[Mees Dekker]

I only use categories, like Ubacher, for two main reasons.

As Sinus pointed out: it is impossible to foresee everything and your own ideas may change as well. If your workflow, ideas or anything are depending on keywords, every change has to be applied to all associated files and a write-back becomes necessary. Especially when these files are on some network drive, this is a rather slow process. Categories on the other hand are only in the database and therefore far easier (and faster) to reorganise, rename, etc.

Only when a client ask for specific keywords being written to the file itself, I convert the necessary categories to keywords and write them back into the files (and send the files off to the client). The only keywords I always apply to any file (on ingesting into IMatch), is my copyright info.

Just my two cents on this item.

[JohnZeman]

If I were to start again, I would take 200-300 files that are typical for me.

And then with these 300 files (of course all copies, not yet the originals) I would build up a workflow.

I would really "play around" for at least a month, much better two or three months, trying this and that. I would not succumb to the temptation of wanting to start "for real" quickly.
What are 3 months for all the time I will then spend working with real files and probably for a very long time?!

I agree with what Markus (sinus) recommended.  Until you've settled upon a method you're happy with to keep your files organized the way you want, start with a small test database with about 200 or so test files and do your experiments on those.

15 years ago when I first started using IMatch I think I trashed my database and recreated it all over again almost every day at first as I slowly figured out how I wanted to do things.

Starting all over again is easy with 200 test images, doing so with 200,000 images would be a massive frustrating project that would take an eternity to do.

[Stefanjan]

Many thanks for the wise advice.

I don't have clients, purely an amateur photographer. Taking family photos and photos for the various organisations I belong and especially for the camera club I belong to.

People, places and events are all important to me. I envisage using face recognition for the first, data recognition for places. Events not sure whether I use categories or events. Categories probably as I probably need these to be more than two levels.

I also name my folders with the date YYYY-mm-dd ¦ folder description so that's another way to search.

Up to now I haven't had a decent system for identifying files within my workflow: Original RAW or Original JPG, Edited in DXO Photolab, Edited in ON1, NIK processed TIF, Full size processed JPG, 1600 Competition entry, Website image, focus stack, exposure stacking etc.

I think it would very useful to me to have a system which allows me to open a version stack and at a glance identify the status of each file in the stack. Sounds like the recommendation is that I use categories rather than keywords for above workflow.

[JohnZeman]

Up to now I haven't had a decent system for identifying files within my workflow: Original RAW or Original JPG, Edited in DXO Photolab, Edited in ON1, NIK processed TIF, Full size processed JPG, 1600 Competition entry, Website image, focus stack, exposure stacking etc.

Along those lines I use the Instructions tag

Code Select Expand

{File.MD.XMP::photoshop\Instructions\Instructions\0}

To record my image processing information.

I also renamed the Instructions tag to

Last Processed & Exported

Then I enter the program(s) I last used to process each individual photo.

That information by itself for each image isn't much use to me but when I use it in a data-driven category I can see in a glance which photos in my database were processed by which program plus I can see in a glance which images haven't been reprocessed in a long time.

[Stefanjan]

Thanks JohnZeman, that sounds like an option I was not aware of but could well do the trick. I'll need to take closer look and work out how that's done.

[lbo]

The main reason I use Keywords AND store them in my files is that I use a variety of software that utilize this information which makes it easy to always have it available even directly from the OS.  For example, just the other day I need to locate an image... sure - I could have fired up Imatch and done a keyword search - but I knew the date range for the image so a quick search in XYPlorer (which I always have open on a 2nd monitor) and I found the image by keyword/date range in 20 seconds.

getting somewhat off-topic, but I'm interested in XYPlorer: Can I search keywords quickly with it? It seems that XYPlorer doesn't use the Windows Search index so it had to read all files to get the information. Am I wrong?

I tested XYPlorer some time ago when Microsoft crippled Windows Search starting with 19H2 and I'm still looking for a better access to the Windows search index.

[Stefanjan]

Hi John Zeman, would you mind giving me some hints as I'm struggling to put this all together.

Along those lines I use the Instructions tag

So do you create a custom metadata panel and rename the Instructions Tag by setting the caption?

The Instructions tag has a padlock preventing from inputting data. Where should I unlock this tag?

When you input Last Processed & Exported, is this freetext every time or are you able to link the tag to the Thesaurus or some other dropdown?

Code Select Expand

{File.MD.XMP::photoshop\Instructions\Instructions\0}

Where do you input this code?

I have tried without success to work this out using the help files. I'd very much appreciate it if could put me on the right track.

[Mario]

The XMP instructions tag is part of the Default Metadata Panel layout shipped with IMatch. You don't need to do anything, just use it.
It is not locked, either.



This tooltip shows when you place the mouse cursor at the little ? icons in front of tags in the MD panel (when a description was provided, which I did for most relevant tags in the Default layout).

[JohnZeman]

I have tried without success to work this out using the help files. I'd very much appreciate it if could put me on the right track.

(I see Mario answered the question while I was typing out my answer.  Well since I've typed it out and made some screenshots to go with it I'll go ahead and post this anyway.)

In the metadata panel click the icon to edit layouts.

The attached screenshots should show you the general way to do it, if you still have questions then by all means ask them.

[JohnZeman]

When you input Last Processed & Exported, is this freetext every time or are you able to link the tag to the Thesaurus or some other dropdown?

I just realized I forgot to answer this question.

I use a Metadata Template to write to the Instructions (Last processed & exported) tag.

The attached screenshot shows the metadata template I use when the processing was done by PhotoLab 4 and Affinity Photo 1.9.

As to your other question about the padlock icon, in essence the padlock means you have chosen a tag that is not meant to written to.  The Instructions tag I use is the writeable XMP Instructions tag.

[Stefanjan]

The XMP instructions tag is part of the Default Metadata Panel layout shipped with IMatch. You don't need to do anything, just use it.
It is not locked, either.

It was locked, see attached. But is now unlocked. Strange!

[Mario]

The lock may be displayed under several conditions: Database read-only, file is currently being read or written etc.