Can Keywords be turned off completely?

[Vincent9756]

It's not that I really want to turn off keywords. I want to stop IMatch from adding the additional XMP lr:Hierarchical tag. I tried setting different options but it's way too complicated. This is what's happening:

I have about 300,000 images. I've been using ACDSee to organize my images with keywords. My images have these tags:

-IPTC:Keywords
-XMP:Subject
-XMP-acdsee:Keywords

I indexed about 10,000 of these images with IMatch. No problem except that it wants to add yet another keyword tag. I wouldn't mind so much but exiftool is extremely slow. Took at least 20 minutes to update those 7,000 images. That's really only a problem because IMatch won't let me update more than 7,000-10,000 images at a time. I have 300,000 images. I have to sit there and attempt to select no more than 10,000 at-a-time, then wait for it to finish and keep repeating this process at least 30 more times.

I'd be somewhat OK with this if IMatch would just update them all at once. But, I really don't need/want this new tag so it's really a complete waste of my time. The tags I have work perfectly for windows desktop search, social media sites and flickr. I don't use adobe lightroom. I'm not a professional photographer. I just have very large image collection that I want to be able to search with IMatch and windows.

I did something crazy:

I made a copy of my images. I used exiftool to strip the standard keyword tags. I loaded some images into IMatch and I don't get any pencils on my thumbnails. Awesome. I then made a data driven category using the tag "-XMP-acdsee:Keywords". That works perfectly. I have my entire keyword hierarchy (it's read only; for searching; no problem) without that new lightroom tag. The only problem with this *crazy* approach is that I'd have to continually update this second copy of images when I add keywords to my original set.

I'd be fine without keywords in IMatch because I want to use it's categories. (I'd like to keep using ACDSee for keywords because it has features that enable to do it very quickly without mistakes. It just can't handle volume. I bought IMatch mostly so I could load and search my entire collection.)

--------------------------------
Once again:

Are there any options I could use to have IMatch completely ignore these tags:

-IPTC:Keywords
-XMP:Subject

or, is there a way to stop IMatch from writing/maintaining that new lightroom tag?

Thanks

[Mario]

All keyword processing in IMatch is based on hierarchical keywords, introduced by LR but now pretty much standard.
IMatch internally works only with XMP hierarchical keywords because they are the superset of the flat keywords used by the IPTC and XMP standard.

If you manipulate metadata (including keywords) IMatch marks the files as pending, to inform you that the metadata you see in IMatch is not the metadata you have in your files. If you write-back that data or not or when is up to you.

acdsee:keywords are a vendor-specific proprietary metadata tag. IMatch does not know anything about these keywords. You can use a metadata template to transform these proprietary keywords added by ACDSee to your files into the standard IPTC or XMP keywords, or hierarchical keywords. I don't know if ACDSee uses any kind of hierarchy in their keywords.

[Vincent9756]

I didn't manipulate any metadata with IMatch. I added a folder with 10,000 images to the database. IMatch immediately marked those images as pending for write-back. It said it wants to write/update the LR hierarchical tag.

As a test, I stripped the old, standard keyword tags (-IPTC:Keywords and -XMP:Subject) from some images, created a new database, and loaded those images again. IMatch didn't want to write the LR tag. (I kept the -XMP-acdsee:Keywords tag so it has nothing to do with acdsee).

I'm asking how can I get IMatch not to force me to write a new tag to 300,000 files that I haven't made any changes to. Are you saying, if I didn't use IMatch to change any keywords, IMatch shouldn't be trying to update my files? If so, there's a bug.

----
I'm not asking you to do anything with acdsee's keyword tag. FYI: It's the exact same format as the LR hierarchical tag, e.g. "location|outdoor|beach". It works fine with a data driven category.

[Mario]

IMatch by default imports existing XMP and legacy IPTC keywords into the hierarchical keywords. That's part of the Metadata Working Group compliance.
After that IMatch may have "more" or "other" keywords for your files, in the hierarchical keywords and also in the XMP and IPTC keywords, depending on your keyword settings under Edit > Preferences > Metadata and your thesaurus.

Importing metadata and creating the initial XMP record also may create other metadata required by XMP or the MWG, e.g. IPTC digests, documentID and instanceID, Metadata Create and Modify timestamps and so on. IMatch creates/writes these automatically, they must not be necessarily listed in the Write-back Pen tooltip.

If your problem is that IMatch now tells you that it has to write-back the files, go to Database > Tools > Clear Processing Queues and clear the write-back queue. Please note that this will mark all your files as "up-to-date" and IMatch then assumes that what's in the file does match what's in the database. No more yellow pens.

When IMatch re-imports a file because you change it in an external application, IMatch will again perform the import, metadata mapping (unless you manually delete the MWG compliance) and then pen will re-appear if the data in the database no longer matches the data in the file.

[Erik]

Stepping back for a moment...

The primary reason you don't want to write the hierarchical keywords is because of time and the limited queue availability for Exiftool processes in IMatch?

Have you thought about just slowly going through the metadata writeback process perhaps folder by folder, even if you just run the process for a few folders per day?

It also occurs to me that the ACDKeywords are essentially the same thing as the Hierarchical ones from Lightroom that are part of the standard.  You could run a command-line instance of EXIFTool outside of IMatch to map the ACDKeywords to the Lightroom Hierarchy through your complete collection, although with 300,000 images you could run into memory issues.  Imatch is still likely to want to write metadata for all your files at least once.  I find that is often the case when files are imported, but once you get the migration done, everything is fast and speedy.

Ultimately, I would think that anything you could do in ACDSee could be done in IM5.  It's thesaurus makes Keyword entry quite simple.  Running two different DAM programs could become problematic, specifically with IM5 rescanning for changes and always wanting to write metadata for the various keyword fields.  Running LR for the purpose of developing even has issues and I don't use it for DAM at all.  Using two different software for DAM purposes would seem like a disaster waiting to happen.

[Vincent9756]

Ultimately, I would think that anything you could do in ACDSee could be done in IM5.  It's thesaurus makes Keyword entry quite simple. 

The thesaurus is nice, but:

1. If I change a keyword in the @Keywords node, the thesaurus doesn't appear to get updated.

2. It takes a lot of screen space if you have a deep hierarchy. This means I have to constantly scroll order to see things and I found myself missing things and making mistakes.

This is why I use acdsee:

1. I define 6 or 7 keyword presets. Keyword presets are like the thesaurus except it gets automatically updated when I rename, add or delete keywords.

2. The presets are not displayed in any kind of tree format. However, you can define a keyword using the "|" separator so the preset keyword can be hierarchical.

3. The possible presets are displayed in a combo-box. I select a preset, "Location" for example. I then see my possibly choices in a small window below the combo-box. It requires very little screen space. Everything I need is constantly in view. That's very helpful in terms of preventing errors.

4. I select a group of images and simply click on the preset word. When an image has a preset keyword, the preset window display that item in boldface type. Everything I need stays in view (no scrolling). Very easy to see at-a-glance. I make far less mistakes with this approach.

Once I'm finished. I verify. I select the keywords in acdsee's keyword tree. I can select multiple nodes *and* specify that I want it to "OR" or "AND" the results. I select all the images and look at the preset panel. It's very easy to see if I've done things correctly. Then, I move on the to next preset with those same images. In this way, I don't forget any keywords that are important to me.

It appears that IMatch always wants to "OR" the results of a multiple selection. This makes it very difficult to verify what I've just done. In order to see any mistakes I might have made, I need to quickly switch between "AND" and "OR".

Another import thing (at least to me) is that, unfortunately, IMatch has 2 category panels: one for producing image results (left side of window) and one for adding images to the category (right side of window by default). ACDSee combines the functionality of both those panels into a single panel. That give me more screen space so that I don't have to keep flipping tabs or devote half my vertical space to an unnecessary panel/view. I can drag/drop, select and tag from the same panel. It's very helpful and useful. Then, I can have max space available on the right hand side of the window for my keyword presets.

So...the bottom line is that I find that I save time and make far less mistakes doing the keywords with acdsee.

The very bad thing about acdsee (at least for me) is that it can't handle many thousands of images. That's why I need IMatch, so I can access my entire collection. I'm trying to use the strengths of each app it order to get the job accurately in the least amount of time.

------------
Maybe I'll initially use exiftool to add the lr:hierarchical tag myself. I can do that over night. But, I may just do what Mario said and clear the pending cache because I don't really need this tag.

As far as I'm aware, Windows, Apple, Google, social media sites and flickr don't care about Adobe and this new keyword is not really a standard (Adobe ignored the MWG standard and invented it themselves). I *need* to be compliant with Windows, Google and flickr so I can also search outside of all the DAM software.

[Vincent9756]

IMatch by default imports existing XMP and legacy IPTC keywords into the hierarchical keywords. That's part of the Metadata Working Group compliance.

Is this statement made here:

http://blog.daminion.net/upcoming-posts/hierarchical-keywords-in-lightroom-be-careful/

wrong:

"So Lightroom uses a new unofficial Lightroom Scheme that stores keywords with the hierarchy preserved (inside the "lr:HierarchicalSubject" field) and the Dublin Core Scheme to store flattened keywords (inside the "dc:subject" field).

The MWG consortium offers a complex new way of storing hierarchical fields in XMP, but that's another story. It would appear that new specification creators don't see any point in providing backward compatibility for existing metadata specifications. I dislike MWG's new method of storing hierarchical keywords mainly because there is no way of determining exactly where the most up to date keywords are: in dc:subject, in "lr:HierarchicalSubject" or in "mwg-kw:Keywords".

Most photo software programs, for example GeoSetter, use "dc:subject" to store Keywords, which means that hierarchical info that you've entered in Lightroom will be invisible to these programs.

Although Lightroom developers probably couldn't care less about other programs (their competitors), this is one of the main reasons that some people prefer to stay with other image cataloging software programs."

It seems to me that Adobe did what acdsee did and stored their keywords in some tag that is not MWG complaint.

All I'm asking is that make support for any software that writes non-standard tags optional -- even acdsee. Because you already have enough options with metadata templates and data driven categories that let me access those non-standard tags.

[Mario]

Hierarchical keywords are a key concept in IMatch and cannot be turned off.
You can decide if you write metadata to files and when.
The MWG has published a fourth way to store (hierarchcial) keywords but it's overly complicated and bloated, and for the typical keyword hierarchy depths used by IMatch users it will blow up the metadata massively. I will look into this later. So far, the hierarchical keywords invented by LR several years ago are widely supported and the best way at this time.

If you want a new feature to suppress write-back for hierarchical keywords, and all the features and functionality which depends on this, please add a feature request for later consideration.

I will also move your post out of the Tutorial board to "General Discussion" because this board is for giving other user tips, not for asking questions.

[Erik]

To answer the questions Vincent made,

Lightroom maps its Hierarchical keywords to the regular subject field just as IM5 does.  These software use the Hierarchical fields as an essentially BASE keywords field and then internally maps those values to the remaining Subject and IPTC Keywords to keep them in-sync.  Of course, you have to process pending writebacks for the mapping to take place.  It's likely that ACDSee does something similar as most online sources do not read hierarchical tags that I am aware of and count on the flat IPTC Keywords or XMP DC:Subject ones for tagging.

As for the blog you linked to... the key for IM5 and LR is that they base themselves on the specific hierarchical keywords tag and then keep the subject and keywords mapped and sync'd so they all "match" and no one has to question which keyword field is the right one (they all should be).  And, if you use a program like GeoSetter, which is not aware of the LR Hierarchy, IM and LR will take a new found and different keyword in the Subject field and map it to the Hierarchical Field, so that all keywords are still in sync. 

Ultimately, you have to do what works for you.

[Vincent9756]

Thanks Erik.

For right now I'm going to do what Mario suggested and clear the processing queue. For my purposes, I only need flat keywords in the standard XMP and IPTC tags. It shouldn't be an issue for me until Windows, OneDrive, flickr, Apple & Android phones support hierarchical keywords. They may never do that. If they do, they may not even use the lightroom tag. So, for right now, updating 300,000 images with that info could be a complete waste of time.

[sinus]

As far as I'm aware, Windows, Apple, Google, social media sites and flickr don't care about Adobe and this new keyword is not really a standard (Adobe ignored the MWG standard and invented it themselves). I *need* to be compliant with Windows, Google and flickr so I can also search outside of all the DAM software.

Well, I think, at least Adobe was and still is very important for the photographic industry. Without Adobe we would not have a LOT of good things for "our" industry (Photoshop as an example), but of course, where light is, there is also some shadows.

But Windows, Google, social media sites and sure not flickr has in the "professional" photo-world the power of Adobe.
At least in the professional world of pictures it is wise, to follow the Adobe-line, if you like or not.
But of course, if YOUR workflow and YOUR clients are happy with your system, that you do, so all is fine.

[Vincent9756]

As far as I'm aware, Windows, Apple, Google, social media sites and flickr don't care about Adobe and this new keyword is not really a standard (Adobe ignored the MWG standard and invented it themselves). I *need* to be compliant with Windows, Google and flickr so I can also search outside of all the DAM software.

Well, I think, at least Adobe was and still is very important for the photographic industry. Without Adobe we would not have a LOT of good things for "our" industry (Photoshop as an example), but of course, where light is, there is also some shadows.

But Windows, Google, social media sites and sure not flickr has in the "professional" photo-world the power of Adobe.
At least in the professional world of pictures it is wise, to follow the Adobe-line, if you like or not.
But of course, if YOUR workflow and YOUR clients are happy with your system, that you do, so all is fine.

I'm sorry I should have said something: I'm not a professional photographer. I completely forgot that IMatch was designed mainly for professionals. It must have been very confusing for some people to wonder why I don't care about Adobe. And now, I can understand why IMatch wants to write that lightroom tag by default.

I do take pictures but I definitely don't have the same needs as an actual photographer would. I just have a very large image collection. ACDSee would work great for me if it could handle volume.

Once again, I apologize.

[Richard]

There is no need for you to apologize. Not all IMatch users are professionals and not all use Adobe. It has been about 15 years since I bought Photoshop the last time.

[Mario]

Even for 'amateur' and corporate users it is important to produce and maintain metadata in their files according go the common standards in use today. That means XMP, that means "Have an eye on what Adobe does, because all other do too" and the Metadata Working Group.

Whether you process files with a RAW converter of some sorts, you upload files to a private web gallery, you upload to Facebook, Flickr, Tumblr or one of the other on-line galleries and social networks, if they are able to process metadata (instead of just stripping it from your files), they usually support the legacy IPTC metadata or the more future proof XMP.

For commercial and professional workflows in the publishing and photographic industry, Adobe makes the rules (to a large extent). Everything that deals with PDF too.

The way IMatch handles metadata, and the bazillion of sometimes strange and confusing options, exist to produce metadata as good and 'rich' as possible, and to make the metadata you work with compatible or 'adaptable' to the majority of applications and services in use today.

Metadata is a mess. Metadata is a moving target. Application and camera vendors seem to come up with new ways to (ab)use it all the time. Look at this and other forums to see how much of a burden proprietary metadata can become after a couple of years. It's best to keep it simple and stick to whatever 'standards' we have. And IMatch allows you to do this, with a minimum of effort.

[lenmerkel]

Metadata is a mess. Metadata is a moving target. Application and camera vendors seem to come up with new ways to (ab)use it all the time. Look at this and other forums to see how much of a burden proprietary metadata can become after a couple of years. It's best to keep it simple and stick to whatever 'standards' we have. And IMatch allows you to do this, with a minimum of effort.

One more quote for the IMatch web site, I think . . . . .