Strategy for version/category management

[Photon]

Using v3.6.0.118 for serious work and testing v5.0.124 for beta evalutation I have general questions about best strategy with versions.
For all my image files managed with IM v3.6 I have typically between 2 and 5 files per image.
These include often one RAW file, always two JPG files and sometimes some processed JPG files.
My requirements and questions:

• I keep all version files together in the same directory.
  The first part of file name is always identical, but the trailing text is different (quite easy for version detection).
• I need to search for categories, but normally I like to find the RAWs only an often only a certain JPG version.
  I know that this is possible with filtering, but my current issue with v3.6 is, that I would need ALL image files with perfectly synchronized categories.
• Do I have to synchronize categories always between all versions? What if a processed jpg files is added later?
  Currently I have not synchronized categories among all version files. How can I achieve this?
  Sometimes only raw is categorized, sometimes only one jpg and sometimes all version files.
  Shall I try to clean this up already in v3.6 or better wait for final v5.0?
• I do not need the category data updated in the image files. Proper storage in the IM database would be sufficient.
  Optional export to XMP sidecars would be nice to have.
• When searching for categories (e.g. a certain person name) I don't want a big counter result for all versioned files, but I like to have a counter for real version sets.
  That means if i have 250 different images of one person, I don't want to have a total counter of 1250 search hits for this person (when e.g. each version set includes 5 files)
• Sometimes my master is RAW file, but quite often just a JPG or a PNG or a video file with slave JPG preview. Can this moving target be solved with a versioning rule in v5.0?
• How to find the best strategy, which meta data shall be synchronized and which not?
  Is it recommended/possible to synchronize all meta data (like ratings, colors, categories, pins, ...) automatically?

Do you understand my principle intention and lack of knowledge?
I followed a similar discussion in https://www.photools.com/community/index.php?topic=191.msg916#msg916.
I am sure IM with current or beta version can do it somehow, but I need a recommendation for best strategy.
What can be prepared already in v3.6, for what is it better to wait for final 5.0?
Thanks for any advice, discussion, ...

Martin

[BenAW]

The versioning setup I'm using is pretty simple:
one dir with all masters and one dir with all versions. Since you already have one dir for all versions, this is pretty similar.
(actually I have a few subdirs to keep the number of images per dir in check)
I only categorize the masters, so I only see the counts per category etc for each version SET.
Versions have no categories assigned to them.

The versioning rules can be restricted to search in only one dir, this speeds things up considerably.
To establish the master / version relations I just select all images in the master dir, and let IM search for versions in the versions dir.
I do not synchronise any metadata at all.
If needed you can do an occasional copy from master the a version to copy eg. the IPTC/XMP data for distribution to relatives.

[Photon]

If only the master (e.g. RAW file) is categorized, how can you find all versions (e.g. all related JPG files) based on categories?
For example I need to generate regularly collections of images for a calendar print, a media show, an online gallery with specific JPG versions.
This can be the full size JPG or a cropped/resized JPG or one edited JPG version.
I like to search/find these files in IMatch and usually drag&drop them to other windows applications for further processing.

I am wondering whether automatic full propagation of categories, labels, stars, pins, ... is the only way?

May be I need an option in IMatch to display all version files, even when a search for category, label, star, ... gives just one file of a version set?
Or can additional or alternative stacking solve this task better?

Martin

[BenAW]

If only the master (e.g. RAW file) is categorized, how can you find all versions (e.g. all related JPG files) based on categories?

Select the master(s) you need, right click and select Versions > Show all Versions (F4, S)

[sinus]

Using v3.6.0.118 for serious work and testing v5.0.124 for beta evalutation I have general questions about best strategy with versions.
For all my image files managed with IM v3.6 I have typically between 2 and 5 files per image.
These include often one RAW file, always two JPG files and sometimes some processed JPG files.
My requirements and questions:

• I keep all version files together in the same directory.
  Martin
  

Hi Martin,
an interesting thread. Well, finally there are users out there, who tries to find a good workflow for they own system.
There are hundreds of workflows out there, but after all, I think, finally for the most of users, we have "only" 2-10 (or so) workflows, and if these are public, a lot of users could pick up an equal workflow to fit the own wishes.
(workflow like nef-jpg or only jpgs and so on).

Yep, me too, I have the same system like you.
In the past, I found this better, because I have also a filenaming-system like you, hence the raws and the versions are neatly placed together, even in Windows explorer. In IM it does not really matter, I guess, if you have one folder or a subfolder.

So, here I agree with you. 

[sinus]

Using v3.6.0.118 for serious work and testing v5.0.124 for beta evalutation I have general questions about best strategy with versions.
For all my image files managed with IM v3.6 I have typically between 2 and 5 files per image.
These include often one RAW file, always two JPG files and sometimes some processed JPG files.
My requirements and questions:

• Do I have to synchronize categories always between all versions? What if a processed jpg files is added later?
  Currently I have not synchronized categories among all version files. How can I achieve this?
  Sometimes only raw is categorized, sometimes only one jpg and sometimes all version files.
  Shall I try to clean this up already in v3.6 or better wait for final v5.0?

Martin

Well, I think it is not necessary to synchronize all images (like Ben, if I understood it correct, does).

But I want to have synchronized them together. I have a Master (raw) and create a jpg, psd or tif. Then I propagate the new image (version) with the same cats like the Master. I like to have a cat, say "Paula Winglock" for all images, not only to the master (I think, this is a kind of personal feeling, I am sure, it is not really necessary).

But, you could have some cats, what are special, and you want not to synchronize. Like, you could have a cat called "Cutted" or "reddish" and the master is not cutted or red, but only the version, because you cutted the jpg very hard and/or reddish. In such cases you must exclude the propagation of these cats.

It is also question with other stuff, say stars (rating). MAYBE a master is not that good to have 4 stars, but you edited the version such good and special, that only this version should get 4 stars.
Now you must decide, should the master also get 4 stars (since without master no 4-stars-version) or not?!

I think, we must decide, what we want and stay with this, and do it always in the same manner. But really necessary in IM5, I guess, it is not, this marvellous program will always give us possibilities to find our images.

[/list]

[sinus]

Using v3.6.0.118 for serious work and testing v5.0.124 for beta evalutation I have general questions about best strategy with versions.
For all my image files managed with IM v3.6 I have typically between 2 and 5 files per image.
These include often one RAW file, always two JPG files and sometimes some processed JPG files.
My requirements and questions:

What can be prepared already in v3.6, for what is it better to wait for final 5.0?

Thanks for any advice, discussion, ...

Martin

A good question. In my case, all what I can prepare in 3.6 I will do (because I know this version very good). But some stuff (in my case rating) I will do in the upcoming new and blinking IM5. 

In IM3 I did not real versions (except with help of cats), but I did a good naming-system, like you, so I will do the versions, I think, in the new IM5.

[Ferdinand]

Using v3.6.0.118 for serious work and testing v5.0.124 for beta evalutation I have general questions about best strategy with versions.
For all my image files managed with IM v3.6 I have typically between 2 and 5 files per image.
These include often one RAW file, always two JPG files and sometimes some processed JPG files.
My requirements and questions:

   
• I keep all version files together in the same directory.
  The first part of file name is always identical, but the trailing text is different (quite easy for version detection).
• I need to search for categories, but normally I like to find the RAWs only an often only a certain JPG version.
  I know that this is possible with filtering, but my current issue with v3.6 is, that I would need ALL image files with perfectly synchronized categories.
• Do I have to synchronize categories always between all versions? What if a processed jpg files is added later?
  Currently I have not synchronized categories among all version files. How can I achieve this?
  Sometimes only raw is categorized, sometimes only one jpg and sometimes all version files.
  Shall I try to clean this up already in v3.6 or better wait for final v5.0?
• I do not need the category data updated in the image files. Proper storage in the IM database would be sufficient.
  Optional export to XMP sidecars would be nice to have.
• When searching for categories (e.g. a certain person name) I don't want a big counter result for all versioned files, but I like to have a counter for real version sets.
  That means if i have 250 different images of one person, I don't want to have a total counter of 1250 search hits for this person (when e.g. each version set includes 5 files)
• Sometimes my master is RAW file, but quite often just a JPG or a PNG or a video file with slave JPG preview. Can this moving target be solved with a versioning rule in v5.0?
• How to find the best strategy, which meta data shall be synchronized and which not?
  Is it recommended/possible to synchronize all meta data (like ratings, colors, categories, pins, ...) automatically?[/l]

There are a lot of questions here Martin.  You will get quite different answers from some users, esp Ben & I, who have quite different workflows.  I will not attempt to answer all your questions at once. 

If you want to synchronise categories between the master and versions in IMatch 3.6, then I have an Image Synchronisation Script on the IMatch 3.6 Wiki that can do this.  Of course it could also be done in IMatch 5 using native versioning.

Should you do this?  That's up to you and what sort of workflow you want.  I do because I find it convenient.  As I understand it, Ben doesn't because he doesn't and he doesn't see the need.  If you want a total counter of 250 search hits for a person rather than 1250 then you might not either.  Because I do synchronise the categories, I would have 1250.  But if I want to see how many master files I have I can filter on masters, and if I want to see how many processed finals I have I can filter on finals.  So it depends on what sort of workflow you want.

You asked how you can find only a certain JPG version when searching for categories.  I guess you need a way to identify that JPG version.  So you will need something to filter on.  I use categories for this, and I also have a character in the version file name to indicate what sort if version it is.  For example, if I want the web version I filter on the web version category, or I could search for the relevant part of the file name.

The problem with your question is that there are many possible approaches to how to manage versions, and they all have their advocates.  My suggestion would be to work out how you want to manage your files in IMatch 5 and prepare for that in IMatch 3.6

[ubacher]

You will want to add at least one of the collections to propagate automatically. This way you can mark your selected (master) files with (let's say) a RED pin and later get all masters and  versions together by displaying all files with RED pins. Since you are adding the pins to master files
they pins will propagate automatically.

I think this is easier than the Show all Versions suggested by BenAW since this way you won't see other, unselected files.

Categorizing the Masters is the way to go. Since propagation is not automatic when a version is added later you will not want to
rely on the version having got the category of the master (Since you might have forgotten to initiate a propagation after adding the version).

Most versatile would be to have the categories synchronized between versions an masters. Ferdinand's Reverse Copy Ratings script
can be modified for this. (I did this crudely for my use)

I too, frequently select images for books, calendars, shows etc. and must then keep track of the final "printable" images. I have not found a
satisfactory way to do this. I tried various ways in IM3 and I hope to soon work out something elegant with all the facilities of IM5.

[BenAW]

I think this is easier than the Show all Versions suggested by BenAW since this way you won't see other, unselected files.

Curious to hear in what way you think your method is easier than just selecting F4,S, and in what way unselected files would show?

[sinus]

You will want to add at least one of the collections to propagate automatically. This way you can mark your selected (master) files with (let's say) a RED pin and later get all masters and  versions together by displaying all files with RED pins. Since you are adding the pins to master files
they pins will propagate automatically.

I think this is easier than the Show all Versions suggested by BenAW since this way you won't see other, unselected files.

To be honest, I do not know, what you exactly mean!? Select red pins for masters could be a good idea, but I cannot see, that it can be easier than using the implemented versions-stuff from IM5, like Ben has suggested.

[Photon]

If only the master (e.g. RAW file) is categorized, how can you find all versions (e.g. all related JPG files) based on categories?

Select the master(s) you need, right click and select Versions > Show all Versions (F4, S)

Thank you BenAW.
But this method does not seem to be really convenient and efficient for fast search and to realize a configurable overview.

Let me use an example:

• You have 12 images for person "Benjamin", 6 of them with is a a set of 5 files (1 x raw, 4 x jpg) and 6 of them without versions (1x jpg).
  In total you have 6 x 5 + 6 = 36 files. Categorized with keyword "Benjamin" are only the 6 master files and the 6 single files.
• You search and find the 12 "Benjamin" images in the category view.
  But you have no chance to view all 36 files together, isn't it?
• The command (F4,S) to show all versions is appropriate for one or a few images, but not for a medium or large number of images.
  This because the resulting search result window is not as clearly arranged as the category view.
  In this search result window, there are also missing the 6 images (with category "Benjamin") which have no version set.

Reading all the previous good comments from other IM users I think we end with the requirement to propagate some relevant categories to all versioned files. If for example the keyword "Benjamin" is propagated automatically to all versions, all 36 images will be visible in the category view. With intelligent filter options this number could be reduced (e.g. hiding raw file extensions or hiding certain file name endings). Good if you have already a consistent file name strategy for versions.

f you want to synchronise categories between the master and versions ... Should you do this?  That's up to you and what sort of workflow you want.  I do because I find it convenient.  As I understand it, Ben doesn't because he doesn't and he doesn't see the need.  If you want a total counter of 250 search hits for a person rather than 1250 then you might not either.  Because I do synchronise the categories, I would have 1250.  But if I want to see how many master files I have I can filter on masters, and if I want to see how many processed finals I have I can filter on finals.  So it depends on what sort of workflow you want. ...

Thank you Ferdidnand for this confirmation. This issue with keyword propagation and resulting "wrong" counters is exactly my concern. Correct counters can be of course a little bit helpful. May be we need a new category feature for v5.0 which allows optional display of all search result related versioned or stacked images regardless of their category? With this option it wouldn't matter whether you propagate or not propagate some categories.

Martin

[Photon]

You will want to add at least one of the collections to propagate automatically. This way you can mark your selected (master) files with (let's say) a RED pin and later get all masters and  versions together by displaying all files with RED pins. Since you are adding the pins to master files
they pins will propagate automatically. I think this is easier than the Show all Versions suggested by BenAW since this way you won't see other, unselected files.

To be honest, I do not know, what you exactly mean!? Select red pins for masters could be a good idea, but I cannot see, that it can be easier than using the implemented versions-stuff from IM5, like Ben has suggested.

If hope I understand the idea of Ubacher correct.
The temporary use of RED pins instead of constant keyword propagation sounds brilliant for me.

Assume you search for keyword "Benjamin" in all your images. You get 12 search results for only the 6 master RAW files and 6 individual JPG files.
Then by ticking the RED pins for all 12 search results this pin is propagated automatically to all versions (specific IMatch file relation setting).
Then you just have to search for all files with a RED pin and you get all (1+4)x6 + 6 = 36 images (including all versioned files from the 6 RAW files).
After all you have to untick the RED pin, because this pin is reserved for future temporary search. Did I understand the idea correct?

Martin

[sinus]

If hope I understand the idea of Ubacher correct.
The temporary use of RED pins instead of constant keyword propagation sounds brilliant for me.

Assume you search for keyword "Benjamin" in all your images. You get 12 search results for only the 6 master RAW files and 6 individual JPG files.
Then by ticking the RED pins for all 12 search results this pin is propagated automatically to all versions (specific IMatch file relation setting).
Then you just have to search for all files with a RED pin and you get all (1+4)x6 + 6 = 36 images (including all versioned files from the 6 RAW files).
After all you have to untick the RED pin, because this pin is reserved for future temporary search. Did I understand the idea correct?

Peter

Thanks for this, Peter.
RED Pins is only one possibility of course. To "collect" a serie temporarly, I think, there is meant to use "bookmarks". Like using in a book, it means temporary.
I do this since years with IM3.

But of course you can use instead bookmarks pins, dots and so on.

There are several possibilities in IMatch, specialy in IM5. It depends all of a clever workflow.
In my case for example I do add always add the keywords to all images. If I have a master (nef) and 5 jpgs and 1 tif, I want to have the key "Bejamin" in all images. But of course, this every user can do, how he wants to do.

[Photon]

Hallo Markus,

The automatic propagation is the key with IM v5.0. This is most important for the decision about best strategy!
With possible automatic keyword propagation for versions the decision (taking into account pros/cons) is probably very individual.

I think I will also propagate my keywords with v5.0. As Ferdinand mentioned you can live with high numbers of counters since you can filter quite efficiently in IMatch. The opposite way to find uncategorized versions for categorized masters does look currently less convenient for me.

Martin

[BenAW]

• The command (F4,S) to show all versions is appropriate for one or a few images, but not for a medium or large number of images.
  This because the resulting search result window is not as clearly arranged as the category view.
  In this search result window, there are also missing the 6 images (with category "Benjamin") which have no version set.

Glad you don't like the layout of the result window either. You may want to support this feature request.
I never had the need to see images without versions in the result window, but this can simply be achieved by using bookmarks, pins or flags in the Collections view. Perhaps a feature request is in order to also show selected images that don't have versions in the result window, only when "Show Originals" is selected. I'll think about this one.

The biggest advantages of my workflow are to me:
- never having to worry whether keywords/categories etc. are propagated or not
- having only one image representing a version set WITHOUT filtering
- counts for images give the number of different images, not including the versions
- when using the filter, I don't have to consider using AND or OR to still keep the versions in or out the selection.