When uploading my MC photos to SmugMug.com, Keywords values that used Backslashes for nesting are not being seen as separate values, the nested values get smashed together into a mess.
I complained, declaring that JRiver Media Center makes helpful use of Back slash delimiters for nesting. SmugMug looked into it, and (amazing) actually modified their upload code to properly read and store Keywords that have Back slashes. BUT in testing, I still found my photo Keywords values mangled upon upload to SmugMug.com.
SmugMug looked further, and said that while I requested and they provided better handling of Back slashes, my photos actually all have Forward slashes, which will not work as desired.
Huh? I'm certainly typing Back slashes for Keywords nesting, and seeing Back slashes in MC. But (below), SmugMug (SM) provided examples of Keywords values extracted from my photos that have Forward slashes and therefore do not work properly. They say all my photos do. How and why?
Using ExifPro to examine MC-tagged photos shows this in Keywords: !Events/Wedding, !Places/US/CA/La Jolla. That's what SmugMug sees too. But what I actually typed and see in MC is this: !Events\Wedding, !Places\US\CA\La Jolla. The Forward slash version of Keywords seems to be created by MC, which then breaks use of those keywords by other apps.
Using a cruder tool, I can dig into the physical files enough to see the same Keywords values stored at least twice, in separate data blocks, one with Back slashes, another with Forward slashes. This might indicate SmugMug is reading the meta data header block "wrong", OR it might indicate that MC's syntax is not "right" so SmugMug doesn't know what part to read and use?
And beyond the slashes question, SmugMug says MC's XML tagging of multiple Keywords values is not per the standard.
Note that the XMP way of using XML for image (and PDF) tagging arose from Adobe in 2001, and in 2012 became an ISO standard (16684-1). Nowadays XMP seems to be used by every modern photo app and site that I encounter, instead of the 1970s IPTC format for metadata (International Press Telecommunications Council method for documenting shared photos, years before digital, of course, but in mid-90s implemented in Photoshop). I think photo apps that put both in each photo is not uncommon.
Adobe has extensive documentation and developer tools for XMP:
https://www.adobe.com/products/xmp.htmlhttps://www.adobe.com/devnet/xmp.htmlBut what does JRiver MediaCenter use? Why does it not seem to work with SmugMug and Adobe Lightroom?
Here are excerpts of comments by a SmugMug product manager (September 11, 2018).
The EXIF format supports hierarchical keywords in the metadata, [JRiver] just need to make sure they export the keywords in that format (specifically calling out the top level keyword and then its children in the proper XMP format).
...In an "XMP" format in which there's a special recipe for how keywords are written into the photo. It often looks something like:
<keyword>KeywordOne</keyword>
<keyword>KeywordTwo</keyword>
<keyword>KeywordThree</keyword>
Each `<keyword>` line tells the software when one keyword ends and the next one begins.
Hierarchical keyword support was later added to image formats and adds a `children` flag that identifies the sub-tags, typically looking something like:
<keyword>KeywordOne</keyword>
<keyword:Children>
<keyword>subKeywordOne</keyword>
<keyword>subKeywordTwo</keyword>
The [JRiver Media Center] photos you uploaded all have XMP metadata that looks like:
<keyword>!Events/Hike</keyword>
<keyword>!People/Bob</keyword>
Because we're reading the keywords based on the metadata spec, which states that keywords are separated by `<keyword>` we see the full string as the keyword.
_Technically_ we're following the metadata spec and reading the metadata properly. However, just because we're doing what the metadata spec says doesn't always mean we're thrilling all of you. That's why we wanted to understand how you and your photo editor store these keywords so that we could do extra processing on the photos to get the outcome you'd like.
I pulled up a few of your recent uploads and it was very easy to understand why your photos are having the issue... It appears your photos are all being uploaded with forward-slashes:
What can I do to have MC preserve my Back slashes in Keywords stored in the JPG header? And shouldn't MC store metadata exactly per what is considered the XMP metadata spec? Why ignore or conflict with Adobe and the 2012 ISO standard? (SmugMug integrates with Adobe Lightroom and claims that tagging format is exactly what SmugMug uses.)
By the way, I used the discussion in SmugMug's forum to strongly brag about how simple and fast it is to use MC for library organization and tagging, far far better than Adobe Lightroom (which does other things, of course, but is clumsy and slow). Unfortunate that my attempt to Spread The Word about Media Center got derailed by discovering the metadata incompatibility.