Closed: Catalog should respect case

[Hayo Baan]

Currently everything in the catalog is made lower case, not only does this look ugly, it also does not allow one to look for errors in spelling (capitalization), perform searches on capitalized words, etc.

The catalog should fully respect case.  I consider this a bug, not a feature request.
It would be nice if searches could (optionally!) be made case insensitive. Given that there is a proper database underlying the catalog, this should be doable, I'm sure.

Note: I've asked this earlier but have now separated the request to be able to better track it.

[carlseibert]

I disagree wholeheartedly.

Case sensitivity should be off by default, if it's available at all. For administrators of DAM systems, people turning on case sensitivity by accident is one of the biggest generators of service tickets. The user can no longer find anything in the expected way, can't do their work and basically freaks.

Very few users are able to actually benefit from case sensitivity in searches. But many insist they want it. It was a big thing in the nineties. Professional librarians were all a clamor for case sensitive search. Once they had it - meh - even people who know how and when to use it rarely do.

Basically, case sensitivity requires the user to know the unknowable. Does the word they're searching for exist, in the particular asset they are seeking, at the beginning of a sentence or in the middle? That's an easy question AFTER you've got the asset in hand, but a little tricky to answer beforehand, when you're formulating your search.

 

[William R Wood]

Sorry, but I agree wholeheartedly with Hayo!

He wants the catalog to respect case everywhere which I totally support. I hate, for example, the fact that my keywords are all in lower case.
But he wants the option to make search case insensitive which eliminates the issues you raised. I suppose search could have case sensitivity off by default with the option to turn it on. I don't like software developers to hard code choices, giving users the option is better.

[syncrasy]

+1 for Hayo's requests: (1) Respect capitalization everywhere and (2) Provide an option to make search "case insensitive." Very standard functionality for a DAM.

Carlsiebert, perhaps you read Hayo's post too hastily and focused only on "search," missing his observation that PMP is turning capitalized words into lower case words throughout the catalog. If I interpret Hayo's (and William's) description correctly, this means that the keyword entered as "Eurasian Lynx (Lynx lynx)" is currently being changed and displayed as "eurasian lynx (lynx lynx)". I don't know if this behavior is a bug or a design flaw, but it is not acceptable.

[carlseibert]

Indeed, maybe I'm missing his point. In my copy of the beta, I see everything in mixed case, however I typed it, keywords included. (build 3029 (9dd0f3c for Mac) So maybe we're talking about two different things.

Oh wait.... NOW I see it.

Ah, you're talking about the tab that shows you the number is instances of given search terms in given fields in the catalog.

And your worry is that that list might actually be the index itself, thus making case sensitive searches impossible, yes? That would be a question for Kirk. And the answer might hinge on performance.

Interface-wise, the tradeoff is that the list would be twice as long and you would have to set two buttons to see most terms. "Lynx" and "lynx" are the same animal, like "car" and "Car".

Functionally, in my experience - which is extensive - if everything is done case-less-ly, that would probably be a favor for 99% of users. Real world, nobody looks for upper case "Car" as opposed to, well, an automobile. And nobody I've ever seen has programmed synonyms for upper and lower case variations of every search term in their keywords. Much more reliable just to ignore case.

The Find function in Photo Mechanic does have selectable case sensitivity. And it does work on contact sheets that are made of catalog assets. So, if you wanted to correct all your lowercase Lynxes or vice versa, you could simply search for all lynxes, and use Find to highlight in the returns the ones that need changing and go from there.

The classic classroom exercise is "Barbara Bush" vs "rose bush". Frankly, I've never had a problem finding a picture of Barbara Bush. It would take a heck of a lot of rose bushes in a collection to swamp the returns. But if it happened, using Find and then suppressing the rose bushes would work just fine.

[Hayo Baan]

I think we're all on the same page now. But just to answer @carlseibert's question:

The problem with the current PM implementation is that the catalog, index and all, is treated as completely being lower case. So there is no way to be able to make the distinction between "Bush" and "bush", the former perhaps the name of a former US president, the latter a plant. Not in queries, nor while browsing or filtering the catalog. These are different things and I need to be able to distinguish between them. All in all I'd say it is pretty clear case should be respected in the database. With the option to make searches behave case sensitive or case insensitive. As a separate option, the filter/browse item lists could be made all lower case just like they are now or also case sensitive (this imo should be a separate option from how you perform searches).

[Michael Naylor]

I've yet to try PM6 Plus, but when I do I'll be importing 20k images which have carefully entered keywords using Proper Case for People, Places and Events, and lower case for objects.   I'll be very disappointed if all my keywords are changed to lowercase.

[Kirk Baker]

I've yet to try PM6 Plus, but when I do I'll be importing 20k images which have carefully entered keywords using Proper Case for People, Places and Events, and lower case for objects.   I'll be very disappointed if all my keywords are changed to lowercase.

That's the way it is at this time.  This may change before final release.

-Kirk

[syncrasy]

I've yet to try PM6 Plus, but when I do I'll be importing 20k images which have carefully entered keywords using Proper Case for People, Places and Events, and lower case for objects.   I'll be very disappointed if all my keywords are changed to lowercase.

That's the way it is at this time.  This may change before final release.

-Kirk

Kirk, I'm following this and other threads (can't beta test myself either), but I have to tell you that this will be a deal breaker for me. I can't use a DAM that doesn't respect case as typed/entered. I hope "may" will become "will."

Actually, I'm somewhat baffled by this thread and wonder if I'm misunderstanding something. I know many PM users are sports photographers who have used PM to tag photos and submit them to agencies, reps, publishers, etc., and presumably PM has always honored case (e.g., it wouldn't change "Steph Curry" to "steph curry" or "Lionel Messi" to "lionel messi"). Is that really what's happening in the database index for PM6+? Or is this just happening in certain views (e.g., keyword lists, but not captions or the index)?

[Hayo Baan]

Mark, it isn’t as bad as you describe. Metadata in the file still respects case as entered. It’s only in the database that everything is made lowercase. This still is highly unwanted behaviour though, at least by me.

[syncrasy]

Mark, it isn’t as bad as you describe. Metadata in the file still respects case as entered. It’s only in the database that everything is made lowercase. This still is highly unwanted behaviour though, at least by me.

Ah, file metadata vs database. Well, since the database is the heart of a catalog (and the source of future metadata), seems bad enough. I'm glad you raised the issue.

[carlseibert]

Syncrasy - Sports photographers put the names of their subjects in the caption and, depending on the agency they work for, maybe the headline. Properly capitalized in both places. Keywords serve a different purpose. And in this case, the "database" actually is the metadata on the files. (That's the beauty of it). The index is just a tool to make searching faster. It's not the data.

I just stopped and figured. In my decade or so as a daily picture editor, I executed over 125,000 searches. (Real, not repetitive. Too many of those to even guess at.) I used case sensitive searching maybe a half dozen times. And probably didn't need to then. Case sensitive searches are dangerous. They do stupid things. If you are searching for "flying", for instance, a case sensitive search will fail if the word begins a sentence. That's like shooting yourself in the head.

Hayo - I actually have searched for Barbara Bush and not in a classroom. In this particular example, since "Barbara" and "Bush" are proper nouns, a case sensitive search could be acceptable. But hardly necessary. The risk if you do it without case sensitivity is that "barbara AND bush" might hit "Barbara Smith tends her rose bush." You won't get every ol' rose bush picture because there were two terms in the search - "barbara" and "bush". There weren't many/any Barbara Smiths with bushes, so there wasn't really be much noise. (And I was working in a collection with over two million assets.) And what if the search was "Barbara AND Bush AND charity"? That would fail if "charity" started a sentence.

On the other hand, even in this rare case where it won't shoot you in the foot/head, case sensitivity doesn't eliminate any risk.  Barbara Smith can still be a very dangerous woman. A sloppy person is at risk of running the wrong Barbara, rose bushes or no. A "Barbara AND Bush" search, case sensitive or not, can still bring back a day-ruining return. "Barbara Jones speaks to Edward Bush" will return for that search. If case sensitivity made the editor who did that search feel more secure, it certainly wasn't doing him or her a favor.

Where a case sensitive search CAN be of help (0.00048% of the time, if my experience is typical) is to filter out an overbroad return. If you get flooded with rose bushes, as I said, You CAN in PM+ filter your return with a case sensitive search using Find. But that only works if you don't accidentally exclude valid returns in the first place.

I hate to be strident here, but really. Professional librarians can sometimes use case sensitivity to good effect. For the rest of us, it's just a dangerous weapon we shouldn't touch.

[Hayo Baan]

Carl,

Yes, for searches, you are probably right; those you will likely do case insensitive most of the time. BUT what I really dislike at the moment that when I see my keywords, locations, etc. in the catalog, they are now all lower case. This doesn't only look horrible (especially for proper names), it also disables you from finding you made a spelling (case) error. And that is bad, since it is the metadata that we later would like to be able to e.g. set the caption with (using the many variables that PM offers for this). Here case is absolutely important!

What I don't understand is why it was deemed necessary to make everything lower case to begin with; databases can perform queries in a both a case sensitive and case insensitive way. It's just a matter of specifying the collation.

[Hayo Baan]

Solved in build 3719.