How do I prevent parts of the file name from being used as a tag?

Opps · April 25, 2009

How can I avoid this? I've tried using file names that consists of only numbers, etc but it seem like no matter what I do parts of the file names end up as tags.

My two wishes for tags:
+ An option to turn this off
+ Support for UTF-8 tags

rainforest1155 · April 25, 2009

Opps wrote:

How can I avoid this? I've tried using file names that consists of only numbers, etc but it seem like no matter what I do parts of the file names end up as tags.

My two wishes for tags:
+ An option to turn this off
+ Support for UTF-8 tags

Hi Opps,

There's no way to turn that feature off, but you have to make a distinction between what gets added to the keywords for the photos and what actually shows up. Take another look - numbers might show in the 'edit keyword' box but they don't actually show as keywords as they are filtered automatically.

If you have any more questions, send us a link to a photo in question.

Sebastian

Opps · May 2, 2009

rainforest1155 wrote:

Hi Opps,

There's no way to turn that feature off, but you have to make a distinction between what gets added to the keywords for the photos and what actually shows up. Take another look - numbers might show in the 'edit keyword' box but they don't actually show as keywords as they are filtered automatically.

If you have any more questions, send us a link to a photo in question.

Sebastian

What I don't understand is why does it pick up those numbers if they are not used. And why doesn't it pick up all numbers ... there has to be some kind of rule for what gets inserted in the keyword field.

My solution for the moment is to clear all keywords after upload.

jfriend · May 2, 2009

Opps wrote:

What I don't understand is why does it pick up those numbers if they are not used. And why doesn't it pick up all numbers ... there has to be some kind of rule for what gets inserted in the keyword field.

My solution for the moment is to clear all keywords after upload.

I don't understand why Smugmug doesn't fix this. It keeps me from making my keywords public because there's so much garbage in them and it adds an extra step to every one of your uploads. It would not be hard to fix. Just take out the code that makes keywords from filenames. It ruins the use of keywords for many of us. Smugmug knows I've been asking about this for a long time and they've even acknowledged that there are better ways for it to work, but there has been no progress on changing it. It is frustrating.

devbobo · May 2, 2009

jfriend wrote:

Smugmug knows I've been asking about this for a long time and they've even acknowledged that there are better ways for it to work, but there has been no progress on changing it. It is frustrating.

come now John, you know this isn't true.

jfriend · May 2, 2009

devbobo wrote:

come now John, you know this isn't true.

I'm confused. Which part isn't true?

devbobo · May 2, 2009

jfriend wrote:

I'm confused. Which part isn't true?

you know that i have been actively working to improve keywords

darryl · May 2, 2009

Opps wrote:

What I don't understand is why does it pick up those numbers if they are not used. And why doesn't it pick up all numbers ... there has to be some kind of rule for what gets inserted in the keyword field.

My solution for the moment is to clear all keywords after upload.

http://www.smugmug.com/help/keywords-tags

If you added keywords in a program like iPhoto or Photoshop before adding the photos to SmugMug, we'll do our best to find those and automatically apply them. Specifically, we look for keywords in the IPTC keywords field. Please note that keywords added on SmugMug do not get embedded in the IPTC data.

We also look at the filenames of your images. If your filename is "Alvin_Birthday3.jpg", for example, we'll add the keywords "Alvin" and "Birthday" automatically. If your filename is "IMG_23987234.jpg", we'll ignore it.

If you click on "I want all the details." you'll see the rules:

* Words beginning with numbers will be ignored unless you enclose them in quotes. Words ending in numbers will have the numbers removed unless the word is enclosed in quotes.
* 1 and 2-character words, and common words such as and are also ignored unless you enclose them in quotes.
* If there is a space in your keyword, like "beautiful sunset", searching for beautiful sunset will not return a handy keyword link. It will still return all photos that match, though, and you can always browse that keyword by clicking on your keyword cloud, or going straight to the URL: http://smugmug.com/keyword/beautiful+sunset

There's examples too.

There's also some useful tips on searching for keywords here: http://wiki.smugmug.net/pages/viewpage.action?pageId=1572960

darryl · May 2, 2009

Opps wrote:

How can I avoid this? I've tried using file names that consists of only numbers, etc but it seem like no matter what I do parts of the file names end up as tags.

My two wishes for tags:
+ An option to turn this off
+ Support for UTF-8 tags

Can you link to some examples of images you've had the problems with?

For me IMG_1234.JPG does not result in any keywords. Also, I believe that according to the "rules", a filename like 1234.JPG would not result in a keyword.

If you're using some renaming images to something like "birthday_party_001.jpg", then I could see you getting extra tags.

Hey jfriend -- are you renaming your images? Any examples? I wonder where the junk tags are coming from. I know that on my son's preschool site we get junk tags when people insist on renaming their files things like "ccns_caltrain_field_trip_001.jpg". But otherwise I don't have this problem.

A while back though I remember Don or somebody trying to compile a list of "standard" file name conventions that different cameras used (like IMG_0123.JPG for Canon, and whatever other cameras use) in an attempt to prevent junk tags.

darryl · May 2, 2009

devbobo wrote:

you know that i have been actively working to improve keywords

To be fair, we don't really have visibility into what anybody at SmugMug is actively working on. When a SmugMug employees says on Dgrin that "yes, we know that's a problem" or "yes, we're working on that", it could mean it could get fixed next month, or next year, or in 3 years.

I understand that's how you've chosen to run things. But many of us have lots of issues (heh), and we can't always keep tabs on what the last thing any employee may have said about any particular issue, especially in an active forum like Dgrin.

I for one didn't remember that you said you were working on it. Last thing I remember hearing you were working on was some API and RSS feed stuff.

It's been interesting to see how a company like Atlassian manages their issues in such a transparent way. They don't necessarily fix something just because a lot people complain about an issue (they even have a voting system) or because it's been sitting there for 5 years (arrrgh). But letting people watch issues (change notifications via e-mail) and comment on them (not just "Me too!", but with useful data).

It's not always fulfilling (did I mention 5-year old tickets?) but it at least lets us confirm we're not alone, and at least give us the illusion of involvement as well as having an impact.

And at least if we had a ticket to look at we wouldn't accuse you of ignoring us, Dev. :-}

Because also, Dgrin is not a bug tracking system, and the searching sometimes sucks.

Andy · May 2, 2009

darryl wrote:

To be fair, we don't really have visibility into what anybody at SmugMug is actively working on. When a SmugMug employees says on Dgrin that "yes, we know that's a problem" or "yes, we're working on that", it could mean it could get fixed next month, or next year, or in 3 years.

I understand that's how you've chosen to run things. But many of us have lots of issues (heh), and we can't always keep tabs on what the last thing any employee may have said about any particular issue, especially in an active forum like Dgrin.

I for one didn't remember that you said you were working on it. Last thing I remember hearing you were working on was some API and RSS feed stuff.

It's been interesting to see how a company like Atlassian manages their issues in such a transparent way. They don't necessarily fix something just because a lot people complain about an issue (they even have a voting system) or because it's been sitting there for 5 years (arrrgh). But letting people watch issues (change notifications via e-mail) and comment on them (not just "Me too!", but with useful data).

It's not always fulfilling (did I mention 5-year old tickets?) but it at least lets us confirm we're not alone, and at least give us the illusion of involvement as well as having an impact.

And at least if we had a ticket to look at we wouldn't accuse you of ignoring us, Dev. :-}

Because also, Dgrin is not a bug tracking system, and the searching sometimes sucks.

It'd be nice to have a really fancy tracking system. We're 47 people doing the work of 150. One day we'll get there

(Chicken and egg thing? Yup, you bet).

In the meantime, post away and we try and answer questions as best as possible.

Thanks!

Andy · May 2, 2009

jfriend wrote:

I don't understand why Smugmug doesn't fix this.
.....(snip....

It is frustrating.

We've assigned a Sorcerer to keywords and improvements are forthcoming. Your patience, and understanding is greatly appreciated. We have a long list of feature requests and improvements, and 7-year track record of delivering on them, as you know, because you've been here for most of those 7 years

Opps · May 2, 2009

darryl wrote:

Can you link to some examples of images you've had the problems with?

Actually no, since I've already deleted the keywords

darryl wrote:

For me IMG_1234.JPG does not result in any keywords. Also, I believe that according to the "rules", a filename like 1234.JPG would not result in a keyword.

I just did a test and uploaded an image with the file name "2009-04-30_19-20-55_1383 11-56-13.jpg" which resulted in the tags "2009; 1383" ... I renamed the file to "DPT 239832.jpg" and got "dpt; 239832" ... I then tried "IMG_1234.jpg" and "DSC_1234.jpg" which both resulted in to tags.

I suppose I have to use a special name pattern for when I'm exporting files to SmugMug but it would be so much easier if there was a "don't make tags from filenames" option.

pilotdave · May 2, 2009

I used to think it would be a good idea to have the ability to shut off this bug...errr... feature. But when i realized how polluted SmugMugs global keyword list is, I've concluded that converting filenames to keywords is pretty much always a bad idea. I'm really looking forward to some improvements to keywording on smugmug. Getting rid of this bug is step one, and will be a giant improvement.

While we're on the subject, I'll explain how the caption/keyword tool could be improved. I add a new gallery to my site pretty much every week. Often I upload pictures to the new gallery two or three different times. For example, I'll create the gallery on a saturday night and upload the first batch of pictures. Then on sunday night I'll upload the next batch.

But I want to get a head start with captions and keywords on Saturday night, so i go ahead and enter all of them. This process usually starts with "Empty all keywords" to get rid of the junk from the filenames. I caption each picture, and then I add the date to the end of each caption, since smugmug doesnt give an automatic way to show the date a picture was taken without pulling up all the exif data (feature request, hint hint).

Ok, then i get to sunday night. Now half my pictures have captions and keywords and the other half has filename junk in the keywords. I can no longer just empty all the keywords to start from scratch. There's no easy way to "select all pictures with a blank caption" (feature request, hint hint). So I have to manually select all the pics... i go with whichever list of pictures is smaller and either choose selected or unselected to be a little mor efficient.

Then theres at least one keyword that applies to all the pictures in the gallery. I already added it to all of saturday's pics, so i need to first remove it from all, then add it to all. Now it'll appear on all the pics without being doubled up on saturday's pics. Would be really nice if i could "select all pictures WHERE keywords does not contain "____". Very much like the add/remove stuff at the top, but specifically for making a selection of pictures.

Finally i got all my captions and keywords entered. Now i want to put the date at the end of sunday's captions. Again, i need to select all of sundays pictures manually, and add to selected, or i need to choose all of saturdays pictures and add to unselected. Would be really nice if there was some kind of tag, like %datetaken% or something, that just automatically got replaced in the caption with the correct date (do i need to point out the feature request?).

So basically it comes down to not converting filenames to keywords, providing a way of automatically filtering what pictures i'm selecting so i dont have to click on each one, and providing an easy way to add repetitive information to captions, like the date or other metadata. This would save me a ton of time every week.

Dave

Andy · May 2, 2009

pilotdave wrote:

Getting rid of this bug

It may not be working as you expect or want, but it is working as it was designed - so it's not technically a bug. I'm not quibbling here - and we're reading and swallowing all the input you and others are giving about filenames.

But know this: if we changed it there'd be equal numbers of folks unhappy about the change.

jfriend · May 2, 2009

Andy wrote:

It may not be working as you expect or want, but it is working as it was designed - so it's not technically a bug. I'm not quibbling here - and we're reading and swallowing all the input you and others are giving about filenames.

But know this: if we changed it there'd be equal numbers of folks unhappy about the change.

Now you're just making excuses. If you have half your customers unhappy one way and half the other way, you engineer a solution for both. It's not rocket science. Worst case you offer a preference, but there are probably even better solutions than that.

This whole conversation could have been over with a happy resolution if you or devbobo just said, "we're actively working on it and a solution should be available in a reasonable amount of time" and that was actually true (I would assume a "reasonable amount of time" is something less than 3 months).

I'm unclear what devbobo is talking about that I should know something is being worked on. If I was told something awhile ago, I don't remember it. Maybe I forgot. Maybe it was awhile ago. I don't remember.

Andy, when you don't tell us customers about when something is going to get done and we go a very long time without any visible movement on something, it is human nature for us to wonder if anything is actually happening and suspect that maybe nothing is. So our only avenue is to ask about it again and see if we can learn more and find out what's going on. You chose to work this way. It makes frustration for customers who see no visible progress and hear no news on it. Those are the consequences of the way you choose to communicate with your customers.

Darryl, I've supplied exhaustive examples to Smugmug before in previous threads. I'm not going to go through it all again. Suffice it to say that my images all come out of my camera with a unique prefix and a number and then end up as this when they go to Smugmug "JLF_3456.jpg". That results in keywords "JLF" and "3456" on every single image which are total junk. I use a unique prefix on each camera (I use three different cameras) so that I can easily sort out which camera an image came from at the file system level and so that I never have filename conflicts between them. I do not touch the filename myself ever. The ONLY keywords I want on my images are the ones I put there on purpose and are in the metadata.

Andy · May 2, 2009

jfriend wrote:

Andy, when you don't tell us customers about when something is going to get done and we go a very long time without any visible movement on something, it is human nature for us to wonder

So years of a track record of listening to our customers (and you, for sure, in particular) get thrown out the window?

We don't give dates or pre-announce. When we do, unforseen circumstances can cause us to have egg on our faces

It's really, really hard for us to satisfy everyone and we can't ever communicate enough for you guys

Devbobo's doing some major work on keywords and it's been in the works for a while. I've sent you private internal emails on the subject. It's a huge engineering effort. I wanna let let our Sorcerers do their job

Thanks for your (and everyone's) patience while we improve things

Oh and thanks for you (and everyone here) telling us how important such changes are. We don't want it any other way.

Oh and if you read the mail I've sent you a while back you'd know that we're not making excuses and we intend to try and make everyone happy

darryl · May 2, 2009

Andy wrote:

It'd be nice to have a really fancy tracking system. We're 47 people doing the work of 150. One day we'll get there (Chicken and egg thing? Yup, you bet).

In the meantime, post away and we try and answer questions as best as possible.

Thanks!

Hey Andy -- I doubt that the JIRA system from Atlassian is that much fancier than what you use in house. The big difference is that Atlassian has decided to make their bugs visible to the outside world:

http://jira.atlassian.com/browse/JRA

darryl · May 2, 2009

Opps wrote:

Actually no, since I've already deleted the keywords

I just did a test and uploaded an image with the file name "2009-04-30_19-20-55_1383 11-56-13.jpg" which resulted in the tags "2009; 1383" ... I renamed the file to "DPT 239832.jpg" and got "dpt; 239832" ... I then tried "IMG_1234.jpg" and "DSC_1234.jpg" which both resulted in to tags.

I suppose I have to use a special name pattern for when I'm exporting files to SmugMug but it would be so much easier if there was a "don't make tags from filenames" option.

I found the [post=869193]old post[/post] from Don that specified exactly what filename formats result in no tags. It's not very long:

DSC#
P#
IMG_#
IMG #
DSC_#
DSCN#
DSC #
DSCF#
PICT#
PICT# #

And some various other strings that are 2 chars or less.

JT also posted a link to a bunch of camera filename formats he found: http://diddly.com/random/about.html

So SmugMug isn't going to be able to deal with purely date-based filenames produced by renaming utilities like Stamp, or custom filenames like John's JLF_1234.JPG.

My last point -- it's interesting but going strictly from Dgrin posts, it seemed like more people were happy than sad when SmugMug broke filename->keyword functionality back in March 2008. I imagine e-mails to support might tell a different story. Or not.

http://dgrin.com/showthread.php?t=85963

Either way, having the option simply be configurable (site-wide on or off, please) would solve the problem for everyone, I think.

darryl · May 2, 2009

jfriend wrote:

Andy, when you don't tell us customers about when something is going to get done and we go a very long time without any visible movement on something, it is human nature for us to wonder if anything is actually happening and suspect that maybe nothing is. So our only avenue is to ask about it again and see if we can learn more and find out what's going on. You chose to work this way. It makes frustration for customers who see no visible progress and hear no news on it. Those are the consequences of the way you choose to communicate with your customers.

It's actually my nature to just forget even if they do tell us they're working on something. Despite appearances, I do *not* hang out on Dgrin every day nor do I remember everything everyone may have said on a given topic. So yeah, I end up asking again.

An open bug tracking system would save everyone some trouble (except of course the poor sap who gets tasked with installing it and migrating all of the existing bugs). I suppose you'd have to decide whether you want to make bug priorites publicly visible, or you could do like Atlassian and apparently ignore those priorities internally. (I imagine they must have a separate list of "must fix" bugs that's separate from "bugs customers think are must fix".)

pilotdave · May 2, 2009

Andy wrote:

It may not be working as you expect or want, but it is working as it was designed - so it's not technically a bug.

I'm well aware. But I'll continue to call it a bug when I feel like complaining.

I liked when that "feature" was broken. Bring back the bug!

Dave

kilroi · May 13, 2009

I have several thousand of my own images that I would like to migrate to smugmug. Each and every one of these images has a distinct filename, for my own personal use on my own computer. I have precisely zero desire for any of these to be keywords. Work with me, and the at least 30 other threads requesting this - come on guys!

jfriend · May 13, 2009

kilroi wrote:

I have several thousand of my own images that I would like to migrate to smugmug. Each and every one of these images has a distinct filename, for my own personal use on my own computer. I have precisely zero desire for any of these to be keywords. Work with me, and the at least 30 other threads requesting this - come on guys!

The only work-around I know of is to make a copy of all your images before you upload them, rename them all to DSC_xxxx.jpg (which gives them suffixes that Smugmug will ignore), upload them with those filenames, then delete the local copies. I think the issue is being worked on, but we haven't seen it yet.

liflander · January 31, 2013

Any updates on keywords?
Hi all.
I know this is an old thread. Have there been any updates on keywords?
Did anyone suggest adding a smugmug feature that renames files upon upload? Does starexplorer do that?

Thanks,
Mark

jemostrom · January 31, 2013

Keywords works fine for me as far as I can see (for quite some time).

EntropicTendencies · January 31, 2013

liflander wrote: »

Hi all.
I know this is an old thread. Have there been any updates on keywords?
Did anyone suggest adding a smugmug feature that renames files upon upload? Does starexplorer do that?

Thanks,
Mark

You can rename when publishing from Lightroom.

I'm not sure I'd recommend it since it could make it difficult if you ever needed to sync back to Lightroom (the filename is used as one of the criteria to match the master image).

If you don't want keywords from filenames you can turn it off in Account Settings > Discovery > Search.

How do I prevent parts of the file name from being used as a tag?

Comments