Google image search thumbnails not matching image

[boating_jo]

Hi folks, need help with Google search images, Coppermine 1.4.5 please and http://www.bymnews.com/photos/index.php
Google search,  site:www.bymnews.com images produces pages of thumbnails where part of the description and not the title is used and then when the image is clicked on a completley different image is then shown and a URL of something different again.

Attached images

[Nibbler]

Coppermine uses relative links (eg. 5th last uploaded file) and this of course changes over time. This has been addressed for the next major release of Coppermine.

Update your gallery, old versions have bugs and security problems.

[boating_jo]

Thank you Nibbler but we updated gallery recently and converted over to pid to get over the changing url problems which makes me even more puzzled at the result of  http://bymnews.com/photos/displayimage.php?pos=-14599 as the URL should be a pid

[Nibbler]

You are 5 versions behind, hence the need to update.

The image in the screenshot is that found at http://www.bymnews.com/photos/displayimage.php?album=lastup&cat=11&pos=4 which is a relative link. I don't see that kind of link at your site now, so maybe you need to wait for google to clean out old pages. It is also possible it found the link from another site.

[marian]

You are 5 versions behind, hence the need to update.

The image in the screenshot is that found at http://www.bymnews.com/photos/displayimage.php?album=lastup&cat=11&pos=4 which is a relative link. I don't see that kind of link at your site now, so maybe you need to wait for google to clean out old pages. It is also possible it found the link from another site.

Maybe I'm being dense and Jo is too, but that picture of Kate was uploaded yesterday, so I do not understand how it can appear with a pos url, when - as you say - that kind of link isn't at the site now??

[Nibbler]

That link is the link to the 5th most recently added pic in category 11.

At the time google crawled the site the 5th most recently added pic in category 11 at that time was http://bymnews.com/photos/displayimage.php?pid=16373 so that is what you see in google. That is what google believes the page to contain. When you go to the real page now you see the 5th most recently added pic now which is Kate.

[marian]

That link is the link to the 5th most recently added pic in category 11.

At the time google crawled the site the 5th most recently added pic in category 11 at that time was http://bymnews.com/photos/displayimage.php?pid=16373 so that is what you see in google. That is what google believes the page to contain. When you go to the real page now you see the 5th most recently added pic now which is Kate.

I still don't understand how pos urls can still work when there shouldn't be any there? Surely clicking on this http://bymnews.com/photos/displayimage.php?pos=-14599 should result in page not found?

[Nibbler]

Why should it?

[marian]

Why should it?

Maybe I'm utterly dense, but I can't see how pasting a non-existent pos url into a browser should reult in anything other than something like "Selected album does not exist", or a 404??

[Nibbler]

It is not 'non-existent' - there is a pic with that id. displayimage.php?pos=-xxx is the standard Coppermine method to get a picture. Unless you specifically disable that code it will give you the pic with that id.

[marian]

It is not 'non-existent' - there is a pic with that id. displayimage.php?pos=-xxx is the standard Coppermine method to get a picture. Unless you specifically disable that code it will give you the pic with that id.

Right!! Thank you very much that is most helpful, it was what I thought we had done when we converted to pid.
Could you also give some advice on why Google only seems to pic up images from part of the description, as Jo mentioned in her first post?

[Nibbler]

You'd need to do some research and find out what method google uses to label the images it finds.

[boating_jo]

Thank you Nibbler for your time and help with these problems 

[boating_jo]

"I have done as you suggested and there seems to be a consensus that google images relies heavily on an alt tag. Does coppermine assign such tags and, if so, how can I check them?"
Thank you again for your patience.

[marian]

I think you hold the key to the google images problem nibbler, even though you don't realise it!
If you put ic and the title of any of your gallery images into a google image search you get the result from your website.
If you put adityamooley Ashwini into a google image search you get "Your search - adityamooley Ashwini - did not match any documents.", even though Aditya has lots of images - with Ashwini in the title - in his gallery.
If you put adityamooley Ashwini into standard google, his website comes up in first place; that is exactly the sort of thing that happens with bymnews!  Put Cougar ace album into ordinary google and there is the BYM Gallery link right at the top; click on Google images and there is zilch!
It's pretty clear, to me, that you have done something right for your gallery, whilst aditya and bym haven't; please try to figure out what that is. I'm sure I am not the only one who loves coppermine, but is frustrated by the Google images issue. A gallery with 25000+ pix, like BYM, should get hundreds of GI referrals a day, we are doing well if we get 10!

[Joachim Müller]

As Nibbler suggested, we have labored on search engine hit improval for the next major release (cpg1.5.x, the current devel version, which has not been released yet and there is no scheduled release date). To improve search engine results, you have to make sure that:

• Pages can be found where google expects them to be. That's why we changed from pos to pid for cpg1.5.x
• Stop Google from spidering irrelevant pages (e.g. the slideshow pages or the rating links) in the first place. This has been addressed by changing some html links to javascript-driven links or form-based. Additionally, we have added the rel="nofollow" attribute for some links. Needs more looking-into though.
• Avoid double content, as this leads to a penalty in Google's page rank. We'll have to stop the bots from spidering meta albums
• Add relevant textual description to pages (using alt tags and <h1> to <h6> as they are considered to be more important and relevant)

You're welcome to post actual code change suggestions.

However, the initial question of this thread has been solved, so I suggest discussing this in another thread.