Using JS for image preview

[learning_brain]

There's no way I can describe this in just a title.

CURRENTLY

I have an image search engine (as some of you know), currently with a unique image preview function (hotlinks the original image).

On click, you're taken to the view image page, which embeds the original image in it (again hot-linked).

This is great for SEO as I have individual optimized pages for each image.

THE PROBLEM

Unfortunately, hot-linking is not considered good practice at the best of times and may now be deemed as being breach of originators copyright (recent court cases ref image search engines and image previews).

As a result, I now need to revert to a more standard thumbnail-only design

(preview on http://www.qualityimagesearch.com/test.php?q=fire+horse)

Currently, this is using a standard plugin light-box script (pretty nice), but it still uses a hot-linked image.


THE QUESTION

What I really want is a thumbnail page (as above preview link) but each click to go to a dynamically created page that can be optimized for SEO, but without grabbing the originators full sized image.

And before you ask, no I can't store 100,000's of full-sized images!!!


Has anyone got any ideas or advice as to which direction I should go?

Many thanks

Rich

 

[learning_brain]

Really??... No ideas??

How about a page that can be optimised for google with a redirect to the original image?

How does that suit SEO?

Richard

 

[essellar]

It's not so much a "no ideas", but that nobody was watching, I think. The recent change to the inactivity policy means that very few people are visiting the forums these days, and I've just been scanning the main page for new items myself (your original post never showed as a new item for me).

I don't think you need to make much in the way of a change, really -- all you need to do is replace the image source with a PHP-based proxy URL that fetches the original image and creates a thumbnail (or at least a reduced-sized image or, barring that, one that is watermarked with the source -- something that would count as "fair use" or "fair dealing" and do proper attribution). The generated image would lie within a link tag pointing to the original (preferably in the context of the original page).

 

[learning_brain]

Thank you!

I like the way that sounds.

I recreate images for the thumbnail creation anyway so that won't be too hard.

I just need to put together a "nice" page for image info and the original image link as suggested.

The only problem I will have here is showing the image in the orininating URL. Whilst I have this, I find this feature a pain on other sites and would prefer to link direct to the image url.

Unless there's a way to show both somehow? I've seen the way google handles it sympathetically but can't figure out how they do it.

Rich

 

[essellar]

Google used to frame, but bust-outs would break that (and too many pages bust out of framesets for some pretty good reasons). I think they're actually fetching and re-creating the page now (or at least serving their chached version). That's just a bit much for your site -- you'd need Google-sized storage or a metric crapload of available bandwidth.

NOTA BENE: IANAL and am not offering legal advice, just a layman's understanding.

The original context is an ideal to shoot for, of course, but practicalities (and a lack of infinitely-deep VC pockets) may mean that sending the user directly to the original image may be the best you can do with the means at hand. It shouldn't be that big a deal, as long as the full-sized image is displayed with the original site's domain in the address bar.

What to do about images hosted with Imgur, Picassa and the like, I don't know. Attribution is a problem because the URL doesn't properly indicate the name, pseudonym or website of the content creator, and without attribution it's hard to make a case for fair use/fair dealing. That's where making a good faith effort for the general case should make a difference -- intent is an element, so the worst that should ever happen if the worst should happen is an order to remove the result. Damages should not follow.

 

[learning_brain]

It's as I feared and you pick up on some very good points I stumbled across along the way.

I appreciate the NB and, for what it's worth, this is my laymans understanding too.

I did write a page caching script that worked pretty well with most sites but failed on far too many. Again though, this may be construed as recreating content... with much the same result.

As you say, I may have to revert to frames - top frame with my banner and lower half split into image/originating page. There are plenty of bust-out scripts on sites now which will make this very tricky.

A bit messy, but without the might of google-sized resources, I guess this is the best I'm going to do.

I also wish I could speed up the crawler, but hey-ho maybe I'll get my own server one day.

As for Picassa and the like, I've always had problems even crawling their sites, so I steer clear.

Many thanks for the views

Richard