Ensure that other photo sharing sites have no reason to resent Seadragon.

  • 3
  • Idea
  • Updated 11 years ago
  • Under Consideration
So this is something of a variation on my earlier request for friendliness towards other image sites like flickr, etc.

It would be nice to be able to have the link on the Seadragon page point to the actual page that the original image appeared on, rather than the hotlink to the image itself. In flickr's case it would be absolutely great to be able to throw the URL to the front page of any flickr photo at seadragon.com and have it find the highest resolution copy of the image hosted on flickr.

You could also parse the license on that photo and decline to convert and host it if it is copyrighted. If more serious integration was attempted such as logging into one's own flickr account, it would be a step towards ensuring that the owner of the image still had the capability of using their own copyrighted images with Seadragon, while turning others away to avoid any legal problems or angry copyright holders looking at Seadragon with an evil eye.

While I'm on this topic, perhaps parsing the author and copyright data out of EXIF tags (or giving the option to fill in copyright data in a fold out blank) could result in a proper copyright/creative commons info in the corner of the Seadragon viewer a la Photosynth.

I haven't read over the terms of use super carefully to know what the legal status is of images converted to DZIs and I'm sure that you all would rather spend your time bringing improvements to the technology itself, rather than pandering to end users and your peers in photos sharing, but I strongly believe that it's incredibly important to have a good relationship with photo sharing sites out there. I would absolutely love to see a Seadragon viewer option on Flickr one day, but that sort of integration in the future will certainly never happen if in the present they build a reputation for Seadragon in their minds as something that simply takes their valuable images that draw viewers to their site and whose only form of linking back to flickr (or other sites) is in the form of a hotlink to the image on their servers.

Again, I love Seadragon. The hard work that I suggest above is only to save your good name. I feel like going the extra mile to show respect and courtesy to your peers would keep this from being one more case for people to resent Microsoft in spite of incredible technology because of some small quirk in presentation/conduct.
Photo of Nathanael Lawrence

Nathanael Lawrence

  • 795 Posts
  • 55 Reply Likes
  • happy

Posted 11 years ago

  • 3
Photo of Daniel Gasienica

Daniel Gasienica

  • 73 Posts
  • 19 Reply Likes
First of all congratulations to the whole team on the successful seadragon.com launch! It's truly beautiful, both in form & function... :)

I absolutely agree with Nathanael, it would be great if in the future seadragon.com could create a better bond with the original image. After all it's called the web.

Personally, I am an avid Flickr user and back in January I was working on this idea to link images on Flickr to their multiresolution counterparts, e.g. DZI, stored anywhere in the cloud. If a Flickr user could give seadragon.com access to her photo stream through the Flickr API, the system could link the original image on Flickr with the converted image through the use of Flickr's machine tag mechanism. By creating that link between the original and the multiresolution version we're opening up many possibilites since the entire metadata of the original is now linked the zoomable image. Through the use of the Flickr API we would then have access to title, tags, copyright, geo, comments etc. from within the viewer.

The following are two examples, the first one being from a test I ran back in January and the second from two minutes ago. If you go to the original image page on Flickr, you can see the machine tags on the right-hand side by clicking on the Show Machine Tags link:

Example (January, 2009)
Original
http://www.flickr.com/photos/gasi/319...

Machine Tag
seadragon:source=http://static.gasi.ch/images/3198854881/image.dzi

Example (August, 2009)
Original
http://www.flickr.com/photos/gasi/369...

Machine Tag
seadragon:source=http://seadragon.com/content/1ax.dzi

So far, obviously, I've been the only one doing this stuff... :)
http://www.flickr.com/photos/tags/sea...
http://www.flickr.com/photos/tags/ope...

Wishful Thinking
As a developer, obviously I'd love to see both crossdomain.xml and clientaccesspolicy.xml on seadragon.com, though I can understand the implications of that. This and an API could enable people to build their very own custom viewing experiences.

Cheers,
Daniel

Further Reading (Machine Tags)
http://www.flickr.com/groups/api/disc...
http://husk.org/code/machine-tag-brow...
http://www.flickr.com/groups/mtags/
http://code.flickr.com/blog/2008/12/1...
http://radio.aol.de/user/underpangs/j...
http://code.flickr.com/blog/2009/07/0...
Photo of Aseem Kishore

Aseem Kishore, Former Employee

  • 702 Posts
  • 124 Reply Likes
Daniel, great stuff as always. As I mentioned in my reply, we do have an API in the works. We'll make sure to include both a crossdomain.xml and clientaccesspolicy.xml in there. I can't wait to see what you do with it once it ships!
Photo of Daniel Gasienica

Daniel Gasienica

  • 73 Posts
  • 19 Reply Likes
Aseem, I just wanted to say that I very much support your decision to release early, release often and iterate on it. It's certainly better than pondering all these extras, which can be added later, and denying people access to such a valuable and useful tool as seadragon.com already is now.
Though, I admit, I'm stoked at the thought of having access to an API. Would that mean developers would also get cross-domain access to the DZI files such as Flickr currently offers access to their images?
Photo of Aseem Kishore

Aseem Kishore, Former Employee

  • 702 Posts
  • 124 Reply Likes
Yep. Not just for SL/Flash but also for Javascript through JSONP. (Though that restricts us to HTTP GET only, no POST, which disappoints my perfectionist side a little bit. =P)
Photo of Nathanael Lawrence

Nathanael Lawrence

  • 795 Posts
  • 55 Reply Likes
About half an hour ago I started entering the seadragon.com/view/ URLs in sequence (http://seadragon.com/view/a, http://seadragon.com/view/b, http://seadragon.com/view/c, etc.), curious to see what other people had been converting.

I've only tested up through http://seadragon.com/view/t so far, but the results confirm in every way that what I was suggesting for convenience's sake (deriving the link to the largest version of the image hosted on the photo sharing site from its default view html page's URL - mainly because I get tired of clicking a minimum of three times to get to any original size flickr image) is not merely a nicety, but absolutely necessary.

The percentage of people pasting links to photo flickr pages is rather overwhelming so far. Somewhat horrified by people's apparent inability to understand, I became curious as to whether they had simply made an initial mistake and thereafter realised their error and corrected it so I began following the links to the original pages/images, copying the actual image link and feeding it back to Seadragon. Unfortunately these have all appeared to convert for the first time, meaning those users who experienced failure to convert an HTML page instead of the image on it (admittedly their fault and not Seadragon's) did not understand what went wrong. It is unknown if they have ever experienced success with other images. (Perhaps you could correlate IP addresses to track success/failure rates for your own assessment.) Additionally there appear to be quite a number of people linking straight to URLs where no large image exists (or is even linked to). (Are they expecting a zoomable version of their homepage?!)

It seems to me that this text seems to need to be posted much more prominently:
"...Make sure it's to the image directly, not to a page that contains it."

I certainly understood right away what I was to do, but apparently the percentage of people whose understanding of getting a link to an object extends only to copying what they see in the address bar, a little help is required. A simple explanation of how to right click on an object to view it's properties, copy its address, or view it alone in the browser window (leaving the direct link in the address bar to copy as they are accustomed to) could go a long way towards improving the success rate.

I'm curious to know the ratio of image URLs to non-image URLs submitted. If you have a way to track that (human error) rather than simply successful conversion to failed conversion (computer error), I think my point must surely be made in the statistics, but perhaps my limited polling has painted a darker picture than is accurate for the whole body of results.

In any case, I felt that it was enough of a problem to warrant comment. I know that you don't want to put off your users by overdoing the explanation up front, but even miniature illustrated tutorials showing the easiest way to get an image URL in the most recent version of each mainline browser (IE8, FF3, Opera 9/10, Safari 4, Chrome 2) on the Create page would go a long way as long as you're taking time out to explain things a bit more. Firefox does something very similar for the install process on the 'Thanks for choosing Firefox' page.

Perhaps stressing image URL on the front page and adding something friendly similar to "Don't know how to find an image's URL yet? Click here to learn." would encourage a higher success rate.
Photo of Aseem Kishore

Aseem Kishore, Former Employee

  • 702 Posts
  • 124 Reply Likes
Yes, we should and hopefully will give more detailed instructions for getting the image from a page. Part of the reason we didn't from the start was we didn't want to insult our users; we weren't sure this would be as significant a problem as it has turned out to be. =) Another reason was actually that it's painfully difficult to get an image URL from IE!
Photo of Aseem Kishore

Aseem Kishore, Former Employee

  • 702 Posts
  • 124 Reply Likes
Official Response
Nathanael, as always, great stuff. And again, we're all on the same wavelength.

We expected from the start that a significant portion of submitted URLs would be to e.g. Flickr pages rather than the image directly. We're well aware of it and want to solve it, but we got some pushback for v1 (and I think rightly so, since it would have been spending time solving a problem we didn't know would necessarily be a major problem, though turns out our expectations were right and it was).

Similarly, linking to the source page of the image as opposed to the raw image itself is also something we talked about -- in my case, I used Wikipedia as an example -- but we again decided to cut it for v1.

But your idea specifically of recognizing all the variants of an image's URLs (e.g. in the case of Flickr, main photo page, each of the size pages, and each of the raw images) as one image -- and linking back to the main source (e.g. the main photo page) -- is a really excellent idea.

Handling copyright info and machine tags and things like that is a bit much to me, and I don't see us pursuing this just yet. This is mainly just because this is stuff very specific to sites like Flickr and not available for any image in general.

But -- exposing an API would let you guys and the community build some of that stuff on top of Seadragon.com. An API exists, but we decided not to expose it publicly for v1. It needs some polish to make sure it's done "right" in order to be forwards-compatible with all of these ideas and features. Hopefully we'll be able to get it out soon. (I've already written what I think is a powerful app on top of it, so I can't wait either! =P)

Thanks for all these ideas, guys, and seriously, keep it coming.
Photo of Daniel Gasienica

Daniel Gasienica

  • 73 Posts
  • 19 Reply Likes
Yeah, I understand your stance on machine tags, as they probably are a feature that at this point is only appealing to metadata fetishists as I am ;P I really appreciate your openness in collaborating with the community and sharing your ideas for the future of seadragon.com
Photo of nbusekist

nbusekist

  • 1 Post
  • 0 Reply Likes
I just signed up with www.mejuba.com. Unlimited storage for both videos and pictures and it's completly free. No monthly limits or qoutas.
Photos and videos are stored in their original formats and sizes and are kept unmodified for backup.
I also like that you can geo-tag your stuff so it shows up on a map. And you can search for stuff on the map - that's cool!.
As the only site i know of it uses a Windows Explorer like navigation with folders - super eacy to use - even has drag and drop!.
I can highly recommend it.