Advertisement
If you have a new account but are having problems posting or verifying your account, please email us on hello@boards.ie for help. Thanks :)
Hello all! Please ensure that you are posting a new thread or question in the appropriate forum. The Feedback forum is overwhelmed with questions that are having to be moved elsewhere. If you need help to verify your account contact hello@boards.ie
Hi there,
There is an issue with role permissions that is being worked on at the moment.
If you are having trouble with access or permissions on regional forums please post here to get access: https://www.boards.ie/discussion/2058365403/you-do-not-have-permission-for-that#latest
There is an issue with role permissions that is being worked on at the moment.
If you are having trouble with access or permissions on regional forums please post here to get access: https://www.boards.ie/discussion/2058365403/you-do-not-have-permission-for-that#latest
Flickr Pro
-
06-03-2011 2:10pm#1
Comments
-
-
-
-
-
-
Advertisement
-
-
-
-
-
-
Advertisement
-
-
-
-
-
-
-
-
-
I'm with kev, I find flickr so easy to navigate, search on, upload, interact. They really need to improve the smart phone ap though. You can't view groups on there and it's fiddly to even view comments.
Pix is a bit bland IMO. Being that it's mostly Irish users you don't get that broader range feel, there's little interaction outside of comments and it's the same people commenting on one another's images all the time. That's fine, alot of friends on there etc but I prefer being just a number on flickr. There's something about someone from china quoting Keats on one of your images that's just really cool
.
I had a free account for years, I think the limit used to be size wise rather than amount of images. I remember 100mb at one stage being the limit. You'd get more in by uploading downsized images.0 -
-
Advertisement
-
-
-
I worry about security on pix.ie, sometimes it's useful for me to upload full res pictures and I think they're universally available on pix.ie whereas on Flickr they aren't. Could be wrong but thats my perception.
Hugh, this has changed as of last year...
When you upload to Pixie, you can select the maximum size of photo that different contact groups can view on the upload page. For instance you can say the Public can only see a max size of 1024px, while your Pixie Friends can view photo up to 1600px and then perhaps you'd like your Family to be able to view the original photos... It's very flexible and you can also retrospectively set these permissions for a given photo or for a given album (using the Share Album link on the album page).
With respect to security, Pixie image URLs are substantially more secure than Flickr. We both use secrets in the image URL to secure the images, but just look at the difference:
Flickr Image URL Format: http://farm{farm-id}.static.flickr.com/{server-id}/{id}_{o-secret}_o.(jpg|gif|png)
Example: farm4.static.flickr.com/3037/2609732632_3d526c90dd_o.jpg
Pixie Image URL Format: http://photos{farm-id}.media.pix.ie/{d1}/{d2}/{file-id}-{account-id}-{image-id}-{size}{orientation}-{o-secret}.{ext}
Example: photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-05616L-EE7A3F9DE5FE4392AD4BFBBEB9BF89A0.jpg
The Flickr URL could in theory be hacked in less than a day as the secret they use is so short. Furthermore, the Flickr URL for all sizes other than the original photo uses the SAME secret:
http://farm{farm-id}.static.flickr.com/{server-id}/{id}_{secret}_[mstzb].jpg
That means that if you know the secret for an 500px photo (everyone has access to this as 500px is their standard size for viewing), you automatically have access to the 1024px photo and so on by just creating your own image URL using the same secret.
With Pixie on the other hand, we use different secrets for every single size: 500px. 640px, 800px, 1024px, 1200px and 1600px which means that you can fine tune access to every single size we offer (and we offer a lot more sizes to choose from too
).
Flickr Sizes:
farm4.static.flickr.com/3037/2609732632_b72c17e185.jpg
farm4.static.flickr.com/3037/2609732632_b72c17e185_z.jpg (640px has same secret as 500px above)
farm4.static.flickr.com/3037/2609732632_b72c17e185_b.jpg (1024px has same secret as 500px above)
farm4.static.flickr.com/3037/2609732632_3d526c90dd_o.jpg
Pixie Sizes:
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-00500L-A174F77A23F54AB49E2C724F3050B364.jpg
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-00640L-C9506BF2589C4C3FA8ADA17952C80DEC.jpg
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-00800L-DA3A66BA0609495D96DEE248E7D8A30F.jpg
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-01024L-88000570397E4FB6BCF248FAEDD68A1A.jpg
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-01200L-7A9E52DD967A4415AC39ED64C701875B.jpg
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-01600L-CBBABC6A295F4DFD85FE415D44CF1150.jpg
photos2.media.pix.ie/6C/49/6C497D1AC44149018D004B1F88A93962-0000324999-0002194985-05616L-EE7A3F9DE5FE4392AD4BFBBEB9BF89A0.jpg
Marcus0 -
-
-
Assuming your secrets are made up of uppercase letters and numbers, that makes for a total search space of 36^32, which is roughly 6.33402867 × 10^49
or,
63340286700000000000000000000000000000000000000000
That's a lot
By way of comparison, the total (estimated) number of atoms in the ENTIRE EARTH is 1.33*10^50 or ...
133000000000000000000000000000000000000000000000000
which is just one order of magnitude greater. So your secret has practically enough addressable space that every atom in the entire planet could have a secret assigned to it :-)
A nameless place I used to work decided we needed 128 character random session ID's for a particular product we were developing, lower & upper case characters, numbers, and some small subset of special characters. I forget the actual figures but the total space far exceeded the total number of atoms in the entire universe. I thought that was a bit overkill.
Flickr's secret isn't SO bad. It's 36^10, which is ...
3.65615844 × 10^15
or
3656158440000000
which, assuming the only access you have is through their webservices interface, and each call takes 100ms (1/10th of a second) could potentially still take you
(((3656158440000000 / 10) /60) /60) /24
days to find the secret in the worst case, or
4.23166486 × 10^9 days = or 11593602.4 years, so not completely insecure.0 -
-
DaireQuinlan wrote: »Assuming your secrets are made up of uppercase letters and numbers, that makes for a total search space of 36^32, which is roughly 6.33402867 × 10^49
or,
63340286700000000000000000000000000000000000000000
That's a lot
By way of comparison, the total (estimated) number of atoms in the ENTIRE EARTH is 1.33*10^50 or ...
133000000000000000000000000000000000000000000000000
which is just one order of magnitude greater. So your secret has practically enough addressable space that every atom in the entire planet could have a secret assigned to it :-)
A nameless place I used to work decided we needed 128 character random session ID's for a particular product we were developing, lower & upper case characters, numbers, and some small subset of special characters. I forget the actual figures but the total space far exceeded the total number of atoms in the entire universe. I thought that was a bit overkill.
Flickr's secret isn't SO bad. It's 36^10, which is ...
3.65615844 × 10^15
or
3656158440000000
which, assuming the only access you have is through their webservices interface, and each call takes 100ms (1/10th of a second) could potentially still take you
(((3656158440000000 / 10) /60) /60) /24
days to find the secret in the worst case, or
4.23166486 × 10^9 days = or 11593602.4 years, so not completely insecure.
Nerd........ quick everyone beat him up!0 -
-
Advertisement
-
-
-
-
DaireQuinlan wrote: »Assuming your secrets are made up of uppercase letters and numbers, that makes for a total search space of 36^32, which is roughly 6.33402867 × 10^49
or,
63340286700000000000000000000000000000000000000000
That's a lot
By way of comparison, the total (estimated) number of atoms in the ENTIRE EARTH is 1.33*10^50 or ...
133000000000000000000000000000000000000000000000000
which is just one order of magnitude greater. So your secret has practically enough addressable space that every atom in the entire planet could have a secret assigned to it :-)
A nameless place I used to work decided we needed 128 character random session ID's for a particular product we were developing, lower & upper case characters, numbers, and some small subset of special characters. I forget the actual figures but the total space far exceeded the total number of atoms in the entire universe. I thought that was a bit overkill.
Flickr's secret isn't SO bad. It's 36^10, which is ...
3.65615844 × 10^15
or
3656158440000000
which, assuming the only access you have is through their webservices interface, and each call takes 100ms (1/10th of a second) could potentially still take you
(((3656158440000000 / 10) /60) /60) /24
days to find the secret in the worst case, or
4.23166486 × 10^9 days = or 11593602.4 years, so not completely insecure.
This is going WAY off topic, but I couldn't let the above go without giving it a different slant
Flickr's secret AFAIK is 32bit i.e 4294967295 rather than 3656158440000000 combinations which is significantly smaller... (Pixie's secret is a 128bit GUID).
Your calculation assumes 100ms to check each possible URL, but in reality we are only looking for the case where Flickr issues a redirect to their File Not Found (or http://l.yimg.com/g/images/photo_unavailable.gif). The hacker would not be interested in downloading any content, they just need to check the first few bytes of the response to identify a "not found condition" which will significantly speed up the process.
Your calculation also serialises the fetches (waits for one to finish before starting the next), not something a hacker would do. A hacker would spin up 10,000 threads and hit Flickr with as many requests as possible from each of the machines participating in the attack.
I've read that Varnish can handle up to 275K (inbound) requests per second on a single machine. But let's say it 10,000 for the sake of argument.
Wouldn't your calculation then become:
(4294967295 / 10000 / 1s) / 60 / 60 / 24 ~ 5 days
And that assumes you have to look at every possible URL, but you'd find the URL on average in half that time since you can stop the attack once the true URL is identified.
That gives us an average time of 2.5 days per machine. Using 2 or more mahcines brings this down to around 1 day.
Also... It's likely Flickr wouldn't even notice this traffic as the packet size is so small and the volume of traffic would be insignificant compared with their global live traffic.0 -
Pix.ie: Marcus wrote: »This is going WAY off topic, but I couldn't let the above go without giving it a different slant
Flickr's secret AFAIK is 32bit i.e 4294967295 rather than 3656158440000000 combinations which is significantly smaller... (Pixie's secret is a 128bit GUID).
.
Actually, yes, you're quite right. I'd assumed that the flickr secrets were just random 10 digit strings with lowercase letters & numbers. I stand corrected, sir
0 -
-
Advertisement
-
-
I was just about to say all that, but you beat me to it... :cool::DDaireQuinlan wrote: »Assuming your secrets are made up of uppercase letters and numbers, that makes for a total search space of 36^32, which is roughly 6.33402867 × 10^49
or,
63340286700000000000000000000000000000000000000000
That's a lot
By way of comparison, the total (estimated) number of atoms in the ENTIRE EARTH is 1.33*10^50 or ...
133000000000000000000000000000000000000000000000000
which is just one order of magnitude greater. So your secret has practically enough addressable space that every atom in the entire planet could have a secret assigned to it :-)
A nameless place I used to work decided we needed 128 character random session ID's for a particular product we were developing, lower & upper case characters, numbers, and some small subset of special characters. I forget the actual figures but the total space far exceeded the total number of atoms in the entire universe. I thought that was a bit overkill.
Flickr's secret isn't SO bad. It's 36^10, which is ...
3.65615844 × 10^15
or
3656158440000000
which, assuming the only access you have is through their webservices interface, and each call takes 100ms (1/10th of a second) could potentially still take you
(((3656158440000000 / 10) /60) /60) /24
days to find the secret in the worst case, or
4.23166486 × 10^9 days = or 11593602.4 years, so not completely insecure.0 -
-
Pix.ie: Marcus wrote: »This is going WAY off topic, but I couldn't let the above go without giving it a different slant
Flickr's secret AFAIK is 32bit i.e 4294967295 rather than 3656158440000000 combinations which is significantly smaller... (Pixie's secret is a 128bit GUID).
Your calculation assumes 100ms to check each possible URL, but in reality we are only looking for the case where Flickr issues a redirect to their File Not Found (or http://l.yimg.com/g/images/photo_unavailable.gif). The hacker would not be interested in downloading any content, they just need to check the first few bytes of the response to identify a "not found condition" which will significantly speed up the process.
Your calculation also serialises the fetches (waits for one to finish before starting the next), not something a hacker would do. A hacker would spin up 10,000 threads and hit Flickr with as many requests as possible from each of the machines participating in the attack.
I've read that Varnish can handle up to 275K (inbound) requests per second on a single machine. But let's say it 10,000 for the sake of argument.
Wouldn't your calculation then become:
(4294967295 / 10000 / 1s) / 60 / 60 / 24 ~ 5 days
And that assumes you have to look at every possible URL, but you'd find the URL on average in half that time since you can stop the attack once the true URL is identified.
That gives us an average time of 2.5 days per machine. Using 2 or more mahcines brings this down to around 1 day.
Also... It's likely Flickr wouldn't even notice this traffic as the packet size is so small and the volume of traffic would be insignificant compared with their global live traffic.
I enjoyed this. The idea of botnets brute-forcing the URLs for high-resolution images of cats is amusing to me.0 -
-
Advertisement
-
-
-
-
defo does yup, they also look a little less sharp. I compared the exact image on both one day when i thought my mind was playing tricks on me.
We (Pixie) introduced colour management support on resized images last year. You should not see any difference other than very slight (some would say indistinguishable) differences due to the different resize algorithms used by Pixie and Flickr.
These differences are a matter of personal taste, some prefer Pixie's more natural images and some prefer Flickr's over sharpned images. We are of the opinion that if you want over sharpened images then you can always sharpen them more before uploading.
Some people have also reported slight differences in shadow areas between the two sites. Again these differences are subjective...0 -
I think if you are seeing differences in images on pix.ie vs flickr it is likely due to the colour management setup in your browser of choice, no?
Read this - http://www.robgalbraith.com/bins/content_page.asp?cid=7-9311-9478
Honesty? I couldn't tell ya, its not really my area but both are side by side in the same browser and theres a notable difference, The flickr one is clearer and sharper.Pix.ie: Marcus wrote: »We (Pixie) introduced colour management support on resized images last year. You should not see any difference other than very slight (some would say indistinguishable) differences due to the different resize algorithms used by Pixie and Flickr.
These differences are a matter of personal taste, some prefer Pixie's more natural images and some prefer Flickr's over sharpned images. We are of the opinion that if you want over sharpened images then you can always sharpen them more before uploading.
Some people have also reported slight differences in shadow areas between the two sites. Again these differences are subjective...
Thats a bit of a "glass half full" view though, isn't it.0 -
If you've got any links handy I'd be happy (and interested) to take a look.Honesty? I couldn't tell ya, its not really my area but both are side by side in the same browser and theres a notable difference, The flickr one is clearer and sharper.
Hands up, yes it was glass half full talk.Thats a bit of a "glass half full" view though, isn't it.
But in all honestly, Flickr users have long complained of oversharpening and over saturation. I actually wrote the image resizing and sharpening algorithms myself as the off the shelf libraries available at the time didn't provide the quality we needed. We tweaked them over and over until we had better results than was available on Flickr. The fact remains that some images need more sharpening and some less. Finding the happy medium is very very difficult and one that will perhaps always evade us. It's also possible that Flickr have made recent enhancements which has put them ahead in the quality game, that's why I'm interested in the comparison you are making with your photos.
Thanks, Marcus0 -
Pix.ie: Marcus wrote: »Flickr users have long complained of oversharpening and over saturation. I actually wrote the image resizing and sharpening algorithms myself
i could be wealthy with some brains
personally i dont like over sharpened digital images0 -
-
Thecageyone wrote: »How are you comparing? Because I've just tried it here , side by side, the exact same image on Flickr and Pix, and see zero difference. Both enlarged on the dark backings. Maybe the default view appears richer on Pix when you use the dark background?
http://www.flickr.com/photos/cagey75/5491582015/lightbox/
http://pix.ie/cagey75/2195944/size/800
Looking closer the version on Flickr actually looks richer and slightly more vibrant.
We've made a change to the sharpening levels which are applied after an image is resized (more on this here). The following photos show the Flickr, Pixie (new) and Pixie (old) versions.
Let us know if you still see significant differences.
0 -
-
Advertisement
![http://farm{farm-id}.static.flickr.com/{server-id}/{id}_{secret}_[mstzb].jpg](http://farm{farm-id}.static.flickr.com/{server-id}/{id}_{secret}_[mstzb].jpg){kind=link}
Advertisement