Happy birthday



On the April 1st, 2019 [no joke] I wrote the first lines of code for Stereopix.
I started by some proofs of concept to check if I would be able to display in different formats and how to show a nice list of thumbnails. The code was improved a lot, but a large part of what I did in early April 2019 is still used today.
It took a few months [on spare time] to have something barely usable and to install everything on the server so that the first "somewhat usable" version was really online in early July 2019. It really took off after a laconic message of Mike Bittner on this mailing list the January 17th, 2020.

By the way, if you want to help for the rent of the server, it is a good timing because it is due in a few weeks for getting Stereopix online for the next year.
The link to the tip platform is on the login page for example (or photolab page if you are connected).
Please, tip only if you can afford it: the project is not fully sustained yet, but there is no urge because I can complement for the moment.

For this anniversary, let's share some usage data.

Here is the amount of data shared on the platform by day.
Each dot represents the number of data added to the platform by day in a logarithmic scale.
It can be a new photo, a comment on a photo, adding a photo to an album or highlighting a photo. The actions that were deleted are not counted here.

Here is the amount of data served by the server.
A hit is a URL asked to the server (it may be a page, a photo, a script and so on).
It does not necessary correspond to all the actual views because some documents are cached in the browser of the visitors and those browsers are allowed to not ask for the resource again if they already know it.
A visit is considered as a group of hits sharing the same IP address and same browser the same day.
However, a lot of robot are crawling the Internet.
For example, the search engine websites use robots to index the pages (I saw Google, Bing, Yandex, Qwant... in the logs)
By trying to remove them, we got more accurate data for real usage as shown in the next image.

We can "zoom" on the active part between the January 17th and today (without crawler robots).
On average on this period, there are 3342 documents loaded per day.
Since the images already viewed are most of the time not loaded every time, the average is "only" of 100MB per day.

To finish, here is the geographical spread of the visitors based on their IP (not highly reliable though):


Join to automatically receive all group messages.