Re: Phereo - revival and community support


As of right now I have access to the AWS S3 buckets containing all the images currently hosted on Phereo. There are also EC2 instances running the site and db to which I also have access.

Nice! This is a very big step toward a revival.

The question is what can we do with this?

First, do emergency repairs to come back to a website that can be visited.
I think that it is only a matter of renewing the certificate for the "api" domain and make the http server use the new certificate.

Then, a diagnostic?
One thing I'd like to know is whether the images published since 2019 are really lost or just not accessible + why (routing errors? bucket limit?)
Also, I wonder which services are running apart from the web frontend and api interface.

After the diagnostic, we can decide if it is easier to repair and maintain, or migrate the images and the database to another system.
Also, could the AWS products be transferred with data, or should a new system be created anyway for the new organization?

For parts that need changes in code, I see the upload system to avoid flash and the mailing system that failed.
An automatic renewing of the certificates could be worth too.

Depending on the amount of work, maybe only freezing the site could be an option. (better than current situation)

If there are more engineers in the crowd who are happy to spend some time figuring out this legacy project, let's have a catch up.

As I said before, I am okay to help, though I have less time now.
You can contact me in private if you wish.

As for image ownership and changes to the site. I think a message on the site informing of the change asking image owners to delete any images before some date they don’t want to be transferred might be sufficient.

This solution might be the less time-consuming option, but it is better to double the message with a mailing to every user and a way to recover their password.

it is amazing that StereoPix is developed by a single programmer

I think it is a little easier nowadays with frameworks (user management, database access, template system, some security tools, and so on were already coded by awesome guys better than me for the site, and a lot of building blocks for the viewer too)
Sure it still remained a lot of work (~6000 lines for the website, ~3000 lines for the viewer [single line of code are not a fair measure, but it is quite easy to count]), but without deadline and moreover without needing to pay the large amount of time spent, it was achievable [I studied the feasibility deeply and coded a proof of concept before starting for real]


Join to automatically receive all group messages.