At C:\ProgramData\BOINC>all_projects_list.xml

the gpugrid project (there are others)

needs to have
<url>https://www.gpugrid.net/gpugrid</url>
instead of
<url>https://www.gpugrid.net</url>

Is there a way to update this w/o having to edit the file? I looked at 8.0.2 and did not see a project list update option.
I do not want to reinstall BOINC to get the latest project list.
The option "modify" is grayed out for BOINC in the windows add / remove applications. That would be a nice way to update a project list.

Possibly the projects do not bother to inform anyone of their URL change other than popup 100 of notices to the user.

I assume I would still have to detach and then re-attach to the project.

After 20+ years since the program inception, I think this should be an automated process that BOINC updates the all_projects_list.xml to add/remove/update project url's without user intervention and without shutting down, updating & reinstalling the whole thing.

BOINC already knows/checks if there's a new program version available to download periodically.

At least every 3 months or preferably a monthly check for outdated project links can't be that hard to implement.

Reading between this thread, the git-hub one and the one over at CPDN I think the problem lies with the projects rather than with BOINC and what is required is for the named project administrator to tell David or whoever deals with keeping the project list maintained that the url has changed. It makes some sense that they won't change it on the basis of some random (i.e. not verifiably part of the project in question) person telling them it needs changing. That said, if the project administrators for want of a better description don't do that it becomes frustrating.

They were surprised their server couldn't handle the load. [rolls eyes emoticon] I guess Oxford uses old 486s form a charity shop.

There is a difference between using data for genuine scientific work and harvesting the whole database to train AI models. The sooner someone hits those companies with a massive law suit for unauthorised use of data the better.

Glenn is going to speak to Andy about the CPDN url. With regards to others, how frequently do project staff peruse their own forums?

In reply to Lucas Dobre's message of 19 May 2025:

I disagree - I'm completely against suing and I'm completely against any restrictions on using data. And AI *is* scientific.

If a server can't handle the download of the whole database in the 21st century, there's something drastically wrong - it's a university, well known for having the fastest internet connections known to mankind. If the server is that slow, why not limit transfer speed per IP or something? It's not rocket science.

That comment shows a lack of understanding of the problem. This is nothing to do with the capacity of the network. We (CPDN) were essentially suffering from a denial-of-service attack. I watched the server logs and the database was getting hammered by requests coming in from multiple URLs through various pages on the boinc website that needed a database request.

The 'data' the server contains is not the results of the batches, that goes to the scientists. The only data stored concerns the configuration of the batches and their outcomes (fail/success), plus data on volunteers.

Glenn, Visiting Scientist/ CPDN.

---

---
CPDN Visiting Scientist

In reply to Dave's message of 19 May 2025:

Glenn is going to speak to Andy about the CPDN url. With regards to others, how frequently do project staff peruse their own forums?

Dave, I was only going to talk to Andy regarding the problem Mr Hucker had with connecting to the project. But this is actually a problem with BoincTasks, it works fine with boincmgr. Nothing to do with CPDN or the URLs the project uses.

---

---
CPDN Visiting Scientist

You need to make it work with boinctasks

That is your opinion. Mine is that boinctasks need to deal with it.

Just because other projects have not posted about what has the same effect as a deliberate DDOS attack does not mean they have not suffered from this. I know for a fact that some other projects have been affected but openly saying so has to be up to them.

Does the call for being logged-in detection form part of the extended BOINC API?
If it does then it is down to BoincTasks to comply with the API.

I can put it in a different way: if the official BOINC Manager works with the Project but BoincTasks doesn't - that's the problem of the BoincTasks, not the Project.

---

BOINC maintainer.
For any insight, check my BOINC Development Blog.

In reply to Lucas Dobre's message of 21 May 2025:

In reply to Vitalii Koshura's message of 21 May 2025:

I can put it in a different way: if the official BOINC Manager works with the Project but BoincTasks doesn't - that's the problem of the BoincTasks, not the Project.

Consider a boinctasks user (there's a lot of us). We connect to 5 or 10 projects. One fails, we run the others instead. CPDN doesn't get the use of our computers, simple. We just don't need a legal agreement there. Hide it in a webpage somewhere else nobody will bother reading it.

Analogy: Projects are competing for our processing time, just like supermarkets are competing for our purchases. Make something more difficult for the customer and they switch.

You are shifting the focus from the application (BoincTasks) to the project, and this is not correct.
There is an API that BOINC provides and Projects support.
If 3rdparty application doesn't support a certain API that is available for years on both BOINC Manager and Project - why should project care about this buggy application?
You are blaming a wrong person.
Push BoincTasks application developer to fix this issue. If the developer don't care - maybe you should then use another application that is compatible with the official BOINC API?

---

BOINC maintainer.
For any insight, check my BOINC Development Blog.