added a comment - - edited
if we choose to hide a user's tasks based on their {{user.show_hosts} flag, then this will make the search less useful, and perhaps eliminating the raison d'etre of this search feature.
I don't see why that should be the case: when someone asks for help, it's certainly no problem to ask that person to un-hide the computers. If we don't respect this setting, we could as well just remove it entirely because it's very easy to derive or plain guess tasks names and thus get access to otherwise private data. I'm strongly against that, in particular since asking the person seeking for help for cooperation should be a no-brainer. Again, it's a choice the "data-owner" has to make; they need to consent.
I believe Richard wrote about a second use case for the task search, the 'anonymous' accounts/hosts that Science United creates have no way of determining user or host information.
Ok, this one is far better justification and we should give this some further thought. OTOH, SU really seems to be broken in a number of ways as its design is apparently based on various assumptions about the downstream projects that simply don't stand the test of reality, particularly in the light of GDPR.
My answer would be no: if you can only search for users with show_hosts=1, there seems to be little point because there is already a way for people to know about their own tasks.
That's a good point you make there. Yet I remember seeing one comment where someone even had a hard time finding his own task in the various tables. So maybe adding a filter or search to the current task overview would improve things. I think this is what BOINC does as well...? OTOH, if total newbies don't know their way around our website yet, letting the community help by task name could have some benefit. But I honestly don't know how often that's the case and hence whether it justifies the effort and potential risks (see previous post). While in doubt, I doubt it.
See Richard's follow-up. https://einsteinathome.org/content/implement-find-task-name#comment-168175
Oliver Behnke Is there a way the client would know the task id (result id)? Or is he correct that the id is server-side only?
I think we discussed that the technical implementation of such a search is easy: it is just a SQL SELECT. But that there may be externalities to having such a database search on our Web site. But what if the form was only available to logged-in users and had a re-CAPTCHA on it?