Some of the HITs posted by these accounts come from Google employees, some from academic researchers working with Google on a project. Most of the Google-affiliated requesters' HIT content is hosted on a Google-owned external server that has 'endor' in its URL (use Is this on endor? to help spot those).
Google's batches often have a hidden limit enforced by their server software. When you've reached your limit for that batch (which may be anywhere from 1 to hundreds, and may vary by a few from one person to another), or if you've done an overly-similar survey before from this or another Google requester, you receive this message after accepting one HIT too many:
This will continue to be the case each time you try to accept another HIT in that batch, until the requester resets the hidden limit - could be hours or days later, you never know. (Since mid-2015, some of the Google accounts' batches have started randomly experiencing temporary 'false caps', and returning one with that message might or might not allow you to accept another working one after it.) Returning one that is displaying this message doesn't trigger a lockout.
Another message you'll sometimes see from Google-hosted HITs is
Reasons this happens can include:
Returns or expirations on HIT assignments that have only ever had the "An error occurred." screen or the "reserved for other workers" screen displayed to you don't trigger/extend lockouts, only on ones that you've had a chance at actually working on. (A working HIT can have been displayed without you ever having actually seen it if you've been pandaing in the background, though.)
Here's what mmmturkeybacon added about that on MTG in Jul 2015:
~~~
The 'Sergey Schmidt' account's name comes from the names of Google co-founder Sergey Brin and chairman Eric Schmidt. The 'Violet Parr' account is named after a character in the CG animation movie 'The Incredibles'. Other Google requester accounts with human-like names are usually the names of real Google employees involved in that project (findable on LinkedIn).
Some projects our work has contributed to include:
Google's batches often have a hidden limit enforced by their server software. When you've reached your limit for that batch (which may be anywhere from 1 to hundreds, and may vary by a few from one person to another), or if you've done an overly-similar survey before from this or another Google requester, you receive this message after accepting one HIT too many:
"Thank you for working on our HITs! Unfortunately, all of the remaining HITs in this project have been reserved for other workers, and there are none left for you to do. Please check again tomorrow for more HITs to be available. (You must click the "Return HIT" button now to return this HIT.)"
This will continue to be the case each time you try to accept another HIT in that batch, until the requester resets the hidden limit - could be hours or days later, you never know. (Since mid-2015, some of the Google accounts' batches have started randomly experiencing temporary 'false caps', and returning one with that message might or might not allow you to accept another working one after it.) Returning one that is displaying this message doesn't trigger a lockout.
Another message you'll sometimes see from Google-hosted HITs is
"An error occurred. Please refresh the current page in your browser."
Reasons this happens can include:
- If you return one of their HITs, you'll get that error on any future HITs from them until the time when the returned HIT would've expired.
- If you accept more than one of their HITs simultaneously, you'll get that error on all but one of them (look through your queue and find the working one); then after you complete that one, another one will become the working one once you refresh it.
- And you'll get the error sometimes 'just because', even when you aren't in one of those situations; in that case, right-click inside the HIT iframe and select 'Reload Frame' (or use this or this), and it will usually work then.
Returns or expirations on HIT assignments that have only ever had the "An error occurred." screen or the "reserved for other workers" screen displayed to you don't trigger/extend lockouts, only on ones that you've had a chance at actually working on. (A working HIT can have been displayed without you ever having actually seen it if you've been pandaing in the background, though.)
Here's what mmmturkeybacon added about that on MTG in Jul 2015:
"It has to do with how they are accepted. The HIT that loads the iframe contents first is the active one. The active one is not necessarily the first one you accepted. So if you hit F5 a bunch of times on a panda link you will accept several without loading the iframe contents, then you might let one fully load, and then panda some more, which would put the active one somewhere in the middle. If you autoreload the panda URL at a rate that lets them fully load then the active one should be first. If you use Turkmaster, then none of them will be active until you look at one of them from your queue."
~~~
The 'Sergey Schmidt' account's name comes from the names of Google co-founder Sergey Brin and chairman Eric Schmidt. The 'Violet Parr' account is named after a character in the CG animation movie 'The Incredibles'. Other Google requester accounts with human-like names are usually the names of real Google employees involved in that project (findable on LinkedIn).
Some projects our work has contributed to include:
Last edited:
I can shuffle posts around when needed here, but it's less hassle to just make some extra posts up-front, and I went with that amount arbitrarily (1 is just a few weeks away from being used, and 10 more as a round number that didn't seem too ridiculous).