Work on Amazon Mechanical Turk, learn from the best, and have fun doing it. Join the crowd today!
Separate names with a comma.
Discussion in 'MTurk Help & Resources' started by electrolyte, Dec 4, 2017.
then why allow short AA timers in the first place?
I've been bitching about the new site since day 1... the redirects are just the fucking worst and losing preferences constantly is maddening. It seems like they have gone out of their way to make the user experience worse in every possible way.
I've seen it intermittently over the last couple of months. Just thought if happened when page didn't load properly for some reason. I'm probably wrong though!
oh god, when you return a job you now have to confirm the cancellation!!!!!!!!
If you scroll up, it was solved. It's the Turkopticon extension causing it.
Just saw this and I've been wondering if I was seeing things. Reminds me of a setting sun whenever I see it. Maddening. Happens every time I go to a new page on the new site.
Disable the extension and it goes away.
I just saw about the extension. I forgot about that extension totally. I had disabled all scripts and it was still happening and I had finally come to accept it was never going away. Glad to have it gone!
I can't find the place on the MTSuite in Chrome to change the site URL.
Please tell me if these will be cleared up:
-Finder doesn't connect to TO.
-There are no "logged hits" in Finder. I have just one huge list from top to bottom.
-Catcher is not inserting the 2 little orange boxes for pand'ing at the top of Finder
Is there a script that makes a search box on all the pages of the new site?
In a future MTS update, not sure if I'll bother converting to a userscript or not.
This whole update and new worker site is a big fat pile of poop. Oh I am sure there were bad apples, etc, but you can't tell me the majority of the mTurk community was "scamming" requesters. In fact, it is more true that requesters were scamming workers, paying 2 dollars per hour on some hits or less, promising big bonuses and the hit taking hours to make 5 dollars. You guys know what I am talking about. There are more "bad apple" requesters out there percentage wise than there are workers. I don't have any stats to back it up, but I am pretty sure that's why we have TO to filter out the crappy ones.
Anyway, this new update is a nightmare. They are taking away AA times so we basically have NO IDEA when we are going to get paid on a hit. It can be hit or miss. Some of us chose to priorities hits with low AA times over others with high AA times (batches) at the same time because getting paid faster is better than waiting up to 30 days to get paid.
Not only that, they've basically targeted scripts and are trying to make it as difficult as possible now. Intentionally or not, I don't know, but redirects from hits you miss are ridiculous. There's no way to get a hit ID from this process and have it saved into something like Panda Crazy, or many others. So miss that hit? TOUGH LUCK!! You ain't gonna put it in a script by hitting a button, you'll have to figure out how to go through and get the ID and pass it through. A cluster F on the next level. Thanks redirects.
That and some qualifications are being hidden, so we can't see if we've been qualified with a "Part 2" hit we can accept by qualification in some cases. It's just not there, but if you go back to the reg site right now you can see you have the qual.
I have never seen a bigger cluster F in my life. I've been working on mTurk for many, many years and while the interface was old and outdated, as well as the site itself, the truth is we managed to get it working the way we needed to.
We have no radical new things, nothing amazingly new, just a new "watered" down website that looks fugly as hell, that contains almost crap all information and is cumbersome to work with with the mass redirects. No amazing APIs to draw from, nothing ground breaking, just a... after all these years we get a version of mTurk that is SO simple that it becomes so annoying to use.
The fact they know we use scripts, and to make a redirect on a missed hit go to a search page and be unable to pull, makes me think they are trying to screw us over somehow. I used to make 60-90 dollars a day, and since I'm trying to adapt to the new worker website, I've made a total of 90 dollars since the start of this week adapting to using a site that won't give is the IDs we need to "queue" a damn hit. So far we have panda crazy able to grab hits from the hit description, expansion, but not directly from a hit that was missed, so if it isn;t on the list below the missed hit, then you're SOL.
As it stands there is NO WORKaround for this once you are on the main site and away from your scraper.
Edit: Currently the only workaround I can think of is to have it pull directly from a scraper, such as HIT Scraper, or MTurk Suites Hit Finder. A script for, let's say, Panda Crazy that enables a button to be added to get the ID directly within the scraper itself and add to Panda Crazy. Because once you're on an external page (mTurks redirects) there's no way to get the ID. That's all I can think of right now.
I'm making 0. There's no panda yet in the Suite. I never work directly from the site.
Are we not able to run pandas and queue HITs anymore, or am I missing something? I've been ignoring the mturk community, so I was absolutely not prepared for the switch.
Panda Crazy functions so long as you load the version from the worker URL and not the old mturk one.
Also regarding AA times, some old HITs like barcodes will redirect me to the old site where the AA time can still be viewed. Is it possible that the AA can still be extrapolated from HITs on the new site or do we know for a fact that's a hard no?
When you say that Panda Crazy functions, can I ask what you mean exactly? Can you still the button to create a Panda when there are no more HITS available? I've been unable to figure out how to do that and it's a critical function. Any help appreciated.
I have always just manually pasted group IDs directly into Panda Crazy's add job prompt and I'm still able to do that even though they're called something different now and all my old ones still work. I just use Panda Crazy as standalone script in Chrome. The 0.5.3 update of Panda Crazy made it so the old site still works, but the PC URL for the new site is https://worker.mturk.com/?filters[search_term]=pandacrazy=on.
And by "Panda Crazy functions" I just mean that I'm able to get it to work correctly.
Thanks for the reply. So they're not called group ID's anymore? It's still the long code in the middle of the link address, correct? Why would you have done it manually when there was just a button you could click? Frustrating.