Skip to content

Commit d27bbbe

Browse files
authored
Update README.md
1 parent 473df2a commit d27bbbe

File tree

1 file changed

+21
-5
lines changed

1 file changed

+21
-5
lines changed

README.md

Lines changed: 21 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -10,13 +10,29 @@ Download the Lite version if your browser or user script manager does not allow
1010

1111
This is a userscript. You must install it with a userscript manager (TamperMonkey for Chrome or Greasemonkey for FireFox).
1212

13-
### MTurk Engine only works on this page: https://www.mturk.com/mturk/findhits?mturkengine
13+
### Mturk Engine only works on this page: https://www.mturk.com/mturk/findhits?mturkengine
1414

15-
## Why use Mturk Engine?
15+
## What is Mturk Engine?
1616

17-
Imagine setting up a watcher to automatically accept a rare HIT when it shows up while simultaneously being notified of any new HITs as they're posted. Imagine then quickly switching to a new HIT, clicking the accept button, and instantly being notified of wther or not your accept went through without having to wait for your browser to refresh or open a new tab. That's what MTurk Enging allows you to do. All in a single tab -- without any page refreshes.
17+
Mturk Engine is a user script that enables you to be more productive on Amazon's Mechanical Turk work platform. You install it with one of the download links above and then navigate to https://www.mturk.com/mturk/findhits?mturkengine while logged in to use it. Mturk Engine brings everything you do on Amazon's Mechanical Turk work platform into a single page. You can search for HITs, accept HITs, add HITs to be auto accepted, monitor your queue, and return HITs all in a single browser tab without ever having to wait for a to page reload.
1818

19-
MTurk Engine brings all of the things that make for an efficient workflow into a single place. You can monitor new HITs in the background, have instant feedback as to whether your accept request went through, block requesters and HITs you're not interested in, set up watchers to automatically accept rare and high-paying HITs or hoard lots of the same HIT, and quickly check your queue when you need to, and never have to wait for your browser to refresh a full page.
19+
Scroll down to the animated gifs and get a sense of what Mturk Engine looks like and what you can do with it.
20+
21+
## Why use Mturk Engine? What does it do that HitScraper/HitFinder & PandaCrazy don't?
22+
23+
### TL;DR: Mturk Engine is a combination of HitFinder + PandaCrazy with queue monitoring, instant accepting of HITs and instant returning of HITs, wrapped in a high-performing, easy-to-use, and accessible UI.
24+
25+
Mturk Engine in its current state can be described as a combination of HitScraper + PandaCrazy, with a few extra features like adding a watcher directly from your search results (watchers automatically accept HITs for you). Scroll down to the "Preview" section to see what that looks like.
26+
27+
You can also accept HITs and know immediately if your accept went through without having to wait for a full page reload and being disappointed when it didn't. Mturk Engine is also super fast, and should handle 0 second delays between searches and 1 billion+ blocked requesters/HITs without slowing down. From what I've seen of the source code of other tools like HitScraper/HitFinder, they use regular expressions to block requesters/hits, meaning that performance will degrade quickly as your block lists get larger. This is not a worry with Mturk Engine.
28+
29+
Mturk Engine uses [Immutable.s Maps](https://facebook.github.io/immutable-js/) under the hood to do blocklist lookups in O(log32n) time for each HIT, meaning that you don't have to worry that your blocklists are slowing things down and you can keep them growing infinitely.
30+
31+
Mturk Engine uses React to update elements on the page, which means that rapidly updating pages in response to changes is faster than other tools. I spend a lot of time making sure Mturk Engine is performant, and [write about some performance related stuff on my blog](https://blog.shovonhasan.com/pattern-for-rendering-lists-of-connected-components-with-react-redux/).
32+
33+
There's also some other technologies used, like [redux-saga](https://github.com/redux-saga/redux-saga) which allows Mturk Engine to simultaneously run your main search while running a bunch of watchers in the background to automatically accept HITs on <1 second intervals (if that's what you need) for maximum productivity.
34+
35+
Imagine setting up a watcher to automatically accept a rare HIT when it shows up while simultaneously being notified of any new HITs as they're posted. Imagine then quickly switching to a new HIT, clicking the accept button, and instantly being notified of wther or not your accept went through without having to wait for your browser to refresh or open a new tab. That's what Mturk Engine allows you to do. All in a single tab -- without any page refreshes.
2036

2137
Mturk Engine is under active development so feel free to suggest features or submit changes, no matter how big or small.
2238

@@ -34,7 +50,7 @@ Mturk Engine is under active development so feel free to suggest features or sub
3450

3551
###
3652

37-
### The MTurk Engine UI
53+
### The Mturk Engine UI
3854
<img src="https://i.imgur.com/JE0M8RB.png" alt="Still image showing the UI for searching for HITs, seeing unread HITS, hiding HITs, and blocking requesters."/>
3955

4056
### Search for HITs and block requesters.

0 commit comments

Comments
 (0)