My ambitious next goal for this project: Needy Question Finder

[viziionary]

This is something that I could quite easily do server side, but I will (once we complete the current things to do) be testing a huge very ambitious feature for this user script, something that is not usually (ever) done via the client side:

I'm going to add a client side crawler that looks around the site for questions tagged with your favorite tags which it deems to be in need of an edit. This will be awesome.

I haven't decided where on the screen the interface for this will go, but it will be minimal, attractive, and intuitive. This will mean activating the script on all SE pages (I think we should expand to support all of SE, it'll be useful everywhere via spell checking being a thing soon, and we can add further rules) and simply switching on the editing functions when the user is at the proper locations. I'm open to suggestions, but I'm thinking this will take the form of a small bar (with 25% opacity) attached to the bottom of the user's screen. This seems to be the least-intrusive place to put it, and the user can simply hover over it and will be able to choose a question to go edit.

The interface will feature (in the style of SE GUI):

• Notifications showing the number of edit-needing questions (answers possibly included in crawl).
• a list of questions that need editing
• the number of edits required on each post

It will work as follows:

• Get the user's username from the SE head bar
• Grab his favorite tags by crawling his info page
• Begin crawling, at a controlled speed as to not slow anything down, until it's queue (25) is filled with questions (from his favorite tags) to be edited
• Stores all information in browser storage so as to not repetitively search the same thing.

Thoughts, comments, improvements, ideas?

[viziionary]

By the way, I think when we finish all current fixes and features, before moving onto this, it would be a good time to go ahead and re-factor into a CommonJS modular system, working on the code as modules and using one of those APIs to load it into a single file.

The new feature idea will make the source too large to work on effectively as it is.

[AstroCB]

I've wanted to write an editing bot since I wrote this userscript initially (see opener.js), but I did not trust it to do anything automated until now (it messed with code too much). This is a good idea, but I'm not sure if we'll know how feasible it is until we try it.

[viziionary]

@AstroCB Not an editing bot. I agree that's not safe nor really necessary. Only a tool that seeks out questions that need edits.

[AstroCB]

@jt0dd Yes: I understand your intent. I was just mentioning what I had been thinking about for its future.

[viziionary]

@AstroCB this is feasible fundamentally, but it's big. I'll need help. Once I lay out the infrastructure for it here, we can divvy up the features. But first, spell checking & in-line support are priority.

[Unihedro]

You could work on this after you finish every burnination request on Meta with the script.

[viziionary]

@Vincentyification what?

[Unihedro]

http://meta.stackoverflow.com/questions/tagged/burninate-request
http://chat.stackoverflow.com/rooms/63635/tag-burnination
http://meta.stackexchange.com/questions/239190/when-to-burninate/239191#239191
http://bit.ly/burnguide

[viziionary]

Is that really a valuable functionality to add? I'm a mediocre SE user, you'll have to cut me some slack. Because that can be automated quite easily, but I'm not sure whether it fits here. The fact that SE or someone didnt already automate the process is beyond me, or at the very least stream-line it.

[viziionary]

They call it a mindless process. Will it bother anyone then if I (rather than giving users the power to do it in the user-script) just write a Node.js crawler that removes all tags listed for it in a matter of hours and package it here as an editing tool?

[AstroCB]

@jt0dd I already have written a couple of userscripts for tag burnination, but my rationale behind not sharing them is that they are pretty dangerous if used improperly (whereas here a mis-edit can be easily rolled back). I believe @Vincentyification's point was that this project could be used by someone who is mass retagging posts to remove tags as it makes cleaning up posts easier, which is part of the burnination process.

[AstroCB]

In fact, that's what opener.js is for: after I finished the original version, I wanted to be able to use it in tandem with my other burnination scripts.

The function in that file should (it, for reasons I'm not sure of, did not work at the time and definitely won't now that inline editing is disabled) open the post whose URL you pass to it, click the edit button and press 'Fix' (activating the main script), submit the edits automatically, and then close the window.

[viziionary]

@AstroCB that sounds dangerous, at least if used for general editing. Here's what I have in mind (and am designing the interface for in Photoshop now):

Spell-checking can get a lot of false positives if the score isn't set to above 0.876 or something above that, but at that level it also misses quite a few fixes. To solve this,one tab in the small interface unit is going to (first, it goes through a 3-part icon animation to show the dictionary's loading process, see relevant icons here: http://imgur.com/KkARbBu,kcjCFVI,bsDqfip,xOStKCi,7SsyOnb,YbdzJgn,4SuHaOs,etPit20) blink to an active state when an edit is made to say basically "Hey, I'm doing stuff, click me if you want".

When the user clicks that, a list slides upward from the bar allowing the user to check or un-check words that have been / can be changed, or change his spell-check setting to higher or lower than the default 87%. That way users can get maximum efficiency from spell-checking if they want to, can disable it altogether, or can let it run automatically at default settings.

There is also a main tab which works sort of in the same way, showing what changes the core editing function has made and allowing the users to check/un-check those changes.

The last two buttons (this is fixed to the bottom of the screen to be the width of the sidebar, and about 50-70 pixels high unless a list is expanded, so it has room for 2 more buttons) will be:

Needy Question Finding - again, a list of 25 questions that need significant editing based on your favorite tags on the current site. No favorite tags = no needy questions.

And I don't have any ideas for the fourth button, so, if you guys want, it really can be an auto-piot burninator. That sounds like something that editors for every single site on the network will be interested in at some point.

So, how that works is obvious. Crawl all questions with that tag and the system goes into auto-pilot mode OR 1 by 1 mode, where the user can just click, click, click. Auto-pilot burnination will only occur in sprints of 25 questions. There will be a key that stops the auto-pilot mode of course besides using the button.

If they mis-use it, there's a system in place to handle that. People > 2000 points will likely not go rogue on the site though. And if they do, well, they could've done the same thing in a few hours anyway. Not to mention writing such a program is extremely simple and anyone could do it, at the cost of losing editing rights or being banned on an account with >2000 score.

Thoughts, ideas, etc?

Remember all of this is handled via a minimalistic, intuitive, non-intrusive small bar attached to the bottom right side of the site's safe area, in-line with the side bar.