if the problematic messages are "conflicting feedback", can we just make them less prevalent? they seem to come in annoying multiples which could perhaps be collapsed
@tripleee they generally have non-linked accounts (so blacklisting users won't help) but reuse the same username a couple of times in a row. If Smokey could blacklist a username in combination with a domain, we'd autoflag a lot more. I think it might even help for other spammers who use the same tactic.
@iBug I have been proposing various ideas around that topic myself but I think it needs more thought, the ones we have so far are not entirely well-defined and/or sustainable
@tripleee That's what makes it so cancerous. They're everywhere. They have blogs, they reply to bug reports, the file bug reports to gain SEO rank, they offer a scholarship for US students (?!)
if we could kill the pharma spam on sight we would nearly be out of a job so I'm thinking there could be some merit to concentrating aggressively on those alone
I'm working as a file manager. My company bought me three safes and assigned me three assistants. I'm busy most of the time and I can't manage the safes myself, so I need to let my assistants access them.
I want to allow all three assistants to open any of the safes independently, without the pr...
@tripleee Well, couldn't you vary the reason weight based on the amount of FP's and TP's already for that keyword in MS? That the more data you get on a certain reason, the more weight it gets?
@micsthepick Body - Position 1-134: <p>Try this. It's a normal travel pillow that converts to allow you to rest your entire body leaning forwards! www.thelapnap.com</p>
@gparyani I tried to implement that at some point but it got hairy, IIRC because the context where you want to ignore the rest of the phrase was lost somewhere along the way
Eh, ultimately whatever hackish solution we do use to prevent it wouldn't be worth the dev time required, especially given that it's currently working as intended
@tripleee but I support adding this as an alias :P
having an alias is nice but the problem is that "why" will still be picked up when you use it in a response which is commentary on a report and not a command
@SmokeDetector I could not understand the where or the why of this
Wow, advids uses a lot of different spam strategies. Copy an existing post + drop in a spam link, or use "hey im new here" with a spam link, or some more complex adding in a bit about explainer videos
@AJ (fun thing: at one point after having a green avatar for more than a year, i switched to one that wasn't green. people freaked out & someone photoshopped my avatar to be green and gave me that image to use as my avatar instead, so as to restore balance & order to the universe.)
@doppelgreener While the thing that I'm about to say probably isn't something that's relevant in this case, it's still something important that we all need to think about: if a particular website is being spammed using a wide variety of different styles, then maybe some of those posts are legit. Spammers are dumb; they don't change their ways. Large differences between posts may be an indicator that we've gotten it wrong.
@angussidney Thanks. I looked through some of the results, and they were copy+paste of posts elsewhere on the internet, but with a sentence thrown in linking to advids and/or its explainer videos. This new one ("hi guys im new here") is a stand-out, but man, who posts that in 2 entirely different websites simultaneously out of context.
[ SmokeDetector | MS ] URL-only title, bad NS for domain in body, bad NS for domain in title, bad keyword in body, bad keyword in title, +5 more: maleintalk.com/vidhigra/ by johnyates on askubuntu.com
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, pattern-matching website in body, +1 more: Keto Trim WEIGHT LOSE by user810048 on askubuntu.com
@gparyani I replied to the mod message and send another message via "contact us" link, but have got no response so far
@gparyani maybe not needed. It isn't very hard for me to answer a few questions on SO and get some rep. The only downside is that it requires slightly more effort and a significantly longer waiting time.
@tripleee es hat gebuggt, sie hat gebuggt, er hat gebuggt, es wird buggen, sie wird buggen, er wird buggen, es wird gebuggt haben, sie wird gebuggt haben, er wird gebuggt haben, es buggte, sie buggte, er buggte
@iBug you can find every kind, Romance languages tend to build long chains of "v de w de x de y de z" whereas Germanic tends to have vswingxungyishzchen with some fugen-particles (what's the English term, compounding particles?) between the parts
so we can't do what you guys did with those words trains, there's only so many nouns you can put next to each other so that they still make sense, and anything longer than 6 syllables (and more than 2 nouns) has to be hyphenated
@iBug a heuristic like "user name consists entirely of keys which are adjacent on the home row" would be fun to play with, but unfortunately probably not very useful
In computer science, an ambiguous grammar is a context-free grammar for which there exists a string that can have more than one leftmost derivation or parse tree, while an unambiguous grammar is a context-free grammar for which every valid string has a unique leftmost derivation or parse tree. Many languages admit both ambiguous and unambiguous grammars, while some languages admit only ambiguous grammars. Any non-empty language admits an ambiguous grammar by taking an unambiguous grammar and introducing a duplicate rule or synonym (the only language without ambiguous grammars is the empty language...
@AndrasDeak then they win, I'm almost certain I have never seen that in practice
"I will spend 10h a day in a scruffy internet cafe for a ridiculously low wage but dammit, I want to plug my custom keyboard into this computer I rent"
@tripleee if you look at it like this: if you spend 10h a day writing spam, you might as well do it comfortably, especially since you can't afford to treat your tendinitis