Very hard to tell on this one but ultimately seems to be spam. Checked the trust badges, etc., but the whole "TrustSeller" thing on the bottom just doesn't seem to be real.
@Catija I love this room. It has changed a lot since I first started coming here. It used to be so quiet, you could hear a pin drop.. then the rabble gradually came in and well... :) waits for someone to reply
@SmokeDetector the OP has legit accounts on the network
@Catija I still forget basic commands and have to check the docs or ask people. There's so many bots I can't keep up with them all, in fact I have about 8 bots muted on the network
@tripleee That pattern looks like it's already caught by Bad ns for domain in body, Bad keyword in body, Bad ns for domain in answer, and Bad keyword in answer; use !!/blacklist-website-force if you really want to do that.
@TetsuyaYamamoto I had some snippets auto-translated. This was clearly spam.
Even without translating it, the likelihood that such a post is not spam is negligible. The only reason why I autotranslated is that I am mod on that site and my flag carries a higher weight..
"In DAIBIO, especially the Family Health Care Handbook, which combines the treatment of East and West Medicine with GS. Duy Duy Ban, Master of Vietnamese Studies and Master Le Khanh Linh and many enthusiastic scientists have been printed at the Medical Publishing House with over 4,000 pages."
Vietnamese currently is not part of any SE language site (English, French, German, Spanish, Portuguese, Chinese, Japanese, Russian etc.), its letters & accents can be optimized to detect spam posts.
you can detect specific accent combinations which are typical for Vietnamese but regardless of techniques, do we want to go to great lengths just to detect one of the approx 8,000 languages which are not supported?
the ones I have tried are not very good at distinguishing between related languages like Danish/Norwegian or Serbian/Macedonian but we probably don't need that level of precision as long as we can tell whether it's Japanese, Russian, Portuguese, Spanish, German, or English (did I miss any?)
@Ferrybig Spanish would have a moderately high percentage too, the words are longer so the percentage will be lower than for a monosyllabic language where most vowels are accented
@tripleee do we want to go to great lengths just to detect one of the approx 8,000 languages which are not supported? – We already detect most of those languages by the non-Latin filter.
@Wrzlprmft probably not "most" but yeah, a filter which detects many languages we don't want is better than a filter which detects a single language we don't want
It would probably easiest to detect texts with a large proportion of Latin characters not in the English alphabet outside of language and localised sites. That would spot Vietnamese, French and Turkish spam too.
@Wrzlprmft I still finding complete list of them. However I know some languages which uses special diacritics and accents of Latin alphabets (mostly in Europe).
@AJ That pattern looks like it's already caught by Pattern-matching website in answer and Pattern-matching website in body; use !!/watch-force if you really want to do that.
@TetsuyaYamamoto However I know some languages which uses special diacritics and accents of Latin alphabets (mostly in Europe). – That’s exactly the point, isn’t it? We can detect many non-English languages by their selection of special characters. And non-English is an indicator for spam (except on those sites where it isn’t). Also, actually, the further you get away from Europe, the more special characters are used (which is to be expected sociolinguistically).
Probably would be easier to have a system that detects all non-standard Latin characters and has a sitewise whitelist. So, allow äöüßÄÖÜẞſ for German, allow áéíóúç… for French, etc.
(And of couse: Do not get excited about one special character.)
@Wrzlprmft it's not like Hawai'ian or Swahili use a lot of special characters, it all depends on the requirements of the language. Even within European languages, different conventions occur -- compare Czech diacritics to Polish digraphs for encoding essentially the same sounds (cz ~ č, sz ~ š, etc)
@SurajRao I created an account and it gave me the association bonus... so it must not matter either way... but it's frustrating when every other site does this:
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, repeated URL at end of long post: How Does AlphaMax 10 Work? by krhilchri on superuser.com
[ SmokeDetector | MS ] Bad NS for domain in body, blacklisted website in body, link at end of body, pattern-matching website in body: Probulus is works? by gloriailliams on workplace.SE
@Mithrandir That is inconsequential - a valid spam deletion carries a much heavier penalty for post bans and the like then heavy downvoting, and outside of the -100 rep you get the -12 from 6 downvotes are refunded when a Q is deleted
So when you have 500 rep and something of yours gets spam deleted, you always have 400, regardless if a mod spam flagged it or 6 users did
