I took a new lesson, SwedishLinqQ Beginner (Eating Out), Part 2
It said that there were 8 New Words , so out of curiosity I checked. I saw there were 13 words highlighted so I made sure I didn’t count duplicates and the total came to 7 ?
I LingQ’d the unknown words and after they were all selected the remaining New Words said 1 which confirmed my count (8-7=1 :P). I did notice that klockan was in the script 3 times but 1 of them had a capital letter so wasn’t sure if that was the reason, also there was a number ‘3’ that was neither underlined or highlighted so am not sure if this is another possible reaon or not. Just feels that the known words count if you hit the blue button might mean your total ‘knowledge’ may be incorrect. If you’re the sort to shout it from the rooftops and say ‘Hey I know 2000 Swedish words !’ and it turn’s out you only know 200 (exaggerated for effect) ,it might make you think you are better than you really are.
Is this a known/unknown issue or is it my lack of understanding on how it works ?
There do seem to be some slight inconsistencies that have cropped up in the New Words number. We are aware of it and are looking into it. We should have it fixed soon. However, I wouldn’t worry too much about the discrepancy. Obviously, the amount the count is off is very slight and should not in any way prevent you from screaming your Known Words total from the rooftops. Or, at the very least adding your badge to your email signature, Facebook page or blog!
LingQ is continuously evolving. Things sometimes go a bit Beta round the edges, but the experience is consistently good.
As regards known/unknown words, we have regular debates about what constitutes a word, whether “wordness” differs across languages, and at what stage you can consider vocabulary “known”. The whole area of semantics seems a bit loosely defined to me, so I prefer to think of LingQ as operating on fuzzy logic. Like a lot of the members
[Watch as I now get flamed by all the computer programmers who will explain to me that fuzzy logic isn’t just getting the algorithm a bit wrong!]
Us computer programmers are generally introverted and we usually don’t flame when we can’t do it anonymously.
That said, the simplest explanation of fuzzy logic is a result does not have to be exact in order to be considered true. Rather than getting an algorithm a bit wrong, its about deriving an algorithm that will give you an answer within the range of acceptable answers, which, if you argue your case right, could apply to the “Known Words” statistic.
