"Just posted “Ludwig: a type-based declarative deep learning toolbox”, a paper describing the design of #Ludwig and the motivations behind its architecture: https://arxiv.org/abs/1909.07930"
The second issue is that the split() function takes an optional "Limit" argument (split /PATTERN/,EXPR,LIMIT)
This causes the example code (split /-/,$str2, split /;/, $str4) to be evaluated as split /-/,$str2,split(/;/, $str4) which evaluates as split /-/,$str2,2
Here, split /;/,$str4 evaluates to 2 because, in scalar context, split returns the length of the resulting list - since there are two elements after splitting $str4 it returns 2.
In this case, the result of splitting $str2 results in two elements so the limit doesn't matter.
If $str2 were 'This-string-has-many-dashes', the result of (split /-/,$str2, split /;/, $str4) would be:
array 3[0] is: This
array 3[1] is: string-has-many-dashes
array 4[0] is:
array 4[1] is:
> I saw a paper go by years ago, on someone who seemed to have outlier excellent memory retention. They could tell you what they did on arbitrary days decades past.
It's called Hyperthymesia [0], the actress Marilu Henner has it and "can remember specific details of virtually every day of her life since she was a small child" [1]
Agreed. I use those features multiple times per day, along with "bookmark all tabs" which shows even lower usage so they'll probably do away with that feature as well.
Like you, I imagine most who use those features are using chromium and not sending those stats to Google.
Elsevier was also recently awarded an "Online peer review and method" patent which earned the August, 2016 "Stupid Patent of the Month" from the EFF [0].
https://twitter.com/w4nderlus7/status/1174244681003208704