for most things in life sampling suffices .. couple of reasons
1. information content in life is low
2. the brain for better or worse is equipped with very good error detection and error recovery processes. so even if you miss data points, its quite easy to interpolate between sampled points. it could be that 1 leads to 2 ..
3. if something is bleeding important the content provider will make sure that its noticeable either by provisioning redudant replicas of the content so that sampling will still guarantee at least one sampled data point, or by making the content strikingly different from the filler material ..
e.g.
1. watching a movie
2. studying for a exam
3. reading a book
4. conversations with people
this doesnt apparently work for tasks where form is more important than content to borrow the discussion from a few posts earlier. sampling content is easier than sampling form .. as some wise man said the devil is in the details ..
Subscribe to:
Post Comments (Atom)
3 comments:
i understand that by sampling u mean the compression of data for storage/communication. if so, the only 'thing' we can 'sample' is 'content'. Form, we can only appreciate, feel, see etc. We cannot store/communicate the Mona Lisa. So what we store/communicate when we talk abt the MonaLisa is the content we posit for the form. Inherently since the MonaLisa is form-oriented, we 'create' content and, depending on our requirements, we sample differently. For example, one man might take the MonaLisa to contain just the single idea 'painting'. For him sampling is quite easy. Another man might find more 'content' - it is the 'painting of a woman', a third, 'a painting of a woman without eyebrows' and so on. Since any content is 'contained' by form, the same holds even in content-rich contexts as a minimum - we could just analyse the form but there is also the content plane to contend with. So as a minimum, sampling any form will require as many 'bits' as we allow for it(the maximum bits required i believe is infinity for anything - nothing can be known completely or rather we can always connect something more to anything and claim that is part of the 'content' of the thing). And the earlier hypothesis of yours holds nicely only now it is true for all objects and with the caveat that all the 'low,' the 'high' and other adjectives are relative. A movie-watcher will always feel there is more to a movie because it works on different planes but a book-reader can retort he has the imagination plane and others too and so there is more info in a book and so on - essentially the problem of mapping n-dimensional infinite planes to one another.
also there was a kind of paradox in ur post - form has more content than content does(using hypothesis 1 as a necessary condition for the requirement of fewer bits of sampling the claim in the end and a fudging up of logic a little to the extent that we are talking non-exact science) :-)
as an afterthought and connecting to an earlier thought .. sampling doesnt work for comedies essentially because humor needs context/memory
madatadam: just to clarify, by sampling i intentionally refer to lossy compression not lossless compression
Post a Comment