[FoRK] Videos of talks / presentations considered harmful (and other peeves)

Jeff Bone jbone at place.org
Tue May 25 05:31:30 PDT 2010

Jim plays the perfect straight man, providing a teachable moment (thank you!)

> Watching the video takes about an hour. Reading the paper requires 
> longer (but could be skimmed faster).

This statement is more or less non-sensical, though in the parenthetical you verge on an important point.

A video is generally going to take different individuals the same amount of time to watch (assuming you don't mess w/ playback speed or something and listen to the presenter in chipmunk mode. ;-)

Reading rates vary wildly between individuals.  Reading to a particular level of comprehension varies wildly between individuals;  reading to different levels of comprehension may vary wildly for a given individual, given different subject matter and differing amounts of context or familiarity.  

The means of covering and learning the content in question is quite different between e.g. video-watching and reading.  The former is more or less strictly linear;  the latter may be done in a linear fashion, but also supports other modes (i.e., skipping around in a breadth- or depth-first manner, etc.) of consumption.  I.e. reading can easily support far more "random access."

Compression / decompression is a useful model of learning.  If a particular reader already "has the dictionary" for a given topic, or a substantial part of it, this process can "zip" along much more quickly than if the reader is completely naive about the subject matter involved.

I find that, for myself, I can generally read a paper on most software related topics to an equivalent level of comprehension as an overview presented in a video that covers the entire paper at a high level of detail in, oh, let's say 10% to 20% the amount of time it takes to watch the video overview.  For other topics, say probability, finance, or certain economics-related topics, the rate's a little slower but still a substantial savings over a video presentation of the same material.  For other topics, say physics, a little slower still.  Reading to complete comprehension for those topics, for me, is usually still faster than watching a video overview of the same material at a high level of detail by some significant amount.  (Again, moreso with software, less so with other areas to which I have some high degree of exposure.  For areas I have little context in, there is often still some small savings.)

Some kinds of information don't follow this pattern;  for example, for me, it's always faster and results in better comprehension to watch a teardown video or instructable about how to perform some hardware-related task (e.g., "replace SSD in MacBook Air") than to read some static presentation of the same material, even with copious pictures, online.  Yet even in this context, having both (video and static text-plus-pictures) is useful;  in this case the pattern would be to sit and watch the video end-to-end, then skim through the textual presentation armed with newly-acquired context, then refer back to the static information while actually performing the task in question.

In any case, comparing the pedagogical applications and value of video and text presentations is an apples-to-anvils endeavor.  Their purpose and use can be quite different.  Only folks who either give or subject themselves to lots of lectures believe lectures to be some optimal way of absorbing information at some particular level of detail for all types of content. ;-)  For myself, I have always found lectures to be relatively extraneous given adequate source material;  indeed, when taking classes the only real value I ever got out of lectures per se was as a kind of live crib sheet to what was going to be on the test. ;-)  (The occasional side-bar interactive conversation with some expert about some particular point notwithstanding;  that's often useful, but facilitating that kind of interaction is rarely the focus / structure of class time, IME.  Supplementary labs and discussion classes to the contrary, of course.)

YMMV.  I'm not actually at war with video / audio / live presentations and teaching modes;  let's have more of it!  But please, let's not kid ourselves that it's a reasonable substitute in any way, for any purpose, for actual source material.  (This thought might undermine the sense of self-worth for some pedants, but c'mon.  Let's get real.)  Source material may be necessary and sufficient in itself for many if not most cases, for many if not most individuals;  rarely is the human presentation alone either necessary or sufficient, and I'd claim that in those rare cases when it *is* either or both necessary and sufficient, we're engaging in "applied folklore" and the content itself and its uses would be better served by both a written and an oral / visual format.  

Finally:  it's a lot easier to get a machine to learn over some content by feeding it text, equations, etc. than by feeding it A/V content...  cf. previous comments about learning-compression equivalence.


More information about the FoRK mailing list