Poor example in "Look Ma! No Network!

[schoon at hsva.de]

sussman at gmail.com wrote on 25.10.2006 22:02:33:

> As a primary book author, allow me to step into your squabble.  :-)

I am merely spending some of my valuable time to make your book
even better than it already is. To me, that is not squabble.

> Johannes:  you're right, the cached text-base isn't strictly necessary
> to send diffs during a commit.

Glad to hear that, I wasn't 100.000% sure myself. (I can only afford 
that type of luxury on really rare occasions.)

> However, that sidebar *doesn't* say that the text-base exists solely
> for the purpose of efficient commits. 

I never claimed the opposite. My point is, and was, that the example
is based on false premises.

> If you read the whole
> paragraph, it says that...
> 
> I don't see how anything is incorrect or distorted here. 

Ok, I should perhaps clarify that the distortion I hinted at lies in
the example that I criticised. When I change subject, I will tell you
so.

> If quoted
> out of context, yes, it sounds like we're claiming the cache is
> necessary for efficient commits.  If you read the whole paragraph,
> we're simply saying that the efficient commits are just an extra bonus
> feature resulting from the cache.

To me, right there, in context, the example says to me, "Hey, if we
did not do this, you would have to send the entire file over. Cool,
isn't it?" I never had to bring it out of context, and I never did.

Maybe I am wasted by reading too much technical literature where each
sentence, every word, is important. Sloppy logic that brings the
critical reader out of balance has always irritated me. The worst book
with respect to this type of abuse of the reader, that I ever had to read, 

is "A First Course in Turbulence" by H. Tennekes, J. L. Lumley. Stay 
clear of that one, whatever you do. 

Yours

Johannes