[sf-lug] Book bloat (was Re: Byfield's "Verdict" on systemd)

[Rick Moen]

Quoting Akkana Peck (akkana at shallowsky.com):

> Just to note that there are other reasons for books to get longer.

Sure.

> After you finish a book, everywhere you look, there are things that
> make you think "Oh, wow! I should have included a section on that!"

And also, you get e-mail requesting you expand scope to include X, Y,
and Z.

Eric Raymond and I fell into that (arguable) pitfall with the essay 'How
to Ask Questions the Smart Way':  We good-naturedly complied with any
number of arriving requests to also address $FOO.  Personally, I think
that process went way too far, and we should have kept tighter rein on
scope.  IMO, the current form of the essay's a bit of a shambling thing,
where it started out being pretty short, clear, and keenly focussed on
what we saw as the core of the problem.

Thing is, it's far, far easier to just throw in more stuff than it is to
redact and rewrite the piece as a whole, so you get Web pieces (and
books) that just accumulate more _stuff_ with each edition, and it's
rarely the case that longer works out to better.

> It sounds like in the case of this specific book, you're saying the
> new material is just publisher-mandated padding, and I'm sure that
> happens.

The reasons for the bloat are IMO far less interesting than the fact
that it happens.

Let's consider Evi Nemeth's system administration book, again.  It
started out as a lithe and focussed guide to system administration on
Unix (or, as it was written old-school style for no compelling reason,
UNIX).  IIRC, in doing so it focussed on commonalities across Solaris,
HP/UX, and a few other commercially significant proprietary Unixes, but
then had complementary coverage of BSDisms.

Then in the middle 1990s, Linux suddenly became a huge fad, the next
edition's page count shot way up, and the title changed from _UNIX
System Administration Handbook_ to _UNIX and Linux System Administration
Handbook_.  And, honestly, was there enough difference between
old-school SysVr4.x-derived Unixes and Linux distros to justify that?  I
really, personally, don't think so.  The Linux world, exercising
influence via the POSIX 2.x committee effort and otherwise, merely had
a unifying effect on Unix in the sense of carving out a middle path that
was a canny compromise, but it wasn't hugely different despite its
kernel and libc coming from hippie coders.  

But Addison-Wesley and Nemeth presumably knew what way they had to bend
in order to better sell the product, so they morphed it as described.

Ms. Nemeth might merely have been bubbling over with enthusiasm to write
all-new chapters about RPM and apt/dpkg, etc., but personally I doubt
it.  Sadly, because of her tragic death in July 2013 when the sailboat 
Niña vanished during a Southern Hemisphere winter storm in the Tasman
Sea, nobody can ask her.