RE: [MLUG - DISCUSSION] web development language of choice

[Jonathan King <EMAIL:PROTECTED>]

[Aside: this stuff is actually related to a problem that is near and dear
 to my heart (mucking around with frequency information when you don't
 really have a good source), which is why I do go on so...]

On Thu, 3 Jan 2002, Ross, Matt wrote:

> Book sales would be only useful if you already knew the answer to
> this.  It would tell you which is harder to find help for online.
> That would make PHP books more heavily purchased than there are PHP
> users, as it's not as well supported yet.

You know, I used to think like this, since it makes a lot of sense. :-)
However, the anecdotal evidence for this seems to be surprisingly weak.

To a great extent, internet-related things that are popular both sell
books and develop good online docs.  In fact, the online docs, if
contributed by the creators, often lead the books by a substantial amount
of time.  I'm not sure I have (since the web) seen anything that had
superb books and no on-line docs.  And I'll argue that when you see the
opposite, it's primarily because nobody uses the tool.  Another
interesting (or irritating) thing is that when you first do see a wad of
books on Foo, most of it is obvious regurgitation of the online resources.
Yes, I'm sure there are exceptions out there, but people really do seem to
buy what they also could get on-line.  I know I do. :-)

OK, so here's the number of titles offered by Amazon in their Computers &
Internet category that mention certain key phrases that just happen to be
the names of popular scripting languages:

  perl   253
   php    43
python    37
   tcl    27
  rexx    12
  ruby     6

Note that this isn't completely fair, since I made no attempt to correct
for whether everything was in print or even take Amazon sales rank into
account.  (Now that could get interesting; we know by Zipf's law that we
should get a shockingly close estimate of overall sales from sales rank
alone, but see below.)

But even this is at least a bit interesting.  I think we would all agree
that the top 4 of these languages have "made it", although tcl is
languishing; rexx is a niche.  Ruby is a language not many have heard
of...yet; the key point here is that a search made this time last year
would have produced at most 1 title.

A frustration with using Amazon rank as a measure of popularity, though,
is that the sales rank information they give is not cumulative sales, but
some weird function of current sales.  A major problem with that then, for
estimating overall sales, is that books tend to sell best when first
published, and then fall off (sometimes pretty rapidly).  Nonetheless, you
can get some idea of "what's hot now" this way.  With that in mind, I was
still surprised by the sales rank data for the top book of these:

php      329 "PHP and MySQL Web Development"
perl     372 "Programming Perl"
tcl     4975 "Exploring Expect..."
python  8309 "Python Essential Reference"
ruby   11053 "Ruby in a Nutshell"
rexx   53174 "MVS TSO:  Commands..."

Note that the php book was published in 2001, while the perl, tcl, python,
and Rexx books are all back in the 90s; the official pub date on the ruby
book is actually Jan 15, although I already have one, so that sales data
there are likely to be weird.  Still, I would *not* have predicted that
the php book (which has the mysql buzzword too, to be fair) would be
out-polling perl.  That's just astonishing.  Also astonishing is how far
tcl has fallen; expect is just a tool that happens to use tcl.  It
would be kind of like the leading C book actually being on the ncurses
library.

But before we get too excited, note that perl has (I think) 4 other books
in the top 1000 list while the next best selling php book is
6000-something.  This matters since overall current sales should be more
closely related to some function like (often exactly) 1/(rank), and
1/329+1/6000 < 1/372 + 1/600.

> Run websearches for .php .pl and .cgi - that might tell you a little
> more, though it imbalances the other direction, as perl has been
> around longer..

Actually, I like to look at unique threads on groups.google.com, for just
that reason.  Now that's an even more interesting one, at this moment in
time (December 1, 2001 through today):

python  3200
perl    2480*
php     2140*
ruby    2110
tcl     2040
rexx     218

(I put *s next to perl and php since I know they have subgroups in their
hierarchy that are important but not included here.)  Again, there are
issues I'm blowing off, but I think most of us should be impressed by how
much attention python is getting.  I suppose that might be in part because
the books have been falling a bit behind these days (similar to your
argument).  Me, I'm surprised that ruby is within a factor of 2 of the big
guys already.

jking

--
To unsubscribe, go to http://mlug.missouri.edu/members/edit.php

Archives are available at http://mlug.missouri.edu/list-archives/