| Markus E Leypold 2007-05-31, 8:04 am |
| >
> 3. OCaml desperately needs a cheap introductory book.
On a more serious note: I agree with this (if we qualify it by "... to
get more usage even by the casual programmer" or something like this:
Every need has to have an implied goal or aim).
But:
- Most people don't realize how much time and effort is necessary to
write a decent book -- and how many units you have to sell to pay
the author (not to talk about the editor/corrector and the tiny
profit the publisher has to make after all).
- We alread have a number of beginners books for OCaml and the
really very concise but rather tight introduction at the beginning
of the manual (that was how I learned OCaml years ago: I started
with what I knew about Scheme and Lisp, learned about the
Hindley-Milner type system from Richard Bird ; Philip Wadler:
"Introduction to Functional Programming" and finally read the
first chapters of the manual. There IS everything there you need
to know, certainly , but there should be some easier way, see the
last point).
- Many introductory books are too long: They start with boring
details and are structured in a way that they are hard to read
cursory.
- What we really need is not another beginners book ("How to
program") but something that can be read by people that already
know another programming language or some of them. What I'd
suggest is a series of short tutorial essays that focus on some
specific aspects of the language (Functions, Datatypes, Data
abstraction, mutable state, references, OO) and sometimes target a
specfic audience (Ocaml for Schemers, Ocaml for C
programmers). Those would be similar to the first chapters of the
manual, have complete code examples, but go into more detail. At
least some of them would be usable as reference documentation for
the aspiring Ocaml novice. Like, e.g. "Arrays in Ocaml": Presently
you have to pick information in use and syntax of arrays from all
over the manual.
Some of the information would be duplicated, of course, but
usually under different headings and in a different context.
> So I think there is plenty of scope for upmarket books targetting
Referring to my 1st point, I don't think there is. My suggestion from
the last point would probably best fly as a community effort in a
wiki-style framework. (Perhaps efforts like this already exist? If not
and there IS interest, I'd be happy to instigate such an effort
myself).
> professionals, like we target scientists and engineers. The financial
> sector have next to nothing and the pharmaceutical industry have next to
> nothing.
They probably have also next-to-nothing demand. Most of them are
afflicted by legacy systems in (ahem) well known languages and that is
where the actual market is seated: Not "How can I build a new an shiny
house with maintenance free plumbing and a biogas plant" but rather
"I've a bloody house I've been building on the cheap and now the
plumbing is gone AGAIN. I URGENTLY need somebody to clean out the
sh*t. It stinks." (Sh*t makes a great market. People pay to get rid
of it or to get it properly contained. It's already there, so (as
opposed to buying something new) the choice of "I don't need that" or
"I don't need that now" does not arise: The only choice remaining is,
wether you want to have the sh*t stinking in you living room, move
into another house (aka "shut down operations") or pay a sh*t expert
to contain the problem (or get tenants w/o a sense of smell:
Admittedly that seems a big strategy presently in the software
market). The market seems sort of guaranteed and much less dependent
on current prosperity than the "buy something new" markets).
> Moreover, mainstream publishers do not know how to sell books on
> non-mainstream languages, so you might as well self-publish as we do.
Mainstream publishers, on the other side, spread the risk on many
publications, so can take on publishing projects that have less than
50% chance to pay for themselves: If they don't pay, they at least add
to the coverage of the catalog (which is a value in itself to some
extent) and some of them pay for the others.
A 50% risk of failure on the other side is much to big a risk to take
in self publishing (if you don't do it as a hobby).
Regards -- Markus
|