Writing solid code book

[Alex Martelli]

post400 wrote:
...

yes Safari seems ok , you get a few days for free to browse but it's
not totally free.I think the electronic version should be free.Some

safari is only free for two weeks; then, if you like the service it
gives you, you can choose to continue a (paid) subscription.

I'm not sure on what basis you intend your use of "should". Do you
mean that you believe it would be in publishers' best interests to
offer electronic versions free -- that it would enhance the sales
of their paper versions producing a greater total profit than having
online books as a for-pay service does? If so, then it would be
interesting to understand on what basis you form that judgment.

O'Reilly, in particular, publishes paper versions both of books that
are also available for free download elsewhere, _and_ of books that
are not thus avialable. Presumably, therefore, they have available
very good historical data to convince them that their policies are in
fact the best currently available to them -- knowing Tim O'Reilly's
ideas, policies, and openness to novelty, you surely can't be
accusing him of "persevering by sheer inertia" in his policies.

Or is your "should" predicated on some ethical objection against
such people as book authors and editors receiving compensation
for their efforts? Speaking as a book author, I can confirm that
the amount of money I earn from writing, divided by the number of
hours I spend writing, is roughly comparable with legal minimum
wage (depending on jurisdiction -- such minima may be higher in
most of Europe than in the US), and *WELL* below the hourly income
I can make by professional consulting. Were I to contemplate the
possibility of would-be freeloaders trying to dictate what I
"should" do (without deigning to mention WHY they think I should),
it might perhaps anger me just enough to deter me from my former
intention of making my next coming book freely available. So, I
would appreciate some explanation of your assertion.

people offer their books for free , Bruce Eckel for example and his
Thinking in Java which seems to be a good book.Maybe more writers
should do the same !

"Thinking in Java" is indeed excellent, and no doubt the feedback
from people who freely downloaded it helped Bruce make it so.
However, I do notice that Bruce's long-planned "Thinking in
Python" seems to be languishing without end. Perhaps this means
that books addressing a huge market for which there are hundreds
of competitors (Java, C++) _are_ indeed profitable to make freely
downloadable (at least if only a minority of authors are so doing,
assuming part of the profit comes from competitive considerations),
but books addressing a smaller market, such as Python, aren't.

As for me, I'm operating on a different assumption (which is not
necessarily contradictory to the possibility outlined in the
previous paragraph) as to what distinguishes books that are best
made freely available online from those which are not; but I have
as yet no experience to back me up (the only "free book" I have
out on the net -- a primer on win32 API GUI programmin in C, in
Italian -- is incomplete and not available in any for-pay form,
only in the free online version).

Several other authors have experience publishing both books that
are also available for free, and ones that aren't. Hearing from
them would be very interesting, at least to me.


Alex

 

[David Mertz]

|> yes Safari seems ok , you get a few days for free to browse but it's
|> not totally free.I think the electronic version should be free.Some

|safari is only free for two weeks; then, if you like the service it
|gives you, you can choose to continue a (paid) subscription.

FWIW, my book, _Text Processing in Python_ (http://gnosis.cx/TPiP) is,
has been, and will remain freely accessible in electronic form for all
readers--even those who have not paid for anything. I really buy into
the "information wants to be free" schtick.

That said, the book is also copyrighted by Addison Wesley; so my
principles are limited by their discretion. For example, I wanted to
make a PDF version of my book available, but only to buyers of the
dead-tree version. Well, without AW to consider, I might have wanted to
do so on an unlimited basis, but I thought that was a compromise. I
even had in mind a little system for using a "What's the first word on
page NNN?" quiz to authorize downloads.

Sadly, AW strongly pooh-pooh'd this idea. I think they're wrong, but
they think leaking a PDF would kill international sales. Obviously,
there -is- a point that testing readers of the original download doesn't
really prevent redistribution of the file. But IMO, the so-called
"pirates" wouldn't be likely buyers anyway.

Ah well... personally, I like the ASCII version better for electronic
access than I would a PDF (well, it's -do- actually... I myself have
the PDF perfectly available on my computer ). When I want to remind
myself of a some function arguments or the like, grep does a really
quick job of reminding me about what I took the effort to research
during writing (of course the standard documentation is likewise useful,
but sometimes I like my own phrasing better... I wrote it, after all
*wink*).

Yours, David...

 

[Jules Dubois]

On 4 Sep 2003 02:40:50 -0700, post400 wrote:

So maybe you (experienced python guys ) [...]

I'm not experienced with Python. I wrote my first forty (or so) lines of
code Thursday morning.

[...] could make some
suggestions from time to time about the must-read books , not
necessarily on python but also on stuff about design , project
management , OOP , etc. be it new stuff or old stuff.

Books I particularly liked:

_Object Oriented Software Construction, 1E_. Meyer. (OOP)
_Design Patterns_. Gamma, Helm, Johnson, Vlissides.
_Refactoring_. Fowler.

For OOP stuff, you might look at the comp.object newsgroup. There are some
Wars of Religion there, but also some good information; Robert Martin
(Uncle Bob) and Martin Fowler post there.

But I really wonder
how a non-programmer could get into python as a first programming
language since there's not much publicity and not many people around
to lure you to it.

Publicity and lure are not the most important things to consider when
choosing a first language.

For education, as opposed to training, I maintain the first two languages a
beginner should learn are Smalltalk and then Common Lisp. Even though
they're not the most widely-used languages, these two will do more to teach
the fundamentals of programming than will Java, Python, Ruby, C++, Perl,
PHP, C#, or (especially) VBA.

 

[Olivier Lefevre]

It's got over 1,000 books -- I doubt they can ALL be best-sellers,

can they?

What I mean is that it seems to have only language or API how-to kind
of books, i.e., those that appeal to working programmers, the largest
audience. A contrario look under Computer Science, AI or Functional
Programming: it's pitiful. Yet the CS book market is also large,
courtesy of undergrad programs. So they may have 1K titles but the
coverage is very narrow nonetheless.

Sure: just like it takes a really long time for the latest movies
to become available at my local video rental shop

I don't think that's relevant. The videos could hurt the theater
release, so the delay is deliberate; in the case of Safari it's
more likely to be mere logistics since printed and online versions
coexists throughout the lifetime of the book.

-- O.L.

 

[John J. Lee]

What I mean is that it seems to have only language or API how-to kind
of books, i.e., those that appeal to working programmers, the largest
audience. A contrario look under Computer Science, AI or Functional
Programming: it's pitiful. Yet the CS book market is also large,

[...]

Isn't that just because that's the kind of stuff O'Reilly publishes?
Addison-Wesley, Wiley &c. tend to do the books that are more on the
software engineering and CS side.

Or does Pearson Education have something to do with it -- I noticed
their site mentioning a 'Safari' a while back, and at the time assumed
that was a copy, but I think Alex mentioned them in this thread. Are
Pearson and O'Reilly owned by a single company, or something? I have
a vague recollection that Pearson owns some other companies...


John

 

[David Mertz]

(e-mail address removed) (John J. Lee) wrote previously:
|Pearson and O'Reilly owned by a single company, or something? I have
|a vague recollection that Pearson owns some other companies...

There's been a huge amount of consolidation in publishing. Pearson is
not the parent of Addison-Wesley, Longman, Benjamin Cumming, New Riders,
Peachpit, Cisco Press, Prentice Hall, SAMS, Que, and several more
imprints. It's really a rather bad thing, although they do not seem to
use the degree of ideological acid-test that, say, MediaCorp
acquisitions do.

I started writing my book for AW, only to find it a Prentice book when
it was published (it still has the AW imprint, but the meaning seems
diluted).

That said, I'm pretty sure that O'Reilly remains one of the few
independent publishers of technical books. I think Manning is also.

Yours, David...

 

[John J. Lee]

(e-mail address removed) (John J. Lee) wrote previously:
|Pearson and O'Reilly owned by a single company, or something? I have
|a vague recollection that Pearson owns some other companies...

There's been a huge amount of consolidation in publishing. Pearson is
not the parent of Addison-Wesley, Longman, Benjamin Cumming, New Riders,
Peachpit, Cisco Press, Prentice Hall, SAMS, Que, and several more
imprints. It's really a rather bad thing,

(That should have been "now the parent...", I assume.)

Wow. :-(

Actually, doesn't Pearson own some stuff outside of publishing too?
That's what ISTR hearing about recently...

although they do not seem to
use the degree of ideological acid-test that, say, MediaCorp
acquisitions do.

What are you referring to there? Is Mediacorp Rupert's company?
What's this about ideological acid-tests?

That said, I'm pretty sure that O'Reilly remains one of the few
independent publishers of technical books. I think Manning is also.

By technical books, do you mean computing books? I've no idea what
goes on with other fields, but I don't doubt they've been
'consolidating' too (and, of course, AW, Longman, Prentice Hall
themselves publish(ed) a broad range).

So, *is* Pearson's 'Safari' the same as O'Reilly's? I thought Alex
had mentioned Pearson in that connection. Too lazy to check.


John

 

[Jules Dubois]

That's the only Python book you think worth having? Or buying?

Speaking as both the co-editor of the Cookbook and the author of the
Nutshell, if I had to choose ONE Python book "worth having or buying"
I think I'd go for the Nutshell. Hard decision, though.[/QUOTE]

I discovered the Engineering library had ordered _Nutshell_, but it hadn't
been delivered from the main library. I requested it and got it the next
day.

However, seeing for the first time Python's syntax in _Nutshell_ was enough
to scare me away. We had to commit today to development in Java or Python.
I had made a firm decision to choose Java, even buying a book on the
Eclipse IDE. Just before sign-ups, we discovered a sample Tkinter
application on the class' internal BBS. After comparing the Tkinter code
to the Swing code I've read, I signed up for the Python development team
five minutes later.

This afternoon, I bought _Cookbook_. When _Nutshell_ is due at the
library, I'll buy it too.

Thanks for your suggestions.

 

[Alan Gauld]

However, seeing for the first time Python's syntax in _Nutshell_ was enough
to scare me away.

Really? Can I ask why? I'd have thought seing Pythons syntax
would be a breath of frsh air to most programmers, it certainly
was for me. Why did you think it was scary?

Curious,

Alan G.

Author of the Learn to Program website
http://www.freenetpages.co.uk/hp/alan.gauld

 

[John J. Lee]

application on the class' internal BBS. After comparing the Tkinter code
to the Swing code I've read, I signed up for the Python development team
five minutes later.

I'll trollishly <wink> repeat a previous assertion of mine: I've never
heard anybody claim that any other GUI framework beats Qt in terms of
good design and power (hence usability & code clarity). Since you're
in an academic environment, licensing isn't an issue for you (assuming
the educational license for Qt3/Win isn't a problem for you -- Qt3/Mac
and Qt3/Unix are both available under the GPL), so I'd certainly go
with Qt in your position. The only fly in the ointment is that PyQt
doesn't yet work natively on the Mac (but Phil is planning to fix that
"with SIP v4").


BTW, I just looked at Phil's SIP roadmap page

http://www.riverbankcomputing.co.uk/sip/roadmap.php


and -- shock, horror -- SIP is actually going to be documented in v4!
We'll believe it when we see it ;-).

(prospective PyQt users should note that SIP is not something you'll
ever use while writing PyQt code, so the lack of documentation isn't
an issue -- SIP is the tool used to wrap Qt for Python)


John

 

[Olivier Lefevre]

Just to be clear, Safari is not limited to O'Reilly, although
O'Reilly created and operates it.

I've no idea what goes on with other fields, but I don't doubt they've
been 'consolidating' too

Yes. In STM publishing, following the purchase of Springer by Wolters
Kluwer, it and Elsevier Reed now control a huge share of the market.

-- O.L.

 

[Geoff Gerrietts]

Quoting Alan Gauld ([email protected]):

Really? Can I ask why? I'd have thought seing Pythons syntax
would be a breath of frsh air to most programmers, it certainly
was for me. Why did you think it was scary?

I know I resisted the idea of significant whitespace for something
like six months, despite evangelical friends, before I turned away
from the dark side.

Perhaps not coincidentally, I began reconsidering my opinion at
approximately the same time I became responsible for extension of a
Perl script written by a fellow who was just coming to understand
use of references in Perl.

--G.

 

[Bengt Richter]

(e-mail address removed) (John J. Lee) wrote previously:
|Pearson and O'Reilly owned by a single company, or something? I have
|a vague recollection that Pearson owns some other companies...

There's been a huge amount of consolidation in publishing. Pearson is
not the parent of Addison-Wesley, Longman, Benjamin Cumming, New Riders,

^^^-- assume you mean "now"

Peachpit, Cisco Press, Prentice Hall, SAMS, Que, and several more
imprints. It's really a rather bad thing, although they do not seem to
use the degree of ideological acid-test that, say, MediaCorp
acquisitions do.

I started writing my book for AW, only to find it a Prentice book when
it was published (it still has the AW imprint, but the meaning seems
diluted).

That said, I'm pretty sure that O'Reilly remains one of the few
independent publishers of technical books. I think Manning is also.

[OT] So instead of a distributed implementation of business, with no
individual way of making billion dollar mistakes, the consolidation
centralizes control and makes billion dollar mistakes both possible
and survivable, and therefore more probable. And if one of these
titanic dinosaurs starts faltering, it is so socially disruptive that
tax payers have to provide survival insurance one way or another.
Evolution at work ;-/

[really OT]
Factoids: Iraq 2002 per capita GDP: ~$2,400. Iraq population: ~24.7 million.
$87bn if distributed to all Iraquis: ~$3,500. per capita. And that's just
the new increment. Sorry. Just trying to get a concept of what these numbers
mean. Or could have meant. (E.g., if you wanted to put 100 tourists on each and
every square mile of Iraq at the same time, with $87bn you could give every tourist
a ~$5,000. travel budget to do it). (87e9/168e3)/100 => 5178.5714285714284)
It would be interesting to compare the effects of the expenditure alternatives
if it were possible. IMO tourism would have been a lot more fun (and good for
airlines too). I suspect peaceful expenditures have more business multipliers
in them, not to mention what happens in people's minds if they can enjoyably meet
as humans rather than agents of some extraneous worry ;-) (Ok, let's not get into
the logistics of sending 16.8 million tourists with $5k each at the same time
to one country ;-) Besides that's only $84bn. We have $3bn in roundoff to spare ;-)

Regards,
Bengt Richter

 

[Lulu of the Lotus-Eaters]

(e-mail address removed) (Bengt Richter) wrote previously:
|[OT] So instead of a distributed implementation of business, with no
|individual way of making billion dollar mistakes, the consolidation
|centralizes control and makes billion dollar mistakes both possible
|and survivable, and therefore more probable.

I don't have an airtight explanation for why it "has to be so", but
empirically[*], media consolidation almost always leads to a narrowing of
ideological diversity. Fewer and fewer different opinions get
expressed, and everything gets mushed together into an "offend no one"
middle.

The trend is perhaps more obvious with political news, movies, music,
etc., but I fear for the same thing in computer books. Rather than
publish occassional books on oddball technologies by authors with their
own take on programming matters, the danger is that every book becomes
"VB.NET for Dummies", or something like that.

It may seem like a silly fear. But back when I was shopping my book (a
couple years back now), one of the publishers I approached was New
Riders. They had published Beazley's excellent _Python Essential
Reference_, as well as quite a number of good programmer-oriented
titles. But the stated "new" focus of New Riders was "graphics design
applications" (I think one of the first steps in mergers was
PeachPit+New Riders). While there is certainly nothing wrong with that
type of book, it really seems to throw away the hard-earned reputation
of New Riders for a certain kind of high-quality programming text--in
favor of something which seems a lot more "middle of the road."

AW still seems to do a lot of good books (not just mine), and maintains
that nice computer-science-ish focus. But I wonder whether a
cookie-cutter pressure will force their titles into narrower molds, over
time.

[*] Maybe Alex will pipe in with a "rational actor" explation of why
what I claim either must be, or cannot be, true.

|[really OT]
|Factoids: Iraq 2002 per capita GDP: ~$2,400. Iraq population: ~24.7 million.
|$87bn if distributed to all Iraquis: ~$3,500. per capita.

Brilliant observations, Bengt!

Yours, Lulu...

 

[Graham Breed]

So, *is* Pearson's 'Safari' the same as O'Reilly's? I thought Alex
had mentioned Pearson in that connection. Too lazy to check.

I've got a URL for Pearson Education's putative Safari:

http://safari.it-minds.com

It diverts to their online bookshop. As the Pearson group's books are
available at O'Reilly's safari ( http://safari.oreilly.com/ ) perhaps
they gave up on having their own one.


Graham

 

[Alan Gauld]

I've got a URL for Pearson Education's putative Safari:

http://safari.it-minds.com

I'm amazed nobody has mentioned InformIT.com yet.

They have a huge range of IT books and articles available. The
articles are free but the books are safari based and need a
subscription. They seem to have more than O'Reilly so far as I
can tell.

InformIT is run in conjunction with Pearson I think.

Alan G.


Author of the Learn to Program website
http://www.freenetpages.co.uk/hp/alan.gauld

 

[Graham Breed]

I'm amazed nobody has mentioned InformIT.com yet.

They have a huge range of IT books and articles available. The
articles are free but the books are safari based and need a
subscription. They seem to have more than O'Reilly so far as I
can tell.

The "New This Week" and "Top Books" are identical, so they look to me
like different brandings of the same service.


Graham

 

[Jules Dubois]

On Fri, 12 Sep 2003 07:20:53 GMT, in article

[Python syntax scary]

Really? Can I ask why?

Upon further meditation (and reading), perhaps it isn't so scary after all.
It's just different from what I'm used to, as are Lisp, Prolog, and ML,
which I also have to learn in a hurry. Having begun to re-read _Nutshell_
and with the great help of another of the project's members, some of the
language features I found inscrutable are now becoming scrutable.

However, I would much rather have an assignment operator rather than an
assignment statement. Some of us C programmers can't think properly
without it!

I'd have thought seing Pythons syntax
would be a breath of frsh air to most programmers, it certainly
was for me.

In _Cookbook_, I find features like the example below to be pleasing:

thenewlist = [x for x in theoldlist if x > 5]

With a BS in Math and some Perl experience, this statement is elegant,
concise, and transparent -- in other words, nearly perfect.

Alan G.

Finally, G. -- may I call you G? We're all members of the same gang now,
aren't we? -- I want you to know your Python web site and Tkinter tutorial
are being recommended to us by our resident Python hacker^w expert.

 

[Jules Dubois]

On Fri, 12 Sep 2003 08:42:22 -0700, in article

I know I resisted the idea of significant whitespace for something
like six months, despite evangelical friends, before I turned away
from the dark side.

I wish I had said that.

I don't have a problem with the whitespace requirement, having programmed
previously in an assembly language which required it. It was one of the
things about Python I liked from the first, except it means line-oriented
syntax with continuations (\).

Perhaps not coincidentally, I began reconsidering my opinion at
approximately the same time I became responsible for extension of a
Perl script written by a fellow who was just coming to understand
use of references in Perl.

Although I think Perl is a good language, it's entirely too easy to program
in write-only mode. C is less powerful, and it's a little harder to create
opaque code... but not impossible as the obfuscated C code contests
demonstrate.

 

[gene tan]

|> yes Safari seems ok , you get a few days for free to browse but it's
|> not totally free.I think the electronic version should be free.Some

Anyone heard if Oreilly planning to do CD bookshelf? Any good ways to
apply moral suasion? I think they have just right amount material:

Prog'g P,
Learn'g P,
nutshell
cookbook
Std. Library
Win32.

The perl, network'g & unix CD bookshelves made my life much easier
(til the search function quit working w/JVM 1.4SE)