[Numpy-discussion] NumPy documentation

Travis Oliphant oliphant at ee.byu.edu
Tue Apr 4 14:18:59 CDT 2006

I received a rather hurtful email today that was very discouraging to me 
personally.  Basically, I was called "lame" and a "wolf" in sheep's 
clothing because I'm charging for documentation.    Fortunately it's the 
first email of that nature I've received.  Others have disagreed with my 
choice to charge for the documentation but at least they've not resorted 
to personal attacks on me and my motivations.   Please know that such 
emails do have an impact.  While I try to build a tough skin, such 
unappreciative statements reduce my enthusiasm for working on NumPy 

My purpose, however, is not to rant about the misguided words of one 
person.  He brought up a point that I want to clarify.  He asked if I 
"would sue" if somebody else wrote documentation for NumPy.   I want to 
be perfectly clear that this is a ridiculous statement that barely 
deserves a response.  Of course I wouldn't.   First of all, it would be 
extreme circumstances indeed for me to resort to that course of action 
(basically a company would have to copy my book and start distributing 
it on a large scale, belligerently).  Second of all, I would love to see 
*more* documentation for NumPy.

If there are other (less vocal) people out there who are not using NumPy 
because of my book, then I certainly feel sorry about that.  Please dig 
in and create the documentation you so urgently want to be free.   I 
will not stand in your way, but may even help.

But please consider that time is money.  Most people are better off 
spending their time on something else and just cooperating with others 
by paying for the book.  But, I'm not going to dislike or have any kind 
of ill feelings with anyone who decides to spend their time on 
"documentation."  In fact, I'll appreciate it just like everyone else.  
I love the growth of the SciPy Wiki.  There are some great recipes and 
examples there.  This is fantastic.  I'm 100% behind this kind of work.  
Rather than write some kind of "replacement" documentation, contribute 
docstrings to the code and recipes to the Wiki.   Then, those that can't 
or won't buy the book will still have plenty of resources to use to 
learn NumPy.

I'm completely behind all forms of "free" information on NumPy / SciPy 
and related tools.  The only reason I have to charge for the 
documentation is that I just don't have the resources to simply donate 
*all* of my time.   I want to thank all of you who have already 
purchased the documentation.   It has been extremely helpful to me 
personally and professionally.   Without you, my time to spend on NumPy 
would have been significantly reduced.   Thank you very much.

Best wishes,


More information about the Numpy-discussion mailing list