BSDDB copyright and licensing restrictions while in use via Python

Warren Postma embed at geocities.com
Mon Feb 14 12:35:46 EST 2000


> > Has the db1.8.5 source code a more liberal license than the 2.0 code
from
> > Sleepycat?
>
> I'm pretty sure that's the case.  But instead of
> trusting my memory, why not download it and
> check check the license yourself:

The only apparent license in 1.8.x is at the top of each source file.
They appear to describe the plain vanilla bsd license. If you modify the
source code, you must
leave the headers intact. If you distribute only a binary you must mention
the use of the product in
your documentation.

It appears the 2.0 license is significantly more restrictive for commercial
users.

I guess I'm free to write and release (for free) a non-official "wbsddb 1.9"
should I find there are any significant
problems in 1.8.

> check out metakit (www.equi4.com)

The problem I have with MetaKit is that it's essentially C++ only. (Unless I
goofed and it really is useable from a C API also). My application is in
straight C because I need a very tiny memory footprint.   C+Python gives me
the most bang per kilobyte, and I'm not considering C++ right now.

For those who are already working in C++ and Python, I would recommend
MetaKit highly.  It appears
to be at an early stage of use and adoption, but it appear stable and well
written from the playing around I've done.

If it's possible to build a straight C wrapper around the C++ version of
Python, and MetaKit doesn't require the C++ Runtime Libraries, or the
IOStreams or C++ Strings template classes, then I could use MetaKit.

> or gadfly (www.chordate.com).

An excellent "in memory database" that implements an astonishingly large SQL
subset, this technology is great whenever you need to provide an SQL
interface to an in-memory data set. Alas, In Memory is not what I need, and
an SQL interpreter in Python is not what I need either, although I'll grant
you that it's really nifty, and when/if I ever need an SQL engine, this is
how I would do it.  I would not "choose" to work with SQL if I don't have
to, but rather, would use this if SQL was required of me.

Warren





More information about the Python-list mailing list