To work. It seems there is some debate about the suitability
of the GPL for OO.o (again). It seems to me that the use of the GPL for
OO.o is deeply unwise for at least the following reasons:
- The SISSL with it's invideous terms - allowing
(antisocial) people/companies to develop their own changes to the
OO.o code without sharing them, and working together with us.
Anyhow the SISSL is now dead, which is great. The LGPL means that
any changes or fixes to the existing code by others in future
-must- be published as they are shipped. This is a huge improvement
to my mind.
- Re-usable components - OO.o contains many pieces (eg.
chart2) that have the potential for re-use across Linux (and
other) desktops. GPLing OO.o, would enfeeble the platform by
not allowing this in many cases. The LGPL allows re-use from
Java, and
any number of other projects / tools. While this is admittedly
not a common use-case now, in future, I'd love to see UNO more
widely deployed [as part of the system], and OO.o's components
re-used.
- The Linux Kernel is a world unto itself. People point
to Linux as a great example of the general applicability of the GPL.
However - this is an unreasonable comparison: the kernel 'links' to
almost nothing. No-one contends that the GNU C library should be
GPL (instead of LGPL) because it interacts directly with the kernel.
OO.o by contrast is a massive linking consumer & producer with
bindings for many languages etc. no problem while LGPLd.
- Competing vs. MS Office - there are a large number of
(small) companies making a living by selling proprietary
plugins, macros, and re-using Office as a component of their
solution. Encouraging these people to not only add support for OO.o
(and hence Free software platforms) - but also to GPL their
(entire) product, seems an unwise disadvantage for Free software.
The LGPL avoids this problem.
- If an end-user is writing macros in eg. Java (ie. a
compiled language), or wishes to password-protect their StarBasic
macros; then if OO.o is GPL'd - (IANAL) - it
appears clear (last sentence) that they infringe the license,
unless they also supply the full source for all their macros. This
is not the case with the LGPL.
Personally, and Corporately, I have no particular axe to grind
here, ( Groupwise/ODMA/COM/Win32 ? ). However, I am concerned about the long
term attractivenes of, and re-use of OO.o. As Fred Brooks infers - if there
is a silver bullet out there, it is code re-use. Thus I applaud Sun's wisdom
in sticking with the LGPL.
Of course - in terms of actually working together with Sun, (and/or
the collab.net infrastructure) life can be incredibly painful, turgid, littered
with (seemingly) arbitrary barriers, conflict and so on, all of which need
fixing (and in many areas things are improving), but licensing is (currently)
the least of my concerns
