[Date Prev][Date Next][Thread Prev][Thread Next][Thread Index]

[XaraXtreme-dev] Cooperating with open source developers

Charles Moir wrote:
But the BIG question is still; what about CDRAW?

But why is this the big question?

The source code of all the product is available with the exception of
CDraw rendering core. This doesn't stop the product working. We make the
binary of this available for free to be distributed with Xara Xtreme.
(And increasing numbers of distros are distributing completely closed
source products such as Flash Player and Adobe Reader, and most Linux
users have these installed even where they are not distributed as part
of the OS). The lack of CDraw source code certainly doesn't stop Xtreme
being progressed and developed by any interested developers if they were
interested in helping us progress the product.

So isn't it a bit of red-herring to be focussed on this small part only?

We've held back on CDraw partly because it's simply not necessary to
have the source code of this to develop, improve, fix bugs, add new
features or all the other things that I'd imagine most users would be
looking for. It's obvious this part is complete, and works perfectly
well (as you can tell using the product), so why the fixation on CDraw?

The original plan and hope was, if you recall, that *together* the
community and Xara would help create a world (and Microsoft and Adobe)
beating product. Well Xara have put a huge amount of time, money and
effort into the product so far, and we've got great result, but there's
not much 'together' in this so far.

So we still hope and look forward to the time where the community can
help us develop the product and achieve that original goal to create the
best graphics product that has ever existed on the Linux (and other)
platforms.  We're still working very hard to make that happen.

You've exhibited some fundamental misconceptions about what motivates
open source developers. This, along with a recent open source faux pas
committed by Xara, helps explain why you're still not getting the
support from the open source community that you hoped for.

Let's get some understanding going. First, there are two steps Xara
must take before its open source efforts can even hope to become

1) Without CDraw, Xara no longer works. Xara needs to acknowledge that
if CDraw is for any reason withdrawn or ceases to be supported, all
development on the open source portions of Xara loses its "immediate
rewards" value to open source developers. To them, this means Xara's
most important core component can no longer be modified, patched for
security issues, compiler-optimized, or ported for new architectures
(and the existing binaries only suit limited needs, and even then only
for a limited time). The potential for some viable CDraw replacement
to eventually fill the void is of zero immediate utility, and so with
their ability to scratch their own itches and receive instant
gratification significantly diminished, most open source developers
will lose interest in Xara. This is crucial to understanding open
source motivations. (Please also realize it's unproductive and
alienating to continue implying there are nefarious "other reasons"
motivating their desire to have CDraw open sourced.)

2) Once step 1 is addressed, Xara must follow up by opening CDraw's
source code. For the very reasons described above, this point is going
to be non-negotiable with the majority of the open source developer
community, and is also a requirement for gaining the full support of
major distros like Debian, Ubuntu, and Fedora which, owing to their
positions in the Linux ecosystem, have considerable ability to make or
break Xara's exposure and adoption prospects. Xara's commercial
concerns are understandable, so why not adopt a QT-style solution to
the problem? Dual-license CDraw: proprietary license for commercial
use, OSI-approved open source license for non-commercial use. This
will leave your competitors in a weaker position than they're in now
because they still must negotiate a commercial license from you if
they wish to use CDraw in their own offerings, and if they choose not
to incorporate CDraw they will face stronger competition once CDraw
has the open source community devoted to improving it.

Finally, on matters of protocol, Xara needs to be more considerate of
downstream packagers. Don't make changes to packages retroactively. I
helped maintain the Xara packages for Gentoo (in the Sunrise overlay
for anyone looking), and when Xara changed the names of existing
release tarballs, it *broke* versions of Xara ebuilds which had been
working perfectly for users. If you need to repackage or rename
existing packages, at least keep the original packages available for
downstream distribution so that users aren't hampered when they try to
install Xara.