[Insight-users] Re: [Insight-developers] IJ Volunteers : THE MATRIX : RED PILL

Stephen R. Aylward Stephen.Aylward at Kitware.com
Sun Sep 17 17:41:10 EDT 2006


Hi,

Regarding open-source, [particularly regarding the Insight Journal (IJ), 
the Insight Software Consortium (ISC), the MICCAI open-source/data 
workshop, and ITK] it seems as though expectations are being formed by 
some that seem unreasonable to others and certain word choices are being 
made some that are offensive to others.

Perhaps the underlying problem is a lack of information:

1) I'd like to start by thanking Luis, Andinet, Zachary, Torsten, 
Gaetan, Julien, Jim Miller, Tobais Heimann, Karthik, Michael Bell, 
Camille Lemen, Henning Meyer, Terry Yoo, and many others for their 
contributions to ITK and open-source.  I don't think such folks get 
thanked often enough.  ITK is a success and will continue to be a 
success because of them and others like them!

2) THE VAST MAJORITY OF THE ITK, IJ, AND ISC WORK, including much of 
Kitware's involvement, IS VOLUNTARY.  Perhaps everyone know this, but 
maybe not.  Volunteers come from academia and industry.  Volunteering 
code/documentation is just one essential part of the volunteer effort. 
I recently learned that the costs of ITK/ISC/IJ volunteering perhaps 
holds true for Kitware employees more than any other group!  I just 
switched form being an assoc prof in a radiology department with tenure 
and nice salary to a Kitware employee, and I was surprised to find that 
most of Kitware's employees were also full or partial ITK/ISC/IJ 
volunteers!!  Kitware is an awesome company, I am extremely lucky to be 
part of their open-source mind-set and the amazing research and support 
that they provide to a diverse set of clients.  But did ya'll know that 
Kitware employees are paid by the hour, AND for the past several years 
most don't get ANY money/hours to work on ITK, the IJ, or the ISC?!?! 
Kitware does have money to support SOME of Luis' time on the ITK user's 
list and a few hours for others at Kitware to provide specific ITK 
deliverables, but nearly all of Kitware's work on ISC/IJ/MICCAI is 
voluntary.  What that means is that every hour of the day, Kitware 
employees, such as Luis, are faced with the choice of personally making 
money by working on a Kitware project, or NOT making money by working on 
the ISC/IJ/MICCAI effort.  I did not face that challenge as a prof in 
academia - my salary was set - it was easy to volunteer during the 
"workday."  Volunteering often personally costs Kitware employees' time 
and money.  For example, Gaetan, because Luis and others were helping 
you integrate your software into the ITK this month, they might actually 
receive a smaller paycheck next month!  Again, many people outside of 
Kitware also face this challenge.  I just wanted to make sure everyone 
knew that this challenged was faced by Kitware employees too.  I had 
mistakenly assumed that Kitware was still getting money to arbitrarily 
spend on all things related to ITK - they're not.

3) Particularly as volunteers (okay, for me it happens even when my 
paycheck is on the line :) ), we (ITK/IJ/ISC/MICCAI volunteers) make 
mistakes and don't always have time for everything that we want to do - 
if this surprises you, then I wish I lived in your world.  If you think 
we're slacking, perhaps consider suggesting a solution or offering to 
help instead of simply pointing out the problem.  We ARE open to 
constructive suggestions - if this surprises you, then I am glad that I 
don't live in your world. :)

4) Not everyone is going to become a "core" ITK developer, but there are 
people out there (reading this?) who will rise to become a "core" ITK 
developer.   Note that there really is not such thing as a "core" 
developer except perhaps in the mind of those who are not "core" 
developers. :)  Yes, flawed logic, but go with me on this - the point is 
that anyone with CVS write access is treated as an equal in terms of CVS 
access rights.   How they choose to volunteer and take the initiative to 
use those rights is up to them - they can stay on the fringe or become 
part of the perceived "core."   We need more "core" developers!  We are 
actively seeking them.  Publish a few papers in the IJ and chances are 
you'll be offered to become one - we do have to know that you're not a 
psycho, anti-open-source terrorist, so we do have the IJ as an informal 
hurdle at this time.

5) The IJ was intended, from the start, to be more than a gateway for 
ITK.  The Charter of the ISC (Insight Software Consortium) and the IJ is 
to promote open-source for medical imaging - not to promote only ITK. 
An open-source community, focused only on ITK, is detrimental to the 
long-term success of open-source.  ITK uses other open-source packages 
(GDCM, VNL, etc.).  Most applications that use ITK also use multiple 
other open-source packages (VTK, FLTK, DCMTK, etc.).  There is MUCH to 
learn from other open-source packages that don't use ITK and that 
perhaps duplicate its functionality.

6) As a broad conduit for open source and open data, the IJ's just 
completed a very successful round of paper handling for the MICCAI 
open-source workshop.  It required significant revisions to the IJ 
software, and it involved 100s of people as authors, reviewers, and 
organizers.  The IJ has received more reviews in the past few weeks than 
every before!   It is becoming a success, but I admit that its 
particular goal of handling ITK papers needs to be improved.   Again, 
should it really be a surprise that the developers of the ITK need more 
help establishing the first-ever open-access software journal with 
automated testing and public review?   I bet it will be another year 
before we get it right.  The hard part is that I don't even know what 
"right" is going to look like.  I've been focusing on finding reviews 
for new papers as they arrive - that is all that I personally have the 
time/money for - and I am about 2 weeks behind.  Again - give us your 
suggestions, remember the broad picture, take initiative, and volunteer!

Thanks for listening to me ramble.  Feel free to cut-and-paste my text 
and point out the flaws of my arguments, but I hope the general message 
still comes across.

Best regards,
Stephen


More information about the Insight-users mailing list