Re: [GMOD-devel] Open Letter to Lincoln Stein RE: BioMart Renaming

[Arek K. <are...@gm...>]

Dear Lincoln:



As you requested on January 24, 2012, I agreed to have a private discussion
with you as I truly believed that you would want to come to an agreement
that would be mutually beneficial, while also being beneficial to the
entire scientific community (including the ICGC).  I am shocked and
disappointed that your proposal did not address my main concerns, which, as
discussed are as follows:



1) The BioMart code base is being renamed at OICR.

2) You intend to split the BioMart code base, thus creating a separate (yet
almost identical) product.



I am also shocked and disappointed that when I privately reiterated my
concerns and reminded you that there is absolutely no need to rename or
split the project since all ICGC-specific customizations can be achieved
through project specific configuration (a solution that has worked for all
BioMart projects throughout the years, including the ICGC), you refused to
engage in any constructive dialogue and plainly stated on February 2, 2012
that you "will not negotiate with me any further on this matter."



Prior to my departure from the institute, I distinctly remember that three
of the five OICR values were:

a) Excellence (scientific and ethical)

b) Innovation; and

c) Teamwork



In my opinion, your current actions seem to be at odds with these values:


a) By choosing to rename the ICGC data management solution (e.g. "SciMart",
"ICGCdb", "franchise"), which has been credited to BioMart (
http://database.oxfordjournals.org/content/2011/bar026.full), the result of
your colleagues ideas, concepts and algorithms, it raises a number of
ethical questions.  None of the other 40+ groups that use BioMart (
http://www.biomart.org/community.html) feel it is necessary to rename it in
order to support their customizations. OICR should not be any different.

b) By splitting the BioMart code base, you are duplicating efforts, and
simply reinventing the wheel.  As a result, rather than having one highly
innovative and unified system (
http://www.nature.com/nmeth/journal/v9/n1/full/nmeth.1815.html), there will
be two almost identical yet incompatible systems, which will impede
existing and future collaborations, and create unnecessary confusion and
problems for the ICGC and the rest of the scientific community.

c) You have decided to terminate a highly successful collaboration with the
BioMart project and rebuffed my many attempts to reinstate a collaboration
with me and/or the other BioMart developers.  Expertise the ICGC DCC could
certainly benefit from seeing that the last functional software release was
in May 2011 - nine months ago.  Had OICR decided to continue to collaborate
with the BioMart team, all of the ICGC-specific customizations (e.g. data
loading and visualization tools) could have been easily implemented months
ago.



Based on all of this, I have no doubt that in the end, you will choose to
rectify the current situation by:



1) No longer attempting to rename BioMart (e.g. "SciMart", "ICGCdb",
"franchise", etc)

2) Restoring all BioMart repository names at OICR to "BioMart" (e.g.
https://github.com/oicr/icgcdb should be renamed
*https://github.com/biomart<https://github.com/oicr/biomart>
)*

3) Ensuring that as long as OICR is using the BioMart code base, all
funding applications and publications (current and future), which describe
ICGC's data management solution, refer to it as BioMart rather than
"SciMart", "ICGCdb", or "franchise", etc.

4) Allowing BioMart experts from outside of OICR to participate in
BioMart's development so that they may guide the ICGC DCC through the
implementation of ICGC customizations.





Very truly yours,



Arek


PS: since some of the emails were deleted from the latest "public" email
thread, I've taken the liberty of reconstructing this thread (below) for
the sake of clarity and transparency.

-- 

Arek Kasprzyk, MD, MSc, PhD
BioMart Project Lead



On Tue, Jan 24, 2012 at 9:01 AM, Arek Kasprzyk <are...@gm...>
 wrote:

> Dear Lincoln,
>
> That is great news!
>
> On a related note, one of the BioMart programmers has come across
> repository for OICR's SciMart:
>
> https://github.com/oicr/SciMart/blob/master/README.md
>
> After having looked at it briefly, the BioMart group and I have a few
> questions:
>
> 1. Below the surface (
> https://github.com/oicr/SciMart/commit/10b3f3c8c55eb56a9129bb299ff6dfd9584f7619#src/org/scimart/queryEngine),
> it appears to be a renamed BioMart.  Can we expect the BioMart name to be
> re-instated?
> 2. The code does not seem to be entirely functional but has acquired ~1GB
> of dependencies? Are we looking in the right place?
> 3. Can BioMart programmers outside of OICR review this code to help take
> it forward?
>
> cheers,
> a
>
> PS. There were some difficulties with my email getting through to the GMOD
> mailing list yesterday
>
>


On Mon, Jan 23, 2012 at 12:05 PM, Lincoln Stein <lin...@gm...>
 wrote:

> All these requests are reasonable, and I would not dream of doing
> otherwise!
>
> Lincoln
>
>
>> 1. Keep the BioMart name and logo clearly displayed on their websites;
>> 2. Acknowledge BioMart in their publications;
>> 3. Refer to BioMart in funding applications, which describe their
>> data management system; and
>> 4. Offer their local modifications to the code base as their
>> contributions to the BioMart project.
>>
>> I expect that you will do the same with the "internal version" of
>> BioMart, which will be used to support the ICGC data portal and other OICR
>> initiatives.
>>
>> Thank you,
>>
>> Arek
>>
>>


On Mon, Jan 23, 2012 at 9:47 AM, Arek Kasprzyk <are...@gm...>
 wrote:

> Dear Lincoln,
>
> The BioMart software that I released in May 2011 and it still in use today
> is fully functional and perfectly scalable as a result of the the parallel
> query engine, which was implemented some time ago.  I believe you may be
> referring to changes to the peripheral package of MLoader tools that
> I scheduled for implementation and integration with the configuration tool
> prior to my departure.  Since a) I fully support these changes; b) this
> does not contradict any of my plans; and c) this issue was never discussed
> with me, I don't see how this could possibly be a reason to split the
> development.
>
> In any case, the OICR is not the first group that has made modifications
> to their local BioMart code base. However, none of these groups felt it was
> necessary to rename their local copy of BioMart to give the impression they
> are using a different product.  In fact, they:
>
> 1. Keep the BioMart name and logo clearly displayed on their websites;
> 2. Acknowledge BioMart in their publications;
> 3. Refer to BioMart in funding applications, which describe their
> data management system; and
> 4. Offer their local modifications to the code base as their contributions
> to the BioMart project.
>
> I expect that you will do the same with the "internal version" of BioMart,
> which will be used to support the ICGC data portal and other OICR
> initiatives.
>
> Thank you,
>
> Arek
>
>


On Thu, Jan 19, 2012 at 2:07 PM, Lincoln Stein <Lin...@oi...>
 wrote:

> Dear Arek,
>
> I'm sorry you have so badly misinterpreted my intentions. I remain a
> strong fan of BioMart and am using it as a data mining engine for several
> projects, including WormBase, Reactome and a variety of cancer projects.
> However, as you know, internal changes to the code base were badly needed
> in order to support the scalability of the ICGC DCC, and OICR's attempts to
> create a constructive collaboration with you to address these issues were
> unsuccessful. Therefore we were forced to make a copy of BioMart for *internal
> use by the ICGC DCC* so that the needed code development could go forward
> without impacting the public code base. We have created a temporary,
> internal name for this version of BioMart in order to avoid confusion, and
> we have absolutely no intention to compete with BioMart's brand, name
> recognition or community.
>
> In fact, OICR has offered you financial support, hardware and
> administrative support in order to continue to support BioMart and its
> community even though you are no longer employed by us. I am perplexed by
> the content of your letter, and saddened by its tone. In the future, I hope
> we can discuss these issues person-to-person, and not in an open forum.
>
> Yours,
>
> Lincoln
>
>


On Thu, Jan 19, 2012 at 10:06 AM, Arek Kasprzyk <are...@gm...>
 wrote:

> Dear Lincoln:
>
> I am concerned about your continued efforts to rename BioMart at OICR.
> These efforts have far reaching negative consequences for many
> organizations and projects including the OICR and ICGC and deprive the
> BioMart project of its well deserved credit. In the last two years, there
> have been two attempts to disassociate the BioMart brand from the ICGC's
> data management solution. The first attempt was made in 2009/2010, when you
> started referring to BioMart in ICGC communications as the "franchise
> database system", a term that was only dropped after my continued
> persistence that the BioMart name be rightfully reinstated. Now that I have
> left the institute, the story is repeating itself and you have decided to
> continue BioMart development under yet another name. Rebranding BioMart (10
> years of your colleague's work), creates the misleading perception that
> OICR/ICGC is using a different product.  As I have noted time and time
> again, I disagree with these actions.
>
>
> I am very proud of BioMart's publication record (
> http://www.biomart.org/publications.html), the size of its community (
> http://www.biomart.org/community.html), and the fact that it is one of
> the most successful and innovative technologies in bioinformatics today. I
> have always been happy to share this success with everyone who have been,
> are and will be part of this open-source community. It goes without saying
> that this applies to you as well as everyone else at OICR who have
> contributed in the past, are presently contributing and/or will contribute
> in the future. Nevertheless, I expect that those who have benefited from
> BioMart use it in a way that is ethical and remains beneficial for the
> entire scientific community.  As mentioned before with respect to using
> and/or developing BioMart, the only options I consider fair and acceptable
> are as follows:
>
>
> 1. Quit BioMart development at OICR and replace it with an alternative
> technology (i.e. not a renamed version of BioMart).
> 2. Quit BioMart development at OICR but remain a BioMart user and let OICR
> benefit from this free, open source technology just like the other 28
> institutes worldwide who are currently using it successfully. You will have
> my full support.
> 3. Continue BioMart development at OICR in collaboration with the BioMart
> project. Your contributions will allow you to prioritize your goals and
> will be dully acknowledged.
>
>
>
> In addition to the issues surrounding the renaming of BioMart at OICR,
> duplicating BioMart and hence having two almost identical code bases is
> completely unnecessary. BioMart's data-agnostic development model assumes
> that the software is generic so that it meets the requirements of a
> principal organization while fulfilling those of other organizations as
> well. As long as these other organizations communicate their requirements
> to the BioMart team (via the mailing list, etc), I include them in my
> development plans and the additional features benefit everyone. Since the
> various group requirements often overlap and/or are already included in my
> vision of BioMart going forward, this has never been a problem to
> achieve. As you recall, when BioMart was funded by the Wellcome Trust and
> hosted at EBI, not only did the software meet Ensembl's data management
> needs but it also benefited other projects including many of your own (e.g.
> GMOD, Reactome, Wormbase, Gramene and HapMap).  More recently, when I
> began working at OICR, my primary mandate was to ensure that BioMart met
> the ICGC's data management needs.  Not only did it successful meet them,
> but it also met the needs of other OICR projects such as PopCure, OSIRIS
> etc, and continued to meet the needs of Ensembl, as well as those of  40
> other database projects worldwide. Another benefit of this development
> model is that it enables projects to utilize existing BioMart databases
> without having to commit any of its resources.  For instance, Ensembl,
> Reactome, Cosmic, PED, BCTC and Intogen are all available through the ICGC
> Data Portal. If the the ICGC's data management solution is no longer using
> the unified BioMart code base, integration of these databases into the ICGC
> Data Portal will become far more difficult. For those who understand the
> BioMart development model, it is clear that it is redundant and harmful to
> create BioMart "duplicates"/"forks"/"copies"/"clones"/"brands" that cause
> confusion and duplication of efforts. In addition, two different BioMart
> code bases create incompatibilities that impede existing and future
> collaborations and destroy the unified system that currently works so well
> for the scientific community.  It is therefore critical that we continue to
> maintain one BioMart project and one BioMart code base.
>
>
>
> For the sake of the BioMart scientific community, which currently includes
> the ICGC and other OICR initiatives, I have made numerous attempts to come
> to an agreement with you with regards to maintaining one BioMart code base,
> however they have consistently been rebuffed or ignored. In our
> correspondence on this matter, you have neither provided convincing
> arguments regarding your decision to rename BioMart, nor have you been able
> to counter any of my arguments against it. Much to my dismay, in reply to
> my latest initiative, dated December 19th you informed me, and I
> quote: "with regards to the options that you have proposed in your
> correspondence of December 12th, 2011, none of them meet with OICR needs
> for the future development of its projects. Consequently, the only
> communications OICR is interested in continuing are in connection with the
> migration of BioMart to an alternative site, separate from OICR".  I must
> say that this type of communication projects an image of OICR that I find
> very disturbing. Should you, a representative of OICR and ICGC not
> support open dialogue and foster scientific collaborations?
>
>
>
> Very truly yours,
>
>
>
> Arek
> --
>
> Arek Kasprzyk, MD, MSc, PhD
> BioMart Project Lead
>