[Rd] OT: authorship and contacts for releasing packages (Re: reshape scaling with large numbers of times/rows)

Hin-Tak Leung hin-tak.leung at cimr.cam.ac.uk
Fri Aug 25 20:53:19 CEST 2006


I think you have fundamentally missed Prof Ripley's point about using 
non-affilated address while requesting help of a substantial nature.
Open-source software doesn't grow on trees.

Often some pieces of open-source software can have substantial
commercial interests, and those who want something done can afford
to pay for it in terms of consulting contracts or sponsorship for 
activities like the useR conference, etc. But some chooses not to,
just because they can hide behind non-affilated addresses. I don't know 
what open-source mailing lists you participates in, but on those with 
substantial commercial insterests (e.g. the databases ones, like MySQL),
people do offer and ask for contractual consultancy tasks.

Mitch Skinner wrote:
<snipped>
> I have to disagree with this one.  I have a hard time imagining anyone
> else at my work getting into the R internals.  However, even though I
> don't own the work I do for my employer, I do feel attached to it enough
> to help people with it down the road even if I change jobs.  So in my
> case the best chance of continuing support is my personal contact
> information.

That's a somewhat arrogant attitude - if an issue with your work
is important to them, they will find someone with suitable skills to 
rectify/extend the work without your help, and such person can usually
be found. It is sad, but you are usually not the only one in the world
who is capable of fixing problems or extending functionalities of
your old work. You are just the cheapest.

If somebody with suitable skills cannot pick up a piece of your old
work where you left off without communication with you, then the work is
not-self-documenting and you have not documented the work properly.
And that, says something about the (lack of) quality of your old work.

I am sure in this context, Prof Ripley would be proud that if he chooses
not to touch R ever again or hear about it from this moment, the
R-project can go on without him, and that's a testiment of the
quality of the work.

> I also disagree with the suggestion that my first messages were
> anonymous because they didn't include an affiliation.  I know that a lot
> of people find it easier to think primarily in terms of organizations,
> but in my opinion one of the big strengths of free/open source software
> developed collaboratively over the network is that individuals can get
> involved.  This may be a little grandiose but I think there's a secular
> shift (especially in software) away from the formal organization and
> toward a less formal community of interested people.  I think the
> individual is primary, rather than the organization.  Point being, it's
> not anonymous because I put my name right on it.

But you were *asking for help* anonymously, not *offering help*
anonymously. If you were *offering help* annonymously (and you
have a stable well-paid job to afford that), I congratulate you.

> As for whether or not that's unprofessional, I would hope that people
> judge the quality of the work by the work itself.  As you know, with
> free software this is much easier to do than it would have been
> otherwise.

...and 2/3 of projects registered on sourceforge never goes beyond the 
"concept" stage. It is like "I have this idea, but I don't the skill, 
and if I throw it out, somebody will do it for me".

> I've seen people on the linux kernel mailing list complaining that
> contributors are often hard to track down after the fact.  Obviously, as
> projects like R grow, they're going to attract contributions from a
> larger number of people.  I'm just trying to spare people the potential
> effort of tracking down the ex-boss or brute-force googling.

...and most of those are the people *asking for help*, for free...
not like "oh, I can't find the contributor, but I have fixed/extended
it anyway, can one of you check and verify this change and commit
it upstream anyway?"...

I do believe Prof Ripley made a valid point: if you ask a question
that is likely of substantial commercial interests, or the answer looks 
as if it will form part of a commercial product or venture, you have
to declare your affliation.

HTL



More information about the R-devel mailing list