Summer of Code 2006
-------------------
Last year for SoC 2005 LispNYC did extremely well, grossly
outperforming organizations far larger than us. This was primarily due
to the _quality_ rather than the quantity of submissions, as LispNYC
mentored the candidates from the beginning to make sure that all
project submissions were: gramicillay korrect, well planned and well
thought out.
This year I think we should follow-up on the same vein in an extended
manner...
Rapid Candidate Prototyping: Starting as quickly as possible (read: two
weeks) that LispNYC set up a program in which candidates start their
projects in return for a guaranteed payment, assuming that they're
accepted by LispNYC. Essentially SoC candidates spend their time
creating a prototype of their desired SoC project.
What the candidate gets:
* Early access to the talent that is the Lisp community.
* A seriously improved chance of getting funded by Google as they
have accomplished a WORKING PROTOTYPE by the application date.
* Guaranteed funding of at least $500.00, either from LispNYC or
Google.
* Smaller commitment quantum. Minimum commitment: the student is
only committed till SoC starts.
What LispNYC gets:
* Early access to the talent pool.
* Much better chance of getting projects funded.
* Filtering of candidates: from this we can more easily determine the
doers from the talkers.
* A working prototype which can be used by other candidates for their
SoC proposal.
* Marketing, marketing, marketing. This early start will most
certainly turn heads and attract the cherry applicants who are not
waiting till the last minute. A thoughtful applicant is a good
applicant.
The Cost of Failure
-------------------
For an unmotivated SoC student who successfully convinces LispNYC and
Google into acceptance of their project, the cost of a failed SoC
project to them is:
* nothing
Ok, loss of esteem by ones peers is offset by the fact that as
students, they have less vested interest in this. Add that to the
fact that they've netted $500.00 and it's a pretty good deal.
However looking at the cost of a failed SoC project to LispNYC, we
have:
* A failed project. Simply put the work is never done and both us
and the community have missed out on a great opportunity.
* Personally wasted effort. LispNYC really tries not to fund idiots
and we put serious effort into ensuring this, but given the sheer
magnitude someone is bound to slip through. So it's sad to see the
hours of time that you've devoted to a project slip into a black hole,
especially when it could have gone to....
* Another candidate's loss of success. For every failed project,
there is another candidate who could have delivered 80% of the desired
features rather than the 100% you asked for.
The Uncertainty Factor
-----------------------
"How hard are sockets?" ...if you've never tired it, then an
appropriate answer would be: "Pretty darn hard."
Why? Because if you don't know, then you're looking into the face of
uncertainty. A great result of the "Ship every two weeks" tenet of
Extreme Programming is a mitigation of uncertainty. In both the
technical and business realms, it's far better to end an unsuccessful
project in two weeks rather than in six months.
Rapid Candidate Prototyping gives us a window into the unknown and
would allow us to mitigate the risks with our projects and push us a
step ahead. Simply put, it makes us be bolder developers because we
now know more of the unknown and can either use this to better target
our release or significantly broaden our technical horizons. Either
way we win.
The Winter of Code
------------------
If successful, this project allows LispNYC to extend Summer of Code way
beyond it's horizon. One can easily foresee that SoC 2007 candidates
would be planning their summer projects in October of 2006. This means
that much more thought and planning go into the projects as well as
code. This raises the bar of entrance and significantly raises the
quality too, because candidates suddenly have many weeks of effort
devoted to their "summer" projects, mentors have something to do over
the long winter months and ultimately more code is written.
As an added bonus, this project conveniently solves the problem of
"What should LispNYC do with the $4500.00 of Google grant money?". So
far, all of our leading ideas involved the funding of projects (see our
long-winded Winter of Lisp discussions), but the biggest problem
project funding, was that there is only enough money to fund one
typical SoC project! Rapid Candidate Prototyping gives us a great
opportunity to spread the love, hedge our bets, get a head start on our
competition and more.
Let us talk,
- Matt Knox and Heow Eide-Goodman
LispNYC.org
Heow wrote:
> Summer of Code 2006
> -------------------
>
> Last year for SoC 2005 LispNYC did extremely well, grossly
> outperforming organizations far larger than us. This was primarily due
> to the _quality_ rather than the quantity of submissions, as LispNYC
> mentored the candidates from the beginning to make sure that all
> project submissions were: gramicillay korrect, well planned and well
> thought out.
>
> This year I think we should follow-up on the same vein in an extended
> manner...
>
> Rapid Candidate Prototyping: Starting as quickly as possible (read: two
> weeks) that LispNYC set up a program in which candidates start their
> projects in return for a guaranteed payment, assuming that they're
> accepted by LispNYC. Essentially SoC candidates spend their time
> creating a prototype of their desired SoC project.
>
>
> What the candidate gets:
>
> * Early access to the talent that is the Lisp community.
> * A seriously improved chance of getting funded by Google as they
> have accomplished a WORKING PROTOTYPE by the application date.
Is that always possible? Hell, that is not even how I for one work.
"Ship every two weeks"? I might spend months on groundwork for later
payoff (scale down as needed to summer-size vs enterprise projects).
> * Guaranteed funding of at least $500.00, either from LispNYC or
> Google.
> * Smaller commitment quantum. Minimum commitment: the student is
> only committed till SoC starts.
>
>
> What LispNYC gets:
>
> * Early access to the talent pool.
> * Much better chance of getting projects funded.
> * Filtering of candidates: from this we can more easily determine the
> doers from the talkers.
> * A working prototype which can be used by other candidates for their
> SoC proposal.
> * Marketing, marketing, marketing. This early start will most
> certainly turn heads and attract the cherry applicants who are not
> waiting till the last minute. A thoughtful applicant is a good
> applicant.
>
>
> The Cost of Failure
> -------------------
>
> For an unmotivated SoC student who successfully convinces LispNYC and
> Google into acceptance of their project, ...
did you mean "who ..convinces LispNYC /but not/ Google..."? Just
guessing based on the rest.
> ...the cost of a failed SoC
> project to them is:
>
> * nothing
>
> Ok, loss of esteem by ones peers is offset by the fact that as
> students, they have less vested interest in this. Add that to the
> fact that they've netted $500.00 and it's a pretty good deal.
So that is how LispNYC has decided to spend the $4500? Of course, the
only outlay (IIUC) is for LispNYC "yes", Google "no" projects, so
probably only a couple.
> As an added bonus, this project conveniently solves the problem of
> "What should LispNYC do with the $4500.00 of Google grant money?".
Ah, OK. Nice problem, btw, $4500. :) I see no harm in saving it until An
Unmistakeably Good Use For It comes along.
The presumption is that money will attract better candidates. I think it
worked the other way last time, and this approach will only increase the
chances of a repeat. Meanwhile, Google has the wherewithal to handle
disgruntled applicants. Hell, does LispNYC even have the legal status to
enter into such contracts? If so, seems to me there will be legal
paperwork upfront and disputes on the backend if things go badly. I was
glad last year that Google was the actual contractor.
ken
--
Cells: http://common-lisp.net/project/cells/
"Have you ever been in a relationship?"
Attorney for Mary Winkler, confessed killer of her
minister husband, when asked if the couple had
marital problems.