University Calendar does not fit the timeline
Harry Cutts
I am a first-year Computer Science student at the University of
Southampton, UK, and I am planning to apply for GSoC 2012. However,
the summer holiday at Southampton University is different to the USA
summer holiday. I would be able to start coding around 24th June
instead of 22nd May. However, my holidays finish on 27th September, so
I would be able to continue coding later than normal.
Is changing the schedule of my project possible, or would I have to
take on a smaller project in order to finish on time?
Thanks,
Harry Cutts
Haris Ibrahim K. V.
These are my personal thoughts regarding this matter. GSoC has a specified timeline and you should abide by it. The milestones being the mid-term evaluation and the final evaluation. You should have results to show during these two deadlines. Otherwise your project maybe marked as incomplete.
You should be able to devote at least 40 hours a week for the project you have undertaken. Even though slight adjustments are possible, the adjustments should be contained within the deadline (mid-August) when Summer of Code 2012 officially ends.
Skipping a month does not seem to be apt for participating in GSoC. So give it much thought before undertaking.
Regards,
Thanks,
Harry Cutts
--
You received this message because you are subscribed to the Google Groups "Google Summer of Code Discuss" group.
To post to this group, send email to google-summer-...@googlegroups.com.
To unsubscribe from this group, send email to google-summer-of-cod...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/google-summer-of-code-discuss?hl=en.
--
David Palacios
David Palacios
Mashiat Sarker Shakkhar
Hi Harry
It really depends on the organization that you are working with. I'd
recommend you to discuss with your potential mentor - don't lose heart
already. Some orgs will allow you to start early and continue working on
the project afterwards. Sometimes you may get by without working for a
few weeks if you put in some work early and work overtime later. It all
boils down to - i. you proving to your org that you are capable of
finishing the project in time; ii. you are committed to the project and
will stay in case the project is not finished in due time; iii. you have
already done some work to support your claim.
Being too strict about timeline in open-source is kinda suicidal for an
org in my opinion (but this is just my opinion).
Best of luck
Shakkhar
Libav
Joel Sherrill
around holidays
during the work period, there are two deadlines in the program which do not
move under any circumstances. You must satisfactorily meet your mid-term
and final goals on time.
IMO mentors are not supposed to pass you on the good faith that you will
finish it later. There are no "incomplete" grades. You pass or fail based on
the situation on the mid-term or final evaluation date.
If you work past the deadline to do something like polish that improves
the work, that is because you like to do it. And I would encourage you
to love the work
and want to stay involved.
I would encourage you to set your goals modestly and schedule your deliverables
so you can meet the deadlines. Don't ignore GSOC for a month but work in smaller
productive periods. Focus in the first month on important deliverables
like design that
can be reviewed or proposed APIs for review. You should have a good
plan as part
of your proposal and that month could be well spent "in thought"
laying the groundwork
with lots of feedback for a huge burst of coding when school ends.
But ignoring GSOC until a couple of weeks before the mid-term evaluation smells
to me like a recipe for disaster.
This is all my personal opinion. It does not reflect Google in anyway. :)
--joel sherrill
RTEMS
Shahzad Saeed
are pursuing your GSoC project. This would be better and practically feasible rather than missing your GSoC program.
--
Shahzad Saeed
Blog: www.TechandProject.com
Author of "Definitive Guide to Google Summer of Code" ebook (Old GSoC students recommended):
www.GSoCguide.com
A. Hunt
whose academic schedule (which has considerably greater priority)
simply does not correspond to what is currently assumed. At Cambridge,
similarly to Southampton, the term ends in the middle of June, the
next term starts in October -- i.e. my vacation is exactly one month
behind the GSOC programming time. It would be ridiculous to expect me
to start programming in the middle of May since that is just before my
exams start, and those (obviously) take precedence. It's not skipping
work -- I'd still spend the same amount of time programming -- simply
starting and finishing at different dates than most of the others.
It's just that the assumed dates do not correspond to reality -- I do
not have the choice or possibility to tell the university to change
this. I don't see how it would be a problem for GSOC or mentoring
organisations to cope with a different set of start/end/evaulation
dates (some mentors might also be affected by similar issues with the
dates). At the very least two alternative schedules should be created
(e.g. one as exists currently, one delayed by a month). The question
is whether GSOC wants to be accomodating for students at universities
with such schedules -- and I doubt such students are a rarity since
most UK universities have a (roughly) similar calendar.
So, is GSOC considering changing this, and if not, why not?
Carol Smith
Cheers,
Aaron Kalair
and face the same issues with regards to timing.
Thanks
Aaron
Matthew Wilkes
Before we get into this: I am not a googler, I don't speak for Google. These are my opinions. And I have a lot of opinions.
On 2012-03-21, at 1959, A. Hunt wrote:
> I don't see how it would be a problem for GSOC or mentoring
> organisations to cope with a different set of start/end/evaulation
> dates (some mentors might also be affected by similar issues with the
> dates).
It wouldn't be much of a problem for the mentor, but the higher up the food chain you go the more complex it gets. The organisations would have to spend a lot more effort on assigning mentors. At the moment we have a set period of time that we can recruit volunteers for, then match them up to where they would do the best good. If we have a sliding timescale then we would start having the best mentor for the job only being available for 75% of the timeline, and we'd have to try and work out if we should put the great mentor on the project that fits her time best, or her interests best.
Then, you have the communication aspect. I've admined for a medium-sized organisation a few times, and keeping the state of all the projects in my head was hard enough without having to keep track of what timetable the student is on too. It might not seem like much, but it makes the difference between being able to ask if they've started the evaluation or not in passing and having to look it up in a calendar. For 1 student that's no work at all, for 100 it's a hell of a lot.
Then, going up another level, you have Google. First off, the software would have to be updated to allow for arbitrary timelines per student, that's a fair bit of work, but it's a one-time thing. More importantly, when certain bits of the timeline are reached money has to change hands. At the moment, that's a big batch job of processing instructions to a payment company, but there are subtleties. For example, you need to be a student on a given date to be eligible, but what if a student's timetable is such that he can't start working for 2 months after the current start date? Maybe even because he is graduating? Should that student still be eligible?
With over a thousand students potentially participating in the programme there is a good chance that there will be a skew of up to 2 months, meaning some students are already past mid-term by the time others are starting. That leaves different payment amounts going to cards depending on what timetable has been agreed upon, making tracking them very difficult. The reason VISA payment cards are used is because something as simple for one person as an international bank transfer becomes prohibitively difficult when you're doing it with a few hundred subtle variations. This would prevent Google being able to treat the cards as one group, effectively obviating that advantage.
Finally, OSPO runs multiple programmes. There is also GCI for high-school students which roughly runs to be an inverted timeline, part of the reasoning for this is that some geographical areas find participating in GSoC disproportionately difficult due to the time of year, so GCI hopefully coincides with the summer vacation for people in the southern hemisphere (and, to a lesser extent, christmas vacations in the northern). This isn't fair on everyone, but it's a step in the right direction at least. Secondly, the amount of work in organising GCI is huge, and the less overlap between it and GSoC the better. I don't know what the current schedule is for the internals, but I imagine the OSPO staff are coordinating the mentor summit and pitching to their bosses to run GCI around the same time – it'd be even harder if they were still having to manage student payments and evaluate mentor org performance.
Now, one way to solve all this would be to get more staff on board to administer the programme in different countries to handle all the local variations, but I can't imagine that would come cheap. Let's do some back of a napkin calculations:
GSoC 2011 cost 7.2*10^6 USD to fund 1115 students over 175 orgs. More than 10% of those students failed, so didn't receive the full payment, but let's simplify here and ignore that. If you use the same stats from previous years you end up with:
7.2*10^6 USD = 1115s + 175o + a
5 * 10^6 USD = 1026s + 150o + a
That means that
89s + 25o = 2.2*10^6 USD
The total payments for a student is 5500 USD (stipend + mentoring org payment), so of that additional 2.2 *million* dollars around 0.49 million dollars went to students. A good portion of the rest will be on the additional travel to the mentoring summit, of course, as will the fact that this hides that admin costs aren't the same year to year, but I think it gives you a scale of just how much it costs to run a programme like this. I can't even begin to enumerate all the myriad things that must happen behind the scenes to make something like GSoC work.
Given the sheer scale of these numbers I'd be surprised if the most cost effective means of improving diversity in participation is to increase the admin overhead, I'd think that what happened in 2011 was right, and that the magic is to try and get as many students involved as possible.
In closing, if someone really cares a lot about this and can come up with an analysis of university terms around the world that shows when the optimal times to run the programme (assuming one, fixed timetable) are to promote a diverse, vibrant mix of participants then I'm sure the big G would send you some schwag. I can promise that Plone would!
Matthew
Konrad Reiche
July, roughly speaking. I've moved more than half of the courses I would
have attend in summer into the previous semester in order to deal with
the required time for this program.
You have to come up with a solution, in the end it has to be
communicated to the organization. After all students were able to attend
from anywhere in the world the last years, I don't think they all took a
break and didn't attend any courses.
Kevin Smith
<konrad...@googlemail.com> wrote:
> Here in Germany the summer term is from April to July with exams in July,
> roughly speaking. I've moved more than half of the courses I would have
> attend in summer into the previous semester in order to deal with the
> required time for this program.
>
> You have to come up with a solution, in the end it has to be communicated to
> the organization. After all students were able to attend from anywhere in
> the world the last years, I don't think they all took a break and didn't
> attend any courses.
I have mentored European students in the past, and have been able to
work things out with them such that they could work within the dates
imposed by the schedule. Obviously saying "I won't start work until
halfway through the summer" wouldn't go very far, but mentors are
usually flexible people who want to work with you if you're flexible
and want to work with them.
YMMV.
BBQ.
/K
A. Hunt
programming/evaluation not be moved back 1 month (while keeping
applications and contact times the same) -- and then the student/
mentor given the power to decide when exactly they start/finish (i.e.
an extra month buffer where the start of programming currently is)?
The major issue being that the first evaluation date is too early for
all those who aren't fully free until June, with the August end date
also being less than ideal for those who want to work for longer. On
the other hand, those who want to work May-August will only be
impacted by having their evaluation submitted to GSOC (and payment)
later than expected. I.e. the later the evaluation date the more
flexibility there is for people from different universities to
participate?
And yes I do have to say "I can't start work until the middle of June"
since I need to pass exams which I cannot move. I am going to start
early with programming (i.e. during the next few weeks) but that's
still quite different from a month of being able to work full time on
the code, and making the first deadline will be a real rush (assuming
I get selected -- if not I'd still be doing the same project, but at a
more relaxed/sensible pace).
On Mar 21, 9:22 pm, Matthew Wilkes <matt...@matthewwilkes.co.uk>
wrote:
Kevin Smith
> And yes I do have to say "I can't start work until the middle of June"
> since I need to pass exams which I cannot move.
This is entirely at odds with
> I am going to start
> early with programming (i.e. during the next few weeks)
This, which is an example of some flexibility you could negotiate with a mentor.
I suspect you'll find very few mentors who don't believe that
school/college/uni (delete as appropriate) work should come first, who
expect you to be able to move your exams, or who are spending time on
GSoC because they /don't/ want to work with students. Most mentors
have probably even been students themselves at some point or another
(I have no stats to back this up).
From my point of view, and I suspect many other mentors, the point of
GSoC is not to get students to do work at Google's expense (it's not
atypical for the overhead of mentoring a student to be greater than
that to just implement the student's project oneself), so I wouldn't
worry too much about how much exams will effect fewer or smaller
deliverables, I'd just go ahead, find mentors for ideas that interest
you and get chatting. Situations like yours don't preclude successful
participation in GSoC.
/K