As some of you may have noticed,
this year's student stipends are calculated based on each student's individual location - the base stipend is set to $6000 for the US and is then modified, with the stipend amounts being clamped at $2400 on the lower and and $6600 on the higher. While there's been a fair share of complaints (in the Facebook group and elsewhere) about the idea of country-based stipends, I think that it's fundamentally good: a flat rate makes the stipend equivalent to a mid-level software development job in some countries and doesn't even pay for rent in others. The PPP modifiers are also fine; they appear to be based on a proper cost of living index and sound about right.
The baseline stipend amount ($6000) is indeed the problem - it is
just low. Stipends in most, if not all, countries are now much lower compared to an average, non-competitive, entry-level software engineering job. In Ukraine, where salaries are as low as they get, a $2400/4=$600/month salary is still lower than the ~$800 median for a generic entry-level job. EU stipends (even for countries where the stipend's increased from last year's) are
still 2-4 times less than the respective average salaries, and the US baseline ($6000/(4*4*40)=~$9/hour) is
still close to minimum wage. And it's not like there's no room to straight up increase the baseline:
a quick Google Sheets-based calculation suggests that (assuming the same country/student distribution as GSoC 2016) every stipend could theoretically be
increased by 43% and the total budget would still match last year's.