A kickstarter approach effectively blocks people who don't have friends & family with spare cash (and whose mothers don't have houses with spare rooms either).
big fan of moving away from anything that buckets individuals into "you are deserving, you are not." creating a group around a specific coding goal provides a nice layer of abstraction away from the individual.
Another potential advantage of a model like this is that aspiring coders who "fail" to find the funding level they were looking for can be put in contact with groups than can help them refine their appplication, skill set, and goals so that they may be able to succeed in finding funding in the future.
This sort of funding paradigm could also be used to support quality educators who want to volunteer to go into schools, non-profits and other areas that are lacking skilled CS educators.
I think this sort of model could also be used "on top of" the more traditional "kickstarter" crowdfunding model. You could still allow traditional, one person to one person funding model, but at the completion of the campaign, match funding groups with individuals who have not raised sufficient funds.
I think asking aspiring coders to run a "kickstarter" style crowd funding campaign for themselves is a mistake for the reasons the jtheory mentions.
I think it would make more sense to allow funders to organize themselves into scholarship groups with shared goals.
Then match aspiring coders to these scholarship groups using a process similar to the "National Resident Matching Program".
Then there is no binary choice of "you are deserving, you are not" about aspiring coders being made by funders. It's more of a "this coder exemplifies our ideals better than this coder"
/edited for clarity