It is definitely possible. I transferred with a 4.0 and still have one until this semester. Started 400 level courses this semester and pretty sure I am going to get 1 or 2 Bs. My Operating Systems professor is extremely difficult. Just retired from IBM, never taught before, and is extremely technical. Theory of Computation is just a legitimately difficult course.

I've heard view model or MVVM (if you're doing 2 way binding) thrown around if you're talking about just a view and a model with no controller. It doesn't really matter what you call it though.

There is nothing as hard in the programming world as figuring out how to get a movie from your iPad to your Mac then cut a few seconds off the end and beginning. Apple puts me into an instant blind rage every time I have to figure some new **** out like this which should be their bread and butter (Grammy wants to get her Ipad video onto her shiny new Mac and edit it a little). It blows my mind how hard they make some of the simplest things.

Is that hard? I've done that and I'm an incompetent when it comes to things like this. Isn't it just automatically imported into iPhoto/iMovies and from there its like a simple cut or something.

Ok here are the steps gleaned through much googling:

1. Install google drive or dropbox (need iOS7 for dropbox) on iPad
2. Upload video to one of those
3. Download video to Mac (the fact that Nana can't just easily just get her movie of her grandkids onto her mac in one step boggles my mind)
4. Open video in iMovie
5. Create *new* iMovie (this is the step I was stuck on for a while)
6. Copy existing movie into new movie (must copy all of it)
7. In new movie, trim frames from the beginning and end through some kind of weird select/crop that can only be learned through googling (and even the guy in a video took a few tries to get it right)
8. Click share button (there is no Save or Save As - had to google this one as well)
9. Share to filesystem (unless you just wan to go straight to youtube)
10. Upload to youtube - but note the Movie will appear immediately in finder, but won't be complete until iMovie pops up a window to tell you so - so if you upload to soon Youtube will go through all the motions then fail in the processing step.

And Voila - just like the commercials with the happy multi-ethnic 20-somethings effortlessly creating beautiful art with their mind-meldingly simple-to-use Apple products.

Does the ipad not have a free app for simple video editing?

If not, you should make a startup out of it and like usual just send me 50% of your profits for the idea.

Good point. Maybe they do.

Does anyone have some advice for getting started in parallel programming? I've taken the follow up to CS50 for systems programming (CS61 through the extension school), but I'd like to get a bit more experience before taking an HPC class next semester. Can I learn the basics on my macbook air or do I need to buy a GPU computer? Should I start by joining an open source project or going through a book, and any recommendations for either?

I've been looking into GPU programming lately. It's my understanding that you should be able to access the macbook's GPU, but which platform you can use depends on the card. CUDA is less flexible, but easier to learn, but it requires an NVidia card. My '10 air has an NVidia card, but my late '12 mac mini uses Intel, so I'd have to use OpenCL to access it. Also, some versions of MacOS had a GPU problem, so I went ahead and upgraded the 'mini to Mavericks, possibly my first OS upgrade ever as I tend to be very lazy about that sort of thing.

That said, I'm strongly considering buying a PC + GPU. It's unfortunate I can't just buy a thunderbolt GPU for the mini, or even a USB GPU, but since most people need high throughout for graphics, no one has been quick to market one. I think you can do this by running windows with bootcamp, but I'd just as soon have the extra machine sitting in the corner crunching numbers.

There's the Udacity class

Yeah, I found the platform a bit tricky to debug/test things out in. Did you go through that class? What did you think?

The default iOS video player can crop videos, all you have to do is drag near the ends of the video until you highlight the part you want to keep, then hit the trim button. Also, I don't understand the whole uploading to dropbox/google drive thing, image capture pops up automatically when I plug my phone in and from there I can download videos/images.

On a Mac? If so please describe how you download images from your iPad. iTunes is the only thing that pops up for me.

No, but I completed 5 other Udacity classes and enjoyed them all a lot. I did the first week of that one but didn't have the time to continue and that post reminded me about it.

You might be right about the platform, all my other classes used the in-browser python, and the browser Cuda seemed about the same, but it might not be as simple to set up an environment on your own computer. I haven't tried.

I certainly agree that if you become adept at one it makes the next one easier. I don't think it hurts not sure if it really helps that much though. I shouldn't discourage your efforts either.

I'll bite, why would you need a GPU to do parallel programming? It is entirely possible that I am under a misconception on what you mean regarding parallel programming. Here is my conception in a simplified example, take a polynomial x**2 + x*2+1. Computing the value for a given x involves two multiplies and 2 additions. However, by factoring it into (x+1)(x+1) only two additions and one multiply are needed. Computing the individual factors can be done in parallel thus speeding up the computation even more.

Lol. I don't actually know. I'm aware that one can learn parallel programming on any modern cpu, but I don't have enough expertise to know whether GPU programming is very different or very important for machine learning (what I'm studying).

Yeah, but this falls apart once you begin to consider n-sized polynomials. There's some cool algorithms for that though.

I would think it would be much better to learn about concurrency before parallel computing. Topics like what the Actor Model is and the short-coming and what functional programming is and how it tries (and fails) to solve the large short-coming of the Actor Model isn't just ivory tower topics anymore. Not only that, you won't sound like an idiot because you use concurrent programming and parallel programming interchangeably.

SICP gives an amazing meta study of this topic... in 1985.

File uploads in one page apps (particularly with the way rails handles file uploads) are a massive pain in the ass.

Handling file uploads is like one of the worst parts of making websites

i've not used it myself, but this site has gotten a lot of press and might be worth checking out:

https://www.inkfilepicker.com/

If there's one thing I've learned, it's that I don't google enough for 3rd party solutions. This is pretty slick.

I have no idea what I'm doing, but I found it no trouble to get file uploads working with the Dojo JavaScript library in my single-page Dojo app. Admittedly I have the luxury of not caring about "other browser" support, but I think it should take care of that anyhow. I used PHP for the backend, don't know Ruby.

I can't imagine paying for a file upload as a service unless I had really specific needs.

What's wrong with using something like https://github.com/blueimp/jQuery-File-Upload?

I took the Udacity CUDA programming class when it first ran and recommend it. It's challenging and time-consuming, but worth taking if you are curious about GPU programming.

As far as applicability to machine learning, Andrew Ng's team used several GPUs to train a huge neural net that ended up recognizing cats on YouTube videos. Found the article:
http://www.wired.com/2013/06/andrew_ng/

adios - The motivation for GPU vs CPU parallel programming is that a modern GPU has 1000+ cores, while CPUs have 4-8.