Sunday, September 14, 2008

Two basic problems

So, I constantly find myself banging my head against the wall because of two problems with the current BCI project. No, nothing juicy. Here they are, and let me know if you have any REAL way around them:

1) Patient data is strictly controlled due to obvious reasons. Why is this a major problem? Most of the really interesting and really applicable solutions to data management require some aspect of working in 'the cloud' - servers that are off site and under the management of some other entity. Examples would be Google, Amazon's S3, Drop Box, Microsoft Live Mesh, Apple's MoleMe (we're sorry - we lost your 'bi'. Here are two months of service for free!). They would be incredibly useful for us, but we can't be upping videos showing the participant's face (fine for our private meetings, but a big no-no for public viewing without signed consent). I use Live Mesh right now, and love it. I'm thinking of possibly switching to DropBox, since it is cross platform and offers similar features (folders that get synced automatically and available from any web browser).

2) Apple's obnoxious resurgence in the academic market and annoying ability to inspire some creative apps. Okay, kidding, somewhat, but every Mac owner I know tries to avoid using any Windows programs, even though they're quick to cite the Mac's ability to run Windows. "I don't want a Mac because there are way more programs available for Windows!" "That's what's great about Mac OS X... (insert latest Steve Jobs keynote speech verbatim)." VMware and Parallels are great, and a huge step beyond the days of VirtualPC (remember SoftPC?), but you're still settling for less than optimal performance. And there's that indescribable feeling that running another OS inside a primary OS just feels like whatever you do in your encapsulated OS is for simple utility. Anyways, the project is pretty evenly split between Macs and PCs (though more Macs at Brown). Macs tend to have better support in newer, flashier, and better thought out services and apps, at a price that is 10x's the comparable ugly PC app with the same features. PCs tend to have 15,000 apps to do any job you might ask, but you might switch around between them a little before settling. Those two cultures are mutually exclusive.

Wrap those two above problems in wanting some way to export everything in some standardized form that won't be obsolete for at least 15 years, and you got a total tech headache.

There are two things that would completely alleviate these issues:
- For issue 1 - strong encryption that is automatic end-to-end, with independently audited offsite servers. In order for this to be viable, I believe some regulations would have to be eased, which would take years, and requiring government specifications would result in ridiculous prices.
- For issue 2 - allow Mac OS to be virtualized. Most PC users aren't against installing MacOS. We just don't want to get a whole damn computer to run it and be stuck with it 24/7. You can't even hackintosh it appropriately (believe me, I tried and gave up after 2 days). I could dual boot with a patched install, but then running virtualization with access to my Windows feels very very risky. Oh, look at that. The hard drive controller isn't supported on Leopard, so every file has one bit shifted randomly. Too bad. No support.

We could implement some self hosted solution, but that requires a full time administrator with all types of certs. As it stands, we have a mix mash of services and techniques, and it all makes me uneasy. I need my tech organized and accessible. Like this (new office space :D )......


No comments:

Post a Comment