So, while we think our hosting management products will be a nice source of revenue for years to come, it's the combination of those tools with cutting edge stuff in the virtualization space that will allow us to lead rather than follow (because no matter how good your products are in an established market, you'll probably always be playing catch up in the mind of the market).

I also think you're underestimating the amount of interesting work that can be found underlying seemingly mundane tasks. I've found that data analysis and visualization can be a really interesting and fun pursuit within systems management software, and it's an area that brings tears of joy to customers (seriously, people wet themselves over graphs and charts...if you also add export to Excel or CSV, there will be fainting).

Regardless of all of that, you have to solve problems that customers want solved. If you're doing that, the harder problems are better, if you are willing and able to solve them faster than your competitors.

But, if you are starting a business rather than a research project, you need to make your decisions based on what people want rather than what you want to work on. You cannot start from "here's a cool technology, let's figure out a way to sell it", at least not generally. You must start from, "Here is a market that is being poorly served, and I can serve this market well." If you can fit, "by building interesting and cutting edge technologies" into that sentence without breaking it, then you've got yourself a winner.

I have been trying to make more interesting out of the apparently mundane; reimplementing the google charting api internally cause we needed to make charts, and even some good ol' csv exports (interestingly, csv export formats that are useful for biz people are not the formats i would prefer as a programmer.. the more you know..) further, I'm focusing on trying to learn the best tools for the job -- picking up vim finally, trying to grok sed/awk, bash scripting where appropriate.. even perl is sneaking into my head.

make stuff that people want is a great mantra.

i suppose i got into programming because it was a process of discovery and a process of invention. keeping it that way seems to require intention.

May I suggest that you roll all that into one neat little endeavor of learning Python... :)

My Python code always ended up being significantly longer and slower than the same task written in Perl. Partly due to CPAN (if you can imagine it, there's at least three libraries for it in CPAN), and partly due to the nature of Perl. Perl is like a DSL built for system administration. Sometimes it gets a little clumsy on non-system administration related tasks, but when you're working in Perl's sweet spot of functionality, it's a very nice experience.

That said, Perl can be intimidating, because it is such a wide ranging language, and some of its features aren't quite right (references, and the way lists and arrays behave, can be confusing, and they're still a source of confusion for me after years of working with Perl). Python has its own quirks, and though I spent a couple of years working almost exclusively in Python, I didn't find it at all painful to come back to Perl.

Ruby, on the other hand, is a language that could probably steal my heart. Between deciding to shutdown my old company and starting up Virtualmin, I was undecided on what to do with myself. I learned some Ruby and tinkered with Ruby on Rails. Loved Ruby instantly. RoR, not as much.

My next project is definitely gonna be a for-profit in some other area. Still deciding where exactly.

Kick it up a notch and get the machine to do that for you =D

http://www.anybots.com

Anybots looks super awesome, by the way. I can't imagine you'll have too hard a time recruiting a young product designer. Seems like a dream job.

- Usually a good solution to a hard problem looks easy to users.

- Simple problems get a lot harder when they're forced to scale. Just keeping things from flying apart at the hinges often involves significant optimization and engineering problems.

On the flip side, it's my guess a lot of hackers consider themselves new-problem-solvers-in-training. A 22 year old computer guy may not be solving the clean water and efficient energy problems of the world, but could conceivably have aspirations to do so when his or her talent is matched with the right street savvy, is matched with the right entrepreneurial network, is matched with the right funding opportunities, and so on. So in a lot of cases it's a question of "when," not "whether."

And if it's people breaking new ground in computer work you're after, you've got to make sure you're in the right ballpark; Plenty of universities are doing remarkable things in CS, but again, it's usually older entrepreneurs (with money and know-how) who get their hands on new tech through the technology commercialization centers.

Which is to say that new problems are being solved (in CS, too), just in different territories and in different ways than we hear about in mainstream tech news.

There are just so many "backwater" areas of the business and science worlds, ripe for optimization, and not enough CS types with strong fundamental skills, attacking them. I guess they are all concentrated in the "hot" companies and focussed on the "hot" areas instead. =)

I was very surprised when I realized this. It was kinda akin to learning Santa wasn't real (edit: but in reverse!)

Its hard to see these opportunities from the surface, I guess they are like deep water fish - so hitch a ride with a submarine =p and start with one based on the contacts you have.

Your point about being in-training is an interesting one, particularly for the readers of this site.

Where do you find out about the remarkable things that people are doing in CS? Also, are you suggesting that most interesting new CS work comes out of universities? If so, does that mean "If you want to solve problems, go to grad school. if you want to make money, start a me-too business?"

The Queue has decent stuff, but the real meat is looking at conference proceedings.

Good question. I'd start in the pages of Technology Review.

If you're working on something in a crowded market, designing your product to stand out/work better than the rest can be just as challenging (and satisfying) as writing something technologically complicated.

I'm getting there, but I am behind schedule. I struggle with whether I should release before I've added reporting in, or if I should wait. If folks are interested, the website has a mailing list signup.

System Error

* Oops! It looks like there was an error.

I'm connecting the load tool to the site, currently, although more still needs to be done.

I wrote an article about this some time ago on my blog: http://www.inter-sections.net/2008/02/11/how-tough-is-your-p...

...no, it is not satisfying.

I guess you could consider the optimisation problem of how to finish your 'less than new' problem most effectively as a potentially interesting meta-problem.

The experience of going to a pharmacy sucks, and is obsolete. The entire medical industry is complacent in allowing 100 billion in unnecessary medical costs because of non-adherence to medications. The doctors don't consider it their job to make their patients adhere. The pharmacist maybe talks to you for a few minutes and hands you a sheet of medical and legal jargon that says, on page 3, how important it is to take medicine. The drug companies also fail to ensure adherence. Yet patients don't know enough to care.

None of these people have stepped up to the plate. So we're gonna do it, and when we solve the problem, we can start upending all the parts of the patient facing medical experience where bad decisions and stupid compromises have been calcified into the shitty experience that anyone who has had a sick family member knows all to well.

We are not exactly a web start up though. We must design some actual stuff people will use too. The computer science part of our project is that all this adherence stuff can actually reduce to a flow of information problem given a few assumptions and the right way of conceptualizing the problem from a psychological and brain science standpoint. The computer science forms the informational backbone along which we can build a medical experience that doesn't suck and isn't full of compromises that suck for the patient. From a computer science point of view, we are applying existing techniques to a new problem, but from the point of view of the medical establishment, we are solving a currently unsolved problem of huge proportions.

Information about your medication is some of the most important information you encounter, because it directly influences how long you live and how well you live. Edward Tufte is an old man now, and the pharmacies are still basically acting like it's 1952 in the way they present this vital information.

I'm not feeling inclined to just hand over all of my cites. There are too many and it would take too long to sort through. But, I'll see if I can dig up a couple of the meta studies that have further links to other research in them later tonight. If you have those and some motivation, you can fill in the rest.

How can you start a project like this without having already encountered all that research? I don't understand. I went and sequestered myself away for a month and researched and read all that stuff. The more I read about the problem, the more I was convinced my solution worked and is desperately needed. But many of the ideas I had intuitively changed when I did the boring work of reading medical journals all the time.

"How can you start a project like this without having already encountered all that research?"

My project actually has to do with parenting, not medication. I'm basically interested in applying models from many areas of academia to parenting, including the research on medical non-adherence. However I'm only interested in ameliorating one or two very specific issues, and medical adherence best practices will probably end up playing only a supporting role to research from other areas of academia. Also I'm still in the kitchen table phase of the project; I just finished my undergrad two days ago so today is really day one in the library.

If you could find the meta studies, that would be excellent though.

EDIT: I'm also interested in whether there are applications of adherence medical research in athletics, in terms of understanding why people don't do stuff (like working out) in general.

Where things are now. Check the cites/links on the bottom

http://jac.oxfordjournals.org/cgi/content/full/55/4/413

Antiretroviral adherence, drug resistance, viral fitness and HIV disease progression: a tangled web is woven. This article is great. Also it has a bunch of cites and links to other relevant issues.

http://www.ajmc.com/files/articlefiles/AJMC1995OctCramer167_...

Partial Medication Adherence, The Enigma in Poor Medical Outcomes. It's old (1995) but still relevant.

That's all for now. Dig through those and their links and you'll find a lot to get started.

Most/many diseases are caused by people not adhering to very basic eating, exercise and health routines to begin with. Couple that with vague superstitious beliefs about Drs just wanting to make money/actually hurting the patient, also the inability to pay for the medicine, etc.

I do wonder, however, if someone is not capable of taking medicine, how will they use a website/other technology any more reliably?

House: Have you been keeping up with her medication?

Woman: She doesn't use that except in emergencies. Pharmaceuticals are very risky.

House: You know what else is risky? NOT BREATHING. That has some minor side effects such as DEATH. Your doctor, after carefully weighing all the options, decided that NOT BREATHING is more dangerous than [insert minor side effect of pills].

(Paraphrased, I don't remember the exact lines.)

repeat for 20 years.

me: realizes doctor prescribes anti-biotics for allergies, the flu, and the common cold.

The problem isn't so much with doctors as it is with patients who see ads for made up conditions on TV and try to talk their way into samples with the talk to your doctor guides to scoring drugs. Ban advertising, ban bribing doctors to push pharma junk, the problem largely dissolves.

(I accidently pushed to downmod instead of upmod you, sorry. I will go upmod 2 others.)

If you think your doctor is not treating you properly seek a second opinion. But you are not qualified to make the judgements about your medication.

A second opinion? can't afford the first. I've gotten pretty good at knowing what's gone wrong with my own body over the years.

I'm no pharmacist for sure, or a doctor. But every time (that I recall) going to a doctor I get a prescription. Why is it I get a prescription for the common cold? or the flu. No prescription can cure those (at least that I'm aware of). Better to tough it out.

why is it exactly that you trust MD's so much? I don't trust them any more than I trust mechanics, and my body is more important than my car, and more complex.

This is responsible for about 10% of hospital and 24% of nursing home admittances, at a cost of 100 billion. There are many 2007 and 2008 medical journal articles about how this is a huge problem desperate for a solution. Its not that people don't want to take the med, in most cases. Its that it takes more then wanting to adhere to make it happen, and the medical industry has no good solution.