Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
Software

Drupal's Dries Buytaert On Drupal 7 55

itwbennett writes "The Drupal community has been working on Drupal 7 for two years, and there are 'hundreds of changes' to show for it, says Drupal creator Dries Buytaert in an interview with ITworld's Esther Schindler on the occasion of Drupal 7 going into Alpha test this week. Most notable for end users are 'some massive usability improvements,' says Buytaert, while site builders will see the greatest changes in the Drupal Content Construction Kit (CCK), which has been moved into the Drupal core. But one thing that hasn't changed is the not-so-easy upgrade path. 'The upgrade path for a Drupal site has never been really easy, to be honest,' Buytaert says. 'We do break backwards compatibility. It's a little bit painful because it requires all of the contributed modules — and there's 4,000-5,000 of them — to make changes.' But Buytaert doesn't think that's all bad. 'Innovation is key. Backwards compatibility limits innovation,' Buytaert contends. 'The rule we have is: We'll break the API if it makes a better API, and if it allows good innovation and progress to be made. Also: The second rule is that we'll never break people's data. We'll always provide an upgrade path for the data.'"

Comment Re:More to the point... (Score 1) 112

You're right, genetic diversity does matter, and maybe I'm a bit cavalier in saying if a drug doesn't work that means cancer is different. But I disagree that mechanism and treatment are distinct.

I would say that treatment and mechanism are related. Cancer has been "cured" in mice dozens of times and the reason those cures don't work in humans is not a dosage or a genetic issue. Rather it's that the drugs used while addressing a particular mechanism miss others. For example drugs that work great in mice for treating cancers with over-abundant EGF receptor also work in humans, but fail after a time because the cancer mutates in humans, but doesn't in mice. It's not a problem with the dose or the genetics, it's just that the mouse provides an incomplete picture.

While yes, mice and humans both develop cancers in senescence, most mouse models of cancer do not work this way because then out of a population of mice, you'd get a few old ones with cancer, and likely they'd be different kinds due to different causes. In other words, you'd have something that's hard to study and wouldn't give you much data as you would only have one or two examples of each particular cancer. Instead most mouse models are bred to be cancer prone, which gets to your genetics point. For the most part, humans aren't bred to be very cancer prone. The models they're talking about here are even worse, they're xenograft models. I.e. you take a tumor from a totally different organism, put it in that organism then see what happens. I think it's safe to say that people never get cancer by getting injected with a chimpanzee tumor.

Cancer is an extraordinarily complex disease. Genetically a mess, subtly different in every case, it's just really hard to study. And while animal models can help, I think it's hard to draw really strong conclusions from them, particularly in organisms that are very different in many ways. I'm just not convinced that zebrafish are a really good model going forward.

I guess my point about this article summarizes as cool trick, what for?

Slashdot Top Deals

"It is easier to fight for principles than to live up to them." -- Alfred Adler

Working...