In the news today:
“Governments, IOC and UN hit by massive cyber attack” (BBC)

How did the attack work? In a mind-numbingly ordinary way:

“An email would be sent to an individual with the right level of access within the system; attached to the message was a piece of malware which would then execute and open a channel to a remote website giving them access…they sometimes embedded themselves in the network and [tried to] spread across different systems within an organisation.”

In short:

• A person with broad authority ran a bit of code.
• The code, operating with this broad authority, wreaked havoc.

Quiz questions:

1. Why did the code inherit the person’s authority?
2. Is there a good reason for allowing this?
3. In the current model of computation, is it easy and natural to grant limited authority to individual computational objects?
4. What alternative model of computation (not an added security layer!) makes it natural to grant limited authority? What is it called? (Links, please.)

Questions for thought and discussion:

1. Why does the current computational model grant authority in this indiscriminate way? How does this lead to “sandboxing”?
2. What would be the main costs and benefits of moving computation toward the alternative model? How would this model play with the existing software base?
3. What are the leading implementations of this model today, at the language and operating system levels? In your opinion, should they be promoted more vigorously?

</lazy_quiz_mode>

Metamodern had vanished at the end of last month while I was traveling, and for a week or so I forwarded it to this stand-in page. As you can see, the blog is now up and running.

The stand-in page outlines a (re)emerging software technology that deserves several orders of magnitude more attention. Current software is mostly built on mud, but there’s is an option to use brick, instead.

Metamodern got wiped and (apparently) replaced by botware a few days ago after a visit logged as

173.234.47.14 – - [31/Mar/2011:16:58:20 -0700] “POST /b/wp-comments-post.php HTTP/1.0″ 200 0 “-” “Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.14) Gecko/2009082707 Firefox/3.0.14 (.NET CLR 3.5.30729)”

From then until now, I’ve been in Meridian, Mississippi, with sporadic internet access and recovery and reconfiguration still in progress.

A bit of snooping reveals that are some not-nice people associated with the region of IP address space around 173.234.47.14, but the real problem, of course, is the fragility of software that allows bits of code to grab (and trash) resources that they had no need to access, and should never have been able to access — not blocked by painful programming discipline, or security alert boxes, etc. but by the nature of the computational medium.

A better model of computation

It turns out that one can structure programming languages so that the creation of an object, together with subsequent messages to it, provides the object with access to resources that are held by the creating or calling object — and nothing more. Each object, in effect, runs on a virtual machine operating in a world that contains a fraction of the machine’s resources, and doesn’t, by default, allow access to arbitrary files, messaging over the internet, etc.

The mess I’m cleaning up, and a lot of pain (and incalculable opportunity) cost in the world, stems from a failure to recognize how much better we can do with languages that aren’t leaky mush.

There’s a Javascript-based implementation of these ideas, Caja, reportedly being adopted by MySpace, Yahoo!, and Google.

Here are some articles on the computational concepts:

• Object-capability model
• Capability-based security

A very big deal

Neither of these articles does a good job of explaining the qualitative differences that result from cleaning up the object model of computation in the way described, or why this matters much.

This line of development is, in fact, very, very, very important, and I urge anyone interested in making computation work better to dig deeper until “very, very, very important” seems like an understatement. What is hard to convey is the magnitude of the opportunity costs of the present morass, because the potential benefits of a better computational world are, by nature, unseen.

Note that most of the criticisms of the capability model are based on either misunderstandings or the idea that it is supposed to solve problems in other domains (for example, human access control). It’s a computational model with powerful security and stability properties woven into its fabric — nothing more, and nothing less.

This news just in:

A ‘buckyball’ — a spherical molecule made up of 60 carbon atoms — has been turned into a vial just big enough to hold a single water molecule….

The authors say that uses for the vial could include acting as a carrier for drugs in the body.
(News item)

As The Onion might ask, What do you think?

As I’ve said, scientists are held to a high standard when talking about scientific results with their peers, and a noticeably different standard when talking about potential applications with everyone else.

This just might be a cumulative problem for the credibility of science regarding (for example) climate change, and maybe even the potential of nanotechnology.

Gary Taubes has started a blog, and his first post nicely summarizes his case against the idea that overeating causes obesity: A tiny caloric imbalance of ~1% (only while actively gaining weight, of course) isn’t a cause of obesity, it’s a consequence of the onset of obesity, which is itself best understood as a consequence of physiological dysregulation of fat storage.

His upcoming book is titled Why We Get Fat: And What to Do About It.

Full disclosure: I’m thin, so my interest in this largely academic — I’m fascinated by how science sometimes goes off the rails (and stays off for decades), even when the failure is both obvious and immensely costly.

As a meta-oriented post, Metamodern is pleased to report a meta-meta-analysis.

In this month’s issue of the CDC-sponsored journal Preventing Chronic Disease, we find, published as a “Systematic Review”:

Quality of Systematic Reviews
of Observational Nontherapeutic Studies

…Of the 145 systematic reviews we found, fewer than half met each quality criterion; 49% reported study flow, 27% assessed gray literature, 2% abstracted sponsorship of individual studies, and none abstracted the disclosure of conflict of interest by the authors of individual studies….

On top of the often wretched methodological quality of individual studies, the compounded effects of data-mining, publication bias, and conflicts of interest, and the notorious problem of drawing causal inferences from epidemiological research, this meta-meta-analysis gives reason to be skeptical of the output of those studies even when filtered and combined by meta-analysis.

Collaborative efforts from investigators and journal editors are needed to improve the quality of systematic reviews.

And, one might add, to improve the quality of the studies that the systematic reviews study.