Understandably, given how differently the camera works, Lytro has tried to avoid using too many conventional photographic terms, both because their effects are somewhat different and because it's primarily trying to target a mass audience who might be put off by such complexities. However, although it captures different information about the light it sees it is, of course, still constrained by the physics of light and optics. So, although the LFC's behavior may seem rather alien at first, it is still working within the same boundaries that you encounter with your current camera.
To make the camera's behavior easier to explain, we're going to coin the term 'refocusability range.' This is comparable to depth-of-field in conventional photography, but instead of defining the range that is acceptably in-focus, it defines the depth over which the output files can be refocused with acceptable sharpness.
In a small-sensor conventional camera, you tend to get depth-of-field that stretches from near the camera, out to infinity. Similarly, the LFC in Everyday mode refocuses its lens as you zoom, to give you refocusability across the same range. The clear distinction to be made is that, whereas on a conventional compact you're capturing everything within depth-of-field at acceptable sharpness, here you're capturing it with the potential to render it with acceptable sharpness.
Creative mode is more like focusing a conventional compact really close to the lens - you get only a shallow depth-of-field. Or, in this case, a shallow refocusability range, concentrated around where you focused. But, unlike Everyday mode, the camera has to physically refocus its lens to achieve this - not something it is terribly quick at doing.
However, regardless of which mode you use, the ability to separate the foreground and background is heavily dependent on how close the subject is. And, because this is still a small sensor camera, your subject needs to be very close to the camera for the end result to exhibit a significant variation in focus in the output file.
What's in the files?
Much of the LFC experience is mediated via Lytro's website - at present there's nowhere else that can interpret the files or present them.
As a result of someone hacking the version of the 'light field' that is sent up to the Lytro site, it has become known that the camera analyses the depth information in each image and the desktop software renders a series of JPEGs representing the key depths in the image. This shouldn't be taken as evidence that the LFC is conducting some kind of Photoshop-esque sleight-of-hand to achieve its effects (demos of parallax-shifting processing disprove that), but does give some idea of the capabilities and limitations of the files that get uploaded.
The files downloaded to your computer (Mac OS 10.6.6 and newer, with Windows support due 'during 2012'), are 16MB raw 'light fields'', containing all the data grabbed by the sensor, along with a metadata sidecar file, which includes details of what depths the JPEG slices should be rendered from.
The advantage of always working from the source data is that when the processing becomes more sophisticated (such as offering the promised all-in-focus processing option or the parallax-shift view) those benefits can be applied to every one of your existing images. The company also promises more advanced features such as the ability to select which 'slices' of the light field are available in the uploaded version.
The downside is, of course, that you are currently dependent on Lytro not just to host, but to provide any form of interactive online access to your own images. The trustworthiness of Lytro - which we've no reason to doubt - and its long-term business viability aside, this presents some real concerns for early adopters. Lytro insists that it does not claim ownership of your pictures that you upload - 'they are yours' - but reserves the right to refuse to host or remove your images if it believe they conflict with the company's terms and conditions.
With Lytro's proprietary technology this goes beyond issues of content distribution to the heart of image ownership. Most users reasonably expect that if they buy a camera, they can control the images they shoot, and the question of whether they will be willing to cede exclusive control over the product's defining feature to a single company remains an open question.
In our earlier interview, Lytro talked about developing an API to allow third-parties to exploit the capabilities of the camera's files, so it'll be interesting to see where that leads.