PocketMatrix

Microsoft has a new web browser in beta called Deepfish. It's quite revolutionary in that it tries to be like Opera. So by "revolutionary" I'm using Microsoft's definition of the term.

It's a restricted beta that just opened today, and already all the download slots are full.

I have a hunch that it uses server-side processing to reformat web pages (scale images, etc). That is the only real reason I can think to limit the downloads so severely - to keep load off the intermediary "proxy" servers.

http://labs.live.com/Deepfish/

Yep, after reading the details, that's exactly what it does:
"client application that leverages a powerful server side technology" Blasted marketdroid-speak. I hate it when the term leverage is used as a verb. You'll only see that in marketing or "synergy" type scenarios.

Anyway, I don't know that I would trust MS to handle all my web traffic through their servers.

Dan East

Well, after looking at the flash demo, this technology is really limited, and for most types of browsing this will be completely useless. Basically all it does is have the server render the entire webpage as an image. Initially it is scaled to a small enough size to fit the PPC / Phone display. You can then pick an area to zoom in on. At first the image is pixelated, because it simply zooms in on the low-res image it already has. It will then retrieve a higher res image from the server.

This is really stupid for a few reasons. In order to read text you will have to zoom in substantially - probably until there is a 1:1 pixel ratio with normal desktop displays. Thus you will be looking at such a small text area that you will have to scroll back and forth constantly to read text as it flows from line to line.

Next, in almost all cases, this would have to consume more bandwidth than downloading the actual page and images. If all you do is look at the thumbnail representation, then you will save bandwidth. But as soon as you zoom in, such large images will be retrieved that any gains will be lost. This is especially true with text pages. Imagine reading a news article. The text would only comprise a few kilobytes worth of data. To download it as an image would require hundreds of times more bandwidth.

Dan East

Would it not make more sense to generate some sort of vector based layout* and send that, along with the text, to the device? You'd lose fine image details but the text and page layout would match with a drastically reduced download size.

* Sorry, throwing out buzz words here...