The documentation is not misleading at all. Maybe if you only read one single paragraph and ignore the rest you will get the wrong idea, but the fact is that most people seem to get it.

And you are saying that TFA is like comparing Access to Oracle, but why are you assuming that everyone knows everything you do about this? A comparison like this between Access and Oracle would not imply that they are equal at all. In fact, the test results speak for themselves.

I'm sure you would like to go on and on about this, but you didn't actually address my comment. Did you understand what I wrote about Chris Messina's misguided blog post?

FWIW I completely agree that Unite and servers like Apache are completely different beasts. But that doesn't mean that it isn't interesting to see comparisons and benchmarks. And nothing beats Unite's ease of use for most people. You won't be hosting the next Facebook, but you will have control over your own data. Several people I know have converted to Opera already based on this first preview of Unite. They love it. It's more to do with the ease of use than that they don't want other people to host their files for the most part, though.

As I said in another post:

I'm surprised to see that people are still linking to this. It's basically full of errors, and was written in rage over all the hype Unite was getting. He was angry about how people just repeated Opera's claims blindly. Kind of like you are blindly referring to his blog post even though it turns out that the post is too inaccurate to really be used for anything.

You really should read some of the comments on the page you are linking to, in order to see people correcting all the misconceptions. For example the misconception that everything goes through a proxy, as you claim it does. Furthermore Chris's comments where fun until Haavard took him down a notch on his own blog, resulting in Chris himself posting on Haavard's blog with a massively different tone.

A more thoughtful take on the subject can be found here:

I'm surprised to see that people are still linking to this. It's basically full of errors, and was written in rage over all the hype Unite was getting. He was angry about how people just repeated Opera's claims blindly. Kind of like you are blindly referring to his blog post even though it turns out that the post is too inaccurate to really be used for anything.

You really should read some of the comments on the page you are linking to, in order to see people correcting all the misconceptions. For example the misconception that everything goes through a proxy, as you claim it does. Furthermore Chris's comments where fun until Haavard took him down a notch on his own blog, resulting in Chris himself posting on Haavard's blog with a massively different tone.

You are comparing apples and oranges. The 10 MB version of Opera has 30 languages included. Compare the basic English version, and the story is quite different.

I don't know about memory usage. In my experience, Opera does better than any other browser there.

What is "just a web server"? He tests both dynamic and static data. PHP and MySQL was just for comparison purposes, to see how the Unite server compares to other web servers.

And yet it's still smaller than the so-called "pure browsers" ;)

If I am not mistaken the OLPC project had to double the amount of RAM to get Firefox to run properly on the system. It looks like Firefox could indeed benefit from losing a bit of excess fat.

As for this story, I doubt that bookmarks, history and those kinds of features are the ones making Firefox run slow. It probably has more to do with the architecture itself.