Hi,
first, thanks to nickf08 for the evo 1.06 question... i've been asking this here and there with no answer for a while
now, just my two cents in this interesting discussion
i've been using evo for a while, have made quite a lot of website with it and will continue to do so, even if evo were due to be left behind by the team (as i said once, i'm in the starting blocks to make my own php5 oo fork of evo with a mootools 1.4 upgrade of the manager if this happened, Susan ?

)
in the meantime, i install and try every single Revo release and, so far, even if some new concepts are really interesting, for many reasons, i'm sorry to say that Evo is still the winner for me...
- performance, i can't get enough of my clients being so stuck by how fast their websites are, when, on my local server, i'm stuck by how slow a simple page with no css, no js but just a content var is rendered in revo (Bob is probably right, this will sure change in a near future)
to be honest, this is particularly true with multilingual websites, but, i compare Evo + Yams with Revo + Babel as being French, most the websites i make are multilingual
- extending, well, i mustn't have read bob's book enough probably

as, personally, i feel installing modules in evo (every single one being php5 oo written) so much simpler than the manager custom pages in revo
- manager, well, here, revo is the clear winner (except for speed but this is improving fast), design and simplicity are really great in revo
- js, well kind of an personal allergy to ext

dojo or mootools (i do love oo javascript...) would have done the trick but well, personal and irrelevant then

and so on... (without speaking of xPDO, i love the pdo part but, god, why xml, why not json or, better, a php config object like in Yii, well, yes, pPDO sounds weird i admit

?)
this being said, i will keep trying revo releases until one makes me think "here we are, it's better than evo", this may happen if Revo had the few features i miss in evo, build in multilanguage support, versionning and workflow
till then i can't see why i would have slower websites when i still haven't hit the wall that says "i can't do this in evo but i can in revo" and, believe me, i've done pretty complicated things in evo with complex user management, software interfacing and so on, the only times i feel evo is not the right solution for me are when a frameworks seems a better one, not another cms, that's to say when it's more about web application than website (event if i also have a evo made website with framework build application aside... this works too and interface with evo like a breeze)
have swing