Ok, some classic FUD that needs debunking.
(a) Is the only real issue here. If you're shooting at f/2.8 or otherwise shooting for scenes where bokeh will need to be considered then you have to be mindful of the potential impact of onion bokeh.
(b) Just about every lens has sharpness falloff towards the edge, including a lot of Canon's "L" glass. If you look at any MTF graph, you'll see this and similarly it will show up in pictures. In this area, Canon's wide angle "L" zooms are atrocious.
(c) If IS is pointless then why does Canon have it on the 24-105, 17-85, etc? For some people that have trouble holding the camera/lens still, IS is a life saver. Lots of people wanted Canon to come out with an IS version of their 24-70, so it is safe to assume that lots of people see benefit in it being there despite it not "slowing the scene down".
(d) nobody has yet come forward with measurements of AF speed, it's all "feeling" based and we're talking differences in 10ths of a second or less. The only people for whom this will be a problem will be the Internet photography experts that shoot from the comfort of their chair at home.
(e) It has yet to be established that there is any significant variation in lens quality from Tamron regarding this lens. Regardless, there is plenty of talk about of people receiving "soft" Canon lenses too. In both instances it is likely a problem solved with AFMA.
No FUD because as I wrote I'll probably buy the Tamron myself for being a lot of "bang for the buck". But I'm sorry to have to correct you...
a) I disagree: the "onion" boekh is *not* an issue because it needs specific conditions to appear, so for many scenes you'll never see it.
Do you see the onion bokeh in any of the sample pictures posted by Tamron for the lens? No.
b) Just because there are other Canon lenses that show a sharpness falloff, it doesn't mean it can be taken out of consideration - because the alternatives to the Tamron 24-70 are Canon's new 24-70ii or primes, but of course both at a higher price.
Bob Atkin's review here:http://www.bobatkins.com/photography/reviews/tamron_24-70_f28_VC_review.html
Seems to disagree with your assertions about sharpness falloff towards the edges being an issue.
And more to the point he says: These problems are fairly easy to correct in post exposure processing. Corner sharpness isn't as good as center sharpness, especially wide open, but again that's pretty much the case for all fast zoom lenses in this range.
It looks like Bob Atkins has a good handle on photography and some real experience with various lenses in this category.
c) The 24-105 has IS because it only has f4 max aperture, and the 17-85 and such are consumer lenses that wouldn't sell w/o IS and are dual-used for shooting still scenes, too. For me, the 24-70 *f2.8* is an event/wedding/docu lens, and since you need about 1/100s min shutter speed for shooting people anyway IS is not needed, esp. if it needs a little time to swing in after af'ing like the Tamron.
On cameras such as the 5D Mark II and those with even more megapixels, the advice is to shoot at 1/2*zoom for sharp results when without IS. So if you've got a 24-70, without IS you would be best advised to shoot at 1/125 or faster in order to get sharp results. What IS does is bring that back a stop or two, allowing you to get sharp results at 1/60 or slower. Of course that doesn't make up for subject movement, but it isn't only the subject that needs to be still, it is the image capture platform as well.
And as Bob Atkins points out:Is it as good as the new Canon 24-70/2.8L II USM? I don't know since I have not yet had the chance to evaluate the Canon lens. I would not be surprised if the Canon lens was somewhat sharper wide open in the corners, but that's just a guess at this point. However even if it is sharper you have to ask the question whether any additional sharpness would be worth an extra $1000 and whether or not some of any additional sharpness might be lost in hand held situations in low light due to the lack of optical stabilization.
d) The Tamron is slower to af esp. at lower light, where even a little makes a difference if you use it for wedding/event w/o preflash or af assist - the photozone review says so, see the "nikonguy" side-by-side comparison, and I tried the lens myself next to a Canon mk1. While the dfference might not be large, it may make a difference if you shoot hundreds of pictures in an hour with it like during a wedding or event.
None of the reviews flag it as being a significant issue.
To reinforce that, you yourself don't know if it will or won't be an issue, you're simply speculating.
And yes, I watched the "nikonguy" review ages ago and came to the conclusion that the issue is negligible as it never impacted any of the photographs that they took.
e) I'm getting my information from Amazon et al customer reviews this thread: http://photography-on-the.net/forum/showthread.php?t=1145772&page=36
Let me quote one back at you...http://www.thephoblographer.com/2012/06/20/review-tamron-24-70mm-f2-8-sp-vc-canon-ef-mount/#more-24771
Quote: This lens, however, has to have been the best I’ve ever tested
It should not be left unsaid that Tamron will calibrate any lens for free inside its warranty period
according to people posting in that thread you linked to above. To me this indicates that even if you do get a lens that isn't right, Tamron will fix it for you for free. What's not to like about that?
If this lens had the name Canon on it and had a different colour scheme, everyone would be singing its praises.