[Scythels]

If you have, few, or simple physical constraints then excellent design is quite possible.  I recently went to a presentation on the design of a 16x zoom lens - 16 - 250mm.  The physical length was constricted to < 250mm, max diameter was 70mm, and the distance from the rear element to the sensor was constricted to > 20mm.  Designs were achieved ranging from f/3.8-f/6 to f/6 constant to f/5-f/6.3 and nine of them were diffraction limited.  In an optical sense, diffraction limited isn't so impressive at f/6, but for a 16x zoom lens it is excellent.  The aberration that drove design was also astigmatism - chromatic aberrations were no trouble at all.

And this is with just 12-15 elements, no aspherical elements, and no use of CAF2 (fluorite). 

 

When you start constraining the design physically you run into heaps of trouble achiving a very good result.  Often it is best to start from scratch at first order, but some may stubbornly insist on trying to make a starting point at an old design fit the task.  This can force you into a bad local maximum in terms of performance, or result in a waste of lenses.  Often, adding more elements doesn't help and you could have achieved the same performance with the original element count.

[Klaus]

While I can't really argue regarding the designs, I have strong doubts that the manufacturing process can follow at the moment.

Some of the Tamron super zoom lenses have very decent results actually but it always felt as if the alignment of the elements are more of a limiting factor than the design itself. While the quality and the wide end is often fine (QC-wise), the evenness of the results (across the frame) tends to fall apart beyond 135mm here. Honestly I have a hard time to believe that such lenses are resilient enough to survive medium use over a few years (but then who cares anyway).

[Scythels]

Well, most of the photographic superzooms are nowhere near diffraction limited.  They can't be becase their pricepoint doesn't really acommodate extensive R&D (the superzoom designs I saw a presentation for took half a year collaborating with one of the best lens designers out there) to get to that performance level, and at the end of the day cheaper usually is better for them.  If you doubled the price to $1,000+ they would be able to accomidate the mechanical complexity to make a very solid performer all around.  In general the profit margin can handle a bit of a squeeze in favor of better quality at a lower price without dropping anything re: QC from canon, nikon, zeiss, leica, schneider, etc.  The production cost for a 70-200/2.8 L II is in the ballpark of $400 for the bulk of the lenses, $100 for the fluorite elements (add a bit for the ones that break in manufacturing), possibly $150-$200 (not my area of specialty) for assembly and mechanical parts... basically it comes in way way below the asking price even with a margin for distributor and manufacture. 

 

Not that much at all is wrong with the 70-200L II, it's got nearly diffraction limited performance everywhere and is great mechanically as well, but it could be sold for $1500 comfortably after a few more years to finish recouperating R&D.