A self cleaning sensor is never necessary but is always nice.
At the end of this post I've described Checking for Sensor Dust. This is not wholly intuitive and this method may be of value to others.
Dust will "pump" through many zoom lenses as they are zoomed in and out. internal volume increases or decreases and air is drawn in and expelled. Some are worse than others at how much dust this sucks in.
I'll add here a note because of a comment by @nwcs who said that zoom lenses do pump dust, but not into the mirror box. I agree that that would seem to be logical - it is not apparent on inspection that there is a clear air path between the zoom body and past the rear element into the mirror box. However, I have an equally logical experience based argument that suggests they do. As with anything observed, you can create a possible explanation after the event, so I've described what I see and what I think may happen at the end under 'Lens sourced dust?:' at the end. Others can decide for themselves and comment is welcome.
I have a Sony SAL18250 18-250mm zoom and when it is fitted the camera seems to acquire sensor dirt much more readily than when a prior Sigma 18-200mm was used. I find that sensor dirt becomes visible within a month of reasonably heavy use.
Professional lenses with environmental sealing should be much better and a prime should also be better — focusing of a prime lens can cause some pumping, but much less than for a "superzoom" zooming.
I have heard it suggested that a mechanism will generate substantial debris of its own. I have no good feel for the truth of this, but I know that my SAL18250 pumps in dirt much more readily than my older Sigma 18200. I have various primes and lower zoom range lenses but, as the 18-250 lives on the camera usually, its effects hide what the others may or may not do.
I previously has a Sony A700 (killed by "hunting" dolphins) and now have a Sony A77 — the latter with a fixed mirror. The A77 fixed mirror has a degree of airspace around it but far less air should get by than with the 'waving fan' mirror of an SLR. I had hoped that the sensor may thus be less dust affected. It seems that the particles are as bad or worse as with the A700 BUT that the tiny hairs and similar of the A700 are completely absent. This may point to the dust being internal mechanism generated. TBD.
It occurred to me that the dust MAY be on the part-silvered unmoving mirror. A quick test of this showed one large spot that moved when the mirror was "blown" but the small particles remained unchanged. More playing to do there.
I do my own sensor cleaning but I am not going to recommend that others do. While it seems a safe enough activity when done with a "correct" mix of proper swab, proper cleaning fluid, care, skill and luck, I won't guarantee how much of that is luck or what your result may be. There is much on the web on this and some pages are very encouraging. I use locally sourced "pure" Isopropyl Alcohol and this is recommended by some reputable commentators as acceptable BUT your results may vary and that is NOT a recommendation.
My experience with the 18-250 zoom is that sensor dust will generally not show below about f/16. At f/22 it is getting rather obvious and at f/32 it shows horridly. This will vary with the lens involved.
Checking for Sensor Dust:
I check for sensor dust as follows:
View frames and swap to-and-from between them.
Zooming in will probably be needed.
Pan across image looking for anything that is not blissfully blue (if sky was used. Dirt will stand out well.
When a candidate is seen, swap to-and-from between the two images.
Sensor dirt will be present identically on both images.
Anything that changes between images is from another source.
With true even illumination there should be nothing else that can do this but the inter-frame comparison checks this.
(Dirt on the lens face will be extremely defocused and will not show as a point image.
The above works very well for finding sensor dirt.
If you do your own cleaning, clean and repeat test.
I find that most dirt can be removed in one cycle but that often some may move and relocate in a whole new pattern.
I clean across and up (or down) so that dirt that is left SHOULD all be in one corner. Often works ;-).
The test without cleaning does no harm at all and tells you if cleaning is needed.
Reducing Sensor Dust Effect — Emergency Action
If you are taking photos and notice that sensor dust is visible and are unable to address it immediately,, increasing aperture to larger than f/16 (more light) will make the effect much less obvious or even completely nonobvious. Not an ideal solution but a useful one.
Just did a quick test. Sal18250 at 250 mm.
Kitchen wall — far from monochrome but 4 seconds of waving produced bright grey.
Exposures around 4s.
Adjust ISO and exposure level to keep times around 4 seconds - very roughly.
f/38 - horrid BUT none of these would show in most photos
f/32 - most still visible but much softened.
f/11 - one dot visible and very soft. The one dot is about 1.5% of sensor height = about 60 pixels !!! or 0.3mm - quite a monster. That one removed with a blower! ;-).
Lens sourced dust?:
I swapped lenses to confirm that dust seen is not lens based.
A77 dust does not seem to be mirror sourced.
Examination of he rear element of a typical lens, and of my Sony SAL18250 in particular, do not seem to show a clear air path. The seal around the lens-body interface is not made to be air tight but is such that you'd expect minimal air flow during normal use.
When changing lenses I am reasonably careful re dust entry but far less so than some people. I tend to orient the body orifice down if possible or horizontal as second choice.
I had a Sigma 18-200mm which I used on Minolta 5D & 7D.
I bought a Sony A700 with Sony SAL18250.
I have used the 18-200 occasionally with the A700.
I bought a Sony A77 with a new Sony SAL18250 (insurance replacement or I would have bought a D700 :-) ).
The Sigma 18-200 used to acquire sensor dirt at a moderate rate. This was both small particles of varying size and shape plus some iems more like hairs or short arcs in appearance. So:
5D / 7D + 18200
A700 + 18250
A77 + new 18250
The A700 + 18250 and used to acquire dirt at a noticeably higher rate than the 18-200 + 5D/7D - maybe 2+ times as fast, with more hairs and arcs noticeable.
The A77 + new 18250 acquires small particles at a substantial rate (needs a clean after 4 months of moderately heavy use but all the dirt is particle shaped - there are no hairs or arcs. this is optically very distinctively different from the A700 + 18250
The A77 has a fixed mirror that has noticeable air path around its edges.
The difference between 18200 and 18250 is pronounced. There is also a change of body but as the new lens involves more hairs and arcs these seem unlikely to originate in the mechanism. It is a possibility.
The difference between the A700 and A77 (same lens type different lenses) is also pronounced, with dirt changing to all particle and no arcs or hairs.
I do not recall checking A77 sensor dirt from new - a sad omission.
Inspection of early photos MAY help.
It is possible that the lack of waving mirror and probable reduction in air movement helps.
It is possible that the mirror blocking almost all of the orifice helps.
It SEEMS that the A700 dirt source was at least partially external.
It SEEMS that the 18250 "causes" more dirt than the 18-200.
It SEEMS that the A77 produces zero arcs and hairs compared to the A700.
One could propose an explanation for almost total internal dust generation plus perhaps admission at lens charge time but Occams Razor suggests that air pressure changes in the lens are conveyed across the rear element "seal" and also that there is a dust path present. This is conceivable and needs more testing.
Having had this issue raised in this detail I will keep a closer watch on it. It may be that the first A77 cleaning will remove dirt from post manufacture (which one would hope was minimal) and that there will be less in future. TBD.