The AIP is not a law, and I think this is nonsense because there are loads of airports on which a failed departure would land you straight into houses.
One tends to not list them openly in case a low grade journo reads the forum and writes some garbage in the local rag saying that a plane could plummet into [randomly select one of: school / church / convent] 
Is it AIP-UK we are talking about here?
I once departed with a SEP in 600 m visibility in fog from an uncontrolled airport. My major concern was not EFATO but staying on the runway and knowing the distance to the runway end.
I agree and will depart in fog if legal and I can track the runway and have confidence no one else is on it. A controlled airport using LVP helps. But it is a subject that tends to be quite divisive.
To some degree at least isnt departing in fog a little like flying at night – if the donkey quits there is a large element of luck involved?
I might put the question a different way – whether on departure or coming through an overcast what clearance between the cloud and the surface do you need to have some hope of adjusting the trajectory to improve your odds? In other words I suspect even over very flat ground if you become visual at say 400 feet there is very little time to do anything but land straight ahead. At 1,000 feet there is actually a reasonable amount of manouevering room left. Some where between the two your options rapidy disappear.
As I said earlier in some ways on departure you are better off than a descent through a low undercast – at least you have some idea of the terrain and when you will become visual if you hvae just gone into cloud. The challenge is gettng the nose down quickly. You also pretty much know what the wind is doing. From above you probably have a far poorer idea of where you will break out, the winds aloft may not give you such a good idea of the surface wind and you might have very little idea of the terrain. Warnings about taking off with low cloud bases should be supported with equal warnings of flying on top of low undercasts which seem to me more dangerous.
Yes, but then most of us are prepared to fly at night….
I know – and in that case bases on departure are reasonably irrelevant if that is your risk assessment.
In the US, there are no takeoff minimums for non certificated part 91 operators. Zero Zero takeoffs are permitted, regardless if they are not particularly bright. They make for good practice under the hood, but I would never consider it in real life. My personal minimums are at least approach minimums for my airport, but they would be useless in the case of an engine failure. It is all about risk and risk management. Same for flying at night, over water and rugged terrain.
In the US, there are no takeoff minimums for non certificated part 91 operators.
That was also the case for an N-reg departure in the UK, till maybe 2-3 years ago.
I would never take off unless I could see the runway to be clear (of vehicles etc) for a sufficient distance, which in practice usually means the whole runway length.
I did do a zero-zero takeoff once, with a then well-known cowboy instructor (who eventually vanished from aviation in a hurry, following some adventures with female students ) just by keeping the heading bug lined up. Evidently, it worked. But it could so easily not have done (had I been solo) because even a 1 deg error is way too much.
In other words I suspect even over very flat ground if you become visual at say 400 feet there is very little time to do anything but land straight ahead.
I think, flying over most of Europe, and unless you are having a really bad day, a 400ft cloudbase will get you to a field.
One can implement plausible risk reductions. Over mountains, if overcast, I fly with a moving map GPS running a topo map, and that has to be vastly better than nothing (nothing = russian roulette). A forest is bad news, as would be a large city. But one spends a tiny % of one’s time over large cities, and same for forests (outside Germany).
Fog aside, the cloudbase almost never touches the ground. I think there is some process which separates the temp and the dewpoint near the ground. Probably it is radiation heating of the cloudbase, because I can’t see the dewpoint being any different.
bases on departure are reasonably irrelevant if that is your risk assessment.
I think it has to be thus if you are flying a single engine plane.
But again one can reduce the risk by not flying much at night. In fact I usually find flying at night very inconvenient because, upon arrival, all one has time for is hunt around for a not totally crappy hotel… It’s handy for flying back home though.
I can’t agree with RobertL that a twin needs 800’. A reasonably capable twin (and by that I really mean any of the serious touring twins – Baron, C3xx, C4xx, Aztec, PA31, later PA34s etc – as opposed to the little trainers) can deliver a NTOFP which is not far off the normal performance in some common SEPs. Yes, you need to be in practice, but if you are, 500fpm is normally quite achievable, especially on the kind of days which produce fog.
I also can’t agree that it’s difficult to maintain the runway centreline in low vis. If you fly commercially and the minima say 175m for takeoff then you wouldn’t keep your job very long if you refused to takeoff in those conditions. I would not use heading, but focus on the centreline. It is then useful to have someone to observe the gauges and dials, of course.
Obviously LVPs and a runway inspection are good ideas, but in their absence you can always backtrack the entire length yourself to check for obstacles and sink-holes.
Low visibility operations are not as safe as VMC, of course, but it’s a question of personal risk management profile.
OP wanted individual opinion on what individual limits we apply, not whether we all agreed! This is a free country.
In any event a typical factored SE rate of climb close to sea level, is between 200 fpm to 400 fpm – with average blue line speeds of around 90 -100 knots, that is only 2 – 4% gradient, once you have cleaned up and feathered.
I tend to like to apply Class B performance criteria when flying 40 year old piston twins, but in truth would not plan to fly low IFR privately these days; this is because Class A passenger seats are available for chump change.
