Yes, flourscent is better. However, special rough service incandescent bulbs for things like shop lights are available.
Would that be semolina or durham? All my bulbs smell like glass. 
Just because that’s a cool name, or is that some sort of pun? If it’s a pun, someone explain it?
One thing I’ve noticed, is the 7 small bulb hanging light over my diningroom table connected to a dimmer switch is often burning out bulbs whilst on. Every month or so there is a small popping sound, and whenever it gets down to three opperational bulbs I go and replace the blown ones.
It’s a semi-obscure reference. “Gravity’s Rainbow”, Pynchon. There’s a very strange little diversion in there relating the story of “Byron the Bulb”, an immortal light bulb. You can actually leaf through a copy, and read it on its own - it’s a funny bit, in a very surrealistic way.
That’s pretty much it. What’s happening to weaken that filament is a process called thermionic emission. What we think of as a solid wire is actually a collection of tungsten atoms stuck together. When you heat it up (as when you put current through it, and resistive heating takes place) the tungsten atoms start to move and jostle. Eventually, by chance, one will gain enough energy to allow it to pop off the surface. It drifts through the mostly-vacuum bulb and eventually strikes the glass envelope. Since this is relatively coolerm, the tungsten atom gives off its excess heat and sticks there. This is why burned-out bulbs have a dark cast on the glass (all that bubbled-off tungsten plating out) and thin portions on the filament (the hottest parts of the filament, where the tungsten has bubbled off the most. It’s a feedback thing, too, since a section of filament that has more tungsten misssing, so its local resistance goes up, and it gets hotter still, boiling off even more tungsten.) It also helps explain why bulbs generally fail when you switch them on. As you point out, that initial surge is what does it in, and if the tungsten has thinned to narrow neck, then the surge will preferentially heat up that spot, and it’ll go right away.
Of course, if you keep the bulb on always, then it will fail when that weakened spot fails, even if the bulb is on. But if you turn it on and off it’s almost guaranteed to die one of those times it’s turned on.
You can extend the life of a filament by putting halogen gas in the bulb (instead of vacuum). The tungsten essentially “bounces off” the halogen gas and is returned to the filament, instead of plating onto the bulb wall. This can slow down the deterioration, but not stop it. (the tungsten atom isn’t going to go back to the same place it’s boiled off from. In fact, it’s more likely to go somewhere else, where it’s cooler. So you still have a deteriorating filament.) Moreover, tungsten-halogen bulbs have to burn “hot” to be efficient, and that can cause problems. At lot of fires have been started by something coming in contact with a hot tungsten-halogen bulb in, say, a torchiere.
The electrical resistance of the filament when cold is pretty low. However, the resistance increases with temperature. A slightly higher applied volatage when the filament is cold will cause a burn-out because of the much higher current that it will abe to push thro. That is one reason why most burn-outs most often occur early in the mornings when the network voltage is most likely to be on the higher side because of low general consumption. Voltage variations while the bulb is ‘on’ also cause surges of currents and may cause burn outs but since the filament resistance is pretty high at the time, they are less likely to fail. The other reasons have been covered in earlier posts.
I doubt that lightbulb has been on since 1901, since I can’t believe that they’ve never had a power outage in all that time. especially during WWII.