Celluloid is given pretty good treatment in
Wikipedia where you will note the Camphor ingredient (up to 30%). Probably the basis for the smell test re celluloid. Fountain pens are even mentioned briefly.
Cellulose Nitrate is explained, and Cellulose Acetate, the more modern less flammable version is also mentioned.
Now to mention heat sources...
I'd like to avoid judgemental terms that usually get mentioned and try to put what I believe into a more neutral perspective. So, please consider the following with an open mind - at least until I finish...
Open Flame heating isn't wrong - it was practised by all professional restorers in the past - and is described (and illustrated) in many of the earlier repair manuals by Sheaffer, Parker etc..., not just by Dubeil.
It has its risks (as do the other methods that follow), but probably the most serious drawback is the tiime it takes to become proficient to the point where one doesn't damage pens on a regular basis. As Ron has said, knowledge of the materials one is working with is essential, and a long apprentiship is normally required to become proficient with open flame, but proficiency can be attained. I also note that these repair persons didn't have heat guns as an alternative.
But if you're good at it - there's nothing wrong in using it.
Heat gun heating isn't wrong - it is practised by most professional restorers currently. It takes less time to master than open flame, and there is a lower probability of an accident. Knowledge of the tool, technique and materials being worked on is still important to success though, and adverse reactions are still possible. Selection of an appropriate tool will minimize the risk, and a lot less experience is necessary if the right tools are used ( infinitely variable heat guns vs embossing tools vs paint strippers).
Immersion in hot water isn't wrong. The temperature is pretty easy to control, although the plastic bag probably introduces local temperature differences unless care is taken to avoid folds, air pockets etc. The risk of fire is practically zero (if not actually zero), temperature measurement may be easier than with the hot air method, and certainly easier than with open flame, but some materials can be damaged by water, so again, knowledge of the technique and materials is important. My guess is that this technique could be much less risk for a beginner, but experience and knowledge is still necessary.
Another technique that I am looking into is an adaptation of the one used by opticians in adjusting plastic eyeglass frames. A small 'Crock Pot' filled with tiny glass beads can be heated to a pretty stable temperature and held accurately (if electronic control is used - not too bad if manual - due to the thermal inertia of the medium), and there is no possibility of damage by water. Ideally, I'd like to use the glass beads that they use, as these don't have sharp edges that might mar the plastic finish, but I might also consider fine sand, or other similar material. Temperature measurement would be dead simple, and the system pretty suitable for the bench. Not sure if there are issues with the beads and nib/feed openings, but I think I have a solution for that too. Enhancements could include electronic control of the temperature. I would suspect that this method represents the shortest learning time and the least risk to the pen regardless of material. One disadvantage is the length of time needed to adjust the temperature if more heat is needed - but if quick response is required, see 'open flame'...

. The heat gun is quicker too - but keeping track of the temp is still an issue.
I trust I haven't offended any particular camp, but I wanted to depart from the right / wrong approach, and to suggest that it really a matter of choices. I wouldn't want to challenge anyone, but I wonder if it were a contest between a 1930's craftsman working with an open flame, an advanced hobbiest today working with a heat gun and thermometer, or a beginning hobbiest working with a cup of hot water, which repair would be completed earlier, and with the least likelihood of failure.
Regards,
Gerry