« first day (1638 days earlier)      last day (589 days later) » 

glS
12:37 PM
@Rammus @user1936752 sorry for the ping. A quick question: I've been meaning to do something about the tags , , . More specifically, it seems to me that having both and only serves to confuse things. Most topics about one will more likely than not involve the other anyway. What would you think about merging them, or making one synonym of the other?
Or do you think it makes sense to keep them both? An alternative could also be to have both redirect to
I'm asking here because this is a rather small and niche issue, so a meta post seemed a bit much. And I pinged you because you asked/answered a bunch of questions on the topic.
 
 
1 hour later…
1:41 PM
@glS It depends on how you want tags to be used. IMO they could all just go under "entropy" and have it very coarse grained. Otherwise I'd say keep them both. I disagree though that most questions about one will include the other. If you're applying these quantities to problems they are used for very different purposes. In quantum crypto I see the min-entropy everywhere and I never encounter the max-entropy.
 
glS
@Rammus I generally think having more or less fine-grained tags like these can be useful, in that if someone is looking for such relatively niche topics, can get a decent idea of what has been discussed via the tag. However, there is a limit: being too fine grained can achieve the opposite. Here, my understanding was that the two quantities are tightly related, and as far as I can tell there's some general duality between them, as per eg arxiv.org/abs/0907.5238
(albeit admittedly this might only refer to duality for the associated conditional entropies)
but if there's specific applications where only one is used, that can be an argument for leaving both things. But I'm worried about how one would then classify things such as relative entropy, max relative entropy, the various min(max) conditional entropies, etc.
having a tag for each variation certainly seems too much. We currently also have , which imo is too specific (should we then use this tag for any question related to any entropic quantities that can be expressed via the relative entropy... which is almost all of them?).
Going the other direction, the math.SE approach, would be to only have , but that tag would catch plenty of questions that have little to do, also techniques-wise, with discourses centered around min/max/smooth entropic quantities
maybe a broader classification between "standard entropic" quantities, and single-shot entropic quantities could make more sense?
so something like and maybe, and have all the more fine-grained alternatives redirect to one of these
 
2:01 PM
The wider landscape of entropic quantities is a bit of a nightmare to tag. I don't see an elegant solution. I would shy away from single-shot-entropies as a tag because I don't know if it is clear to people which category they are in.
 
glS
@Rammus I can see that people wouldn't immediately think about a name like that, but that can be solved via synonyms. Effectively, one would type, say, , and would end up using (or whatever alternative we'd be using)
ok so then alternative: use and and that's it? Is it fair to say that most discussions about min/max entropies are effectively about (specific applications of) Renyi entropies?
 
Honestly I don't see a good option. I think relative-entropy is worth keeping as a separate tag though because it's not really an entropy.
Min/max entropies are defined as limits of Renyi entropies so yeah they would fit there
 
glS
2:46 PM
@Rammus mh. Maybe let's wait for more questions on the topic to appear before deciding whether to change it then. To get a better idea on the best course of action. Thanks for the help!
 

« first day (1638 days earlier)      last day (589 days later) »