is mathematical interval notation inherently clash...
# randomh
Hullaballoonatic
11/14/2018, 8:42 PMis mathematical interval notation inherently clashy with typical code syntax?
[1,10)1 until 10d
dalexander
11/14/2018, 8:47 PMI think mismatched braces/brackets/parentheses/etc. are unlikely to be well received even if there’s a basis in the mathematics community for some of them.
h
Hullaballoonatic
11/14/2018, 8:49 PMyeah, that's my worry, though i do think there could be a good demand for number types restricted by range, e.g.
var x: Double by (-3, 4]😱 1
r
Ruckus
11/14/2018, 8:52 PM@Hullaballoonatic Scala much? 🧌
Ruckus
11/14/2018, 8:54 PMIn all seriousness, Range requirements aren't really something you want in the type definition in general purpose languages.
h
Hullaballoonatic
11/14/2018, 8:54 PMthere's little reason to think that a feature existent in a language you don't prefer couldn't be an effective addition to one that you do. refusing to ask such questions out of principle seems silly.
unless i misjudge your intent. in which case' i'm sorry
Hullaballoonatic
11/14/2018, 8:55 PM"Range requirements aren't really something you want in the type definition in general purpose languages"
i'm intrigued to know why that is?
r
Ruckus
11/14/2018, 8:56 PMI agree completely with your point in the generic sense, I was talking specifically about this feature.
h
Hullaballoonatic
11/14/2018, 8:56 PMgotcha!
r
Ruckus
11/14/2018, 8:57 PMRange checking is just a form of data validation, and that can be so nuanced and use case specific (or general and vague), it shouldn't be part of the type system, beyond the fundamental limitations of the type (as computers, unlike math, are finite machines).
h
Hullaballoonatic
11/14/2018, 8:58 PMi can envision that point pretty easily: e.g. if a value is set outside the delegated range, does it default to the closest allowed value, or simply throw an exception?
Hullaballoonatic
11/14/2018, 8:59 PMi appreciate sharing of your thoughts
r
Ruckus
11/14/2018, 9:01 PMExactly. I've caught myself in a similar trap on many occasions, where I think "I wish it was smart enough to do..." or "I wish there was a notation for..." and after more consideration realize there's very good reason it shouldn't (and funnily enough it almost always has to do with some form of data validation).
h
Hullaballoonatic
11/14/2018, 9:04 PMis the problem too varied to solve with a single word like
restricteda
aarjav
11/14/2018, 9:22 PMNot quite what you're looking for but if you're just interested in the way it looks, you can use something like this plugin: https://plugins.jetbrains.com/plugin/9320-advanced-java-folding The code syntax stays the same but the ide can sugar it
g
gildor
11/15/2018, 12:54 AMRange in your example even not a part of type system (you cannot understand range knowing only type), rather just forced runtime validation
Because of this, such validation already can be achieved using classes, just write RangedDouble and use any semantics which you want (crash or use min/max for values outside of type)
t
thomasnield
11/15/2018, 3:18 PMMaybe programming should have been invented before math. Notation would be so different.
🤔 4
s
Shawn
11/22/2018, 2:39 PMwe don’t have a think emoji strong enough for that one
2️⃣ 2
❌ 1
🤔 2
3 Views