Detect Whether Ans is a String or a Number

[bfr]

In Ans is a string or a number (real or complex). Make a program that puts 0 or 1 in Ans depending on whether initially Ans was a string or a number. Note that this should work for any string or number.

I'm going to keep the size of mine hidden for now.
Mine is 30 bytes (after subtracting the length of the title from the memory menu size). ;P

[Weregoose]

19 bytes say any string and only zero will produce 1.

[bfr]

Yeah, I have a version that works like yours (with 0 being treated as a string) that is 18 bytes (19 bytes without subtracting the length of the program name).

I guess I might as well give the size for my robust version that works for "any string or number," which is 30 bytes (you can confirm that I haven't improved it since yesterday by highlighting the line after the "I'm going to keep..." line in my original post, which was last edited yesterday, if you really care ).

[Galandros]

I have tried to solve this and haven't reach far...

But in the process I provoked some odd errors and curious behaviours...

[DarkerLine]

I've gotten a version that works for any string or number in 23 bytes (returning 0 for a string and 1 for a number).

[bfr]

After reading your reply, I have a version that is 23 bytes large and returns 1 for a number and 0 for a string. However, strictly speaking, it doesn't work for any number or string...although exception is really insignificant. I'll PM you or something about the exception.

[Builderboy2005]

I am completely at a loss as to how to achieve this. I was able to get a basic technique that only worked for non-decimal non-zero non-negative numbers, but that was a very hacky way and used over 300 bytes. It would seem you can only use operators that can take either a string or a umber as arguments? The only ones i have found are the + operator, and the = and != operators. Am i on the right track? If not, could somebody bump me in the right direction?

[bfr]

You're on the right track. You can highlight the following facts (maybe just the first would be enough) that give hints about the way I made my versions. You pretty much just have to know these in order to make this routine the way I did.

Useful fact: You can store a string to a list variable (normally for attaching a formula to a list).

Another useful fact: You can store to the list _LA simply by storing to the A character alone. For example: {0,1}->A .

Also, the (small) exception I was talking about in my 23-byte version, and possibly in DarkerLine's 23-byte version too: The fact that the + operator can be applied to strings and numbers is useful. However, an error is returned when one attempts to use the + operator with the "empty string" (which can be created by entering a quote on the home screen and then pressing enter). It's kind of a really insignificant and cheap exception though. =P I did not use the + operator in by 30-byte version that works for any number and any string (well, it works for any number and string as far as I know) .

[Builderboy2005]

Ahhhh, I knew both of those two facts, but failed to put them together! That is unimaginably clever!

[ztrumpet]

This is unimaginably clever! I can't figure this out at all. Those who have fiured it out are amazing. Great job!

After an hour of screaming at my calculator I've only come up with an 8 byte thing of code that works for every input except for the number zero...

[calcdude84se]

Until that was explained, I, also, was at a total loss.

[bfr]

Here's my 30-byte solution that works for any input:

[Weregoose]

My smaller program was one byte larger than your smaller program for one reason: it's hard to use not( on a complex number.

[bfr]

Ah, I didn't know about that...thanks for informing me. I didn't really test my 18-byte and 23-byte ones anyway, though, because they already had at least one exception.

[ztrumpet]

Very nice brf! Insorak (on Omnimaga) had that small solution also. Great job!

[IAmACalculator]

There's also the lame 68k version:
getType(ans(1))

[Galandros]