My Cousin

My cousin

You are trapped in an elevator with the person on your lockscreen. Who is it?

Reblog with who you get stuck with~

A friend of mine was complaining about a problem they were having with their statistics class and I offered to help. They had copied and pasted the example code from their lecture and changed some variable names, but after a while we figured out that the csv file their code was using for data storage had a variable whose name had parentheses in it, so R was trying to interpret it as a function. Anyway the punchline is their TA had designed a coding assignment without making sure they could solve it.

Jesus

I spent almost two days debugging and failing to locally reproduce a bug in some CSV getting exported.

When I finally got the prod db dump, it turns out that what looked like a completely different error is in fact and off-by-one shift in some rows, which is because some client inputs contained semicolons, probably for the first time in years. And not sanitizing properly is not even our fault, the relevant code is between 8 and 10 years old, wayyy before we were maintainers.

Today's lesson is: when the client emails you about an error, don't believe their interpretation of what happened.

not to be weird or anything but we could be kissing and dancing together to slow music in my dimly lit kitchen at 1:00am if you even care

Linux thinkpad / math / transfem

these things coincide oddly often (it’s me I’m the coincidence)

Anyone have any recommendations for laptops? My plan is to install Linux on it, and my budget is ~$500

friendly reminder that you can (hopefully) choose at least some of the things that you allow in your life and “because I really don’t want to latch onto that as a special interest/hyperfixation” is a valid reason to not want something in your life.

Approaching the R^infinity question from a vector space PoV, the unit sphere in l_2 with the standard inner product structure is not compact; just surround each “standard basis vector” (ie (1,0,0,…), (0,1,0,0,…) etc. ) with an open ball of radius sqrt(2) / 2. Since the l_2 norm is nice, none of those open balls contains more than one of the “standard basis vactors,” so that is an open cover of an infinite sphere with no finite subcover.

Of course, l_2 is a complete metric space, so if the l_2 topology can be reconciled with R^infinity in a natural way (I dunno about that part, but I wouldn’t be shocked) the R^infinity sphere in *that* topology has a non-compact closed subset, and is thus not compact. (The open cover with no finite subcover being that one we found for the l_2 sphere together with the (necessarily open) complement of the l_2 sphere.

Thank you for adding to my infinity post, and for doing it kindly! My mutual asked me for a fun fact about infinity, so I do the same to you!

One of my favourite things involving infinity is the construction of some interesting topological spaces!

You can construct the Long Line using the first uncountable cardinal, which is an example of a sequentially compact space that isn't compact!

Sometimes it's also possible to make "infinite dimensional" versions of some common topological spaces! This can be done if there is an inductive way of defining those spaces. For example, we can construct Sⁿ by gluing two copies of Dⁿ to Sⁿ along their boundaries. For S² this is exactly the construction of taking the equator (S¹) and gluing each hemisphere to the equator. Because this is inductive, we can define S^∞ to be the space we get at "the end" of the process. The formal way of doing this is with a categorical limit!

One fun fact about S^∞ is that it is contractible, that is we can deform it into a point (formally, S^∞ is homotopy equivalent to the one point space). But none of the finite spheres are!