Do we have a canonical answer for "Why doesn't a=b; f[b_]:=a; f[2]" output 2?

@xzczd Pattern is HoldFirst

@b3m2a1 It's a=b, not b=a…

@xzczd Oh then it's just because of the injection. b_ doesn't bind 2 to b, it injects 2 by lexical replacement.

I'm asking this because I think it's not trivial, and if we can find a canonical answer for it, we can mark this post as duplicate: mathematica.stackexchange.com/questions/172903/…

@xzczd this seems like it might be that: mathematica.stackexchange.com/questions/17767/…

Leonid answered it and to my mind for Mathematica programming that's as close to canonical as we get

@b3m2a1 Hmm… not that close, I should say. Let's wait for a while to see if someone can find a closer one.

Love these Michael Trott posts. Don't understand a substantial part of the content, and have just a small understanding of the motivation, but they really show Mathematica potential and distinction from other environments.

Another good one they did that was quite interesting was the Hidden Figures post: blog.wolfram.com/2017/02/24/…

@Edmund bought the ticket! (just 15 € below 3 months ago cost, but had a feeling that it was just about to go up... as other companies options were starting to disappear)

Very interestingly, while choosing the seats, half of the airplane on the 4 flights was already filled... 5 months in advance!!

community.wolfram.com/groups/-/m/t/1333342

I'm going to drop this here again: If anyone is doing graph stuff with Mathematica, any testing that you could do for IGraph/M would be most appreciated. There is no doubt there are bugs there. Just try to break stuff and let me know about problems.

MailSystemMap as a built-in function? The refusal to break Mathematica into modules is ridiculous beyond belief. There are so many advantages to modularity (and so many disadvantages to the lack of it) that I just can't understand why they do this.

For example, if the design of Graph turns out to be just unsustainable (I'm specifically thinking of the lack of support for properties in multigraphs) then they could just introduce a new Graph module, and phase the old one out gradually. If the two functionality groups lived in different contexts, then they could be used concurrently when necessary.

With the current state of things, there could never be another System`Graph ever again. If there's a new one, it would have to be NewGraph or Graph2 or similar, and the same for all the new function names that come with it.

There is only one set of packages you should use! And in case they will need to introduce conflicting frameworks, it is a good opportunity to have WolframGraph <3

There are just too many warning signs. The more I think about it, the more I'm convinced that the days of WL are numbered. It's accelerating off a cliff.

M has very nice things, but that parts that are broken are really broken, and they are fundamental to the system.

@Szabolcs I agree 100% with everything but your future :)

Please, How to computing the exact value of Exp[-1000]?

@Betatron it is exact, you are up to this I guess: Exp[-1000.], read about N and related docs.

@Szabolcs, it's interesting to read that. I've been thinking similar things recently - in particular that I've invested a lot of time and learning into M, but it's gradually becoming less and less relevant in the technical computing world (or at least my part of it).

So I've been making a conscious effort to NOT use M all the time, to try and get competent in a couple of other languages so that I'm not left behind.

I like M and I expect there will always be a place for it in my toolkit, but realistically it's not going to be my main computational workhorse a couple of years from now.

@Szabolcs This all day. So much bloat. It's ridiculous. And they want to demodularize the paclet system...

@SimonWoods I think that most of us is thinking more or less the same thing... which is a pity. What is scaring is the speed at which such a community feeling spreads. I really think that we are at the moment of all or nothing... I mean, it is no longer M OR another tool, but instead, how will M integrate into the other tools market, to stay relevant, since the wrong integration will mean JUST other tools...

@P.Fonseca Good to hear. I just submitted a presentation for a talk. Hopefully it gets accepted.

Wolfram changed their product presentation to a knowledge based language marketing philosophy. This seems logical, as it is, to my understanding, the only one selling this argument. But I think that the argument is far from being enough to justify its position in the market, while competition has already covered for way more than the 10% that the 90% needs to do their/our job (and in a nice and continuously growing way).

@Edmund Great! Complicated stuff, or just funny stuff? I’m also submitting something, but I haven’t figured out what yet (have two months to do so :-)

@SimonWoods and this is why I like Michael Trott type of blog posts, because they kind of show the type of production where Mathematica shines, helping to keep alive it’s reason and my motivation

@P.Fonseca Not complicated stuff. Just an application in an industry not often presented

Thinking planning on submitting a second one as well to hedge my bet. That will be more on the lines of citizen science