Once you have made a question one of your favorites, how do you get back to it? I noticed that when you click "Questions" you go a new page with categories "newest," "faq," "votes, "active," and "unanswered." Shouldn't there be a "favorites" category too? And, wouldn't in be cool to be able to...
Well, that is... odd. Are you on Windows? I would not be surprised if this was partly down to bad hinting on Microsoft's part. Some of their fonts are just atrocious.
What I want very badly is a rasterized version of every Courier letter (in a resolution I want) with the same amount on space like it appears in the text.
I teach a college class on the basic use of Mathematica. I wonder what folks would think if I steered them towards this site? One possibility is that I just mention the site (which I've actually done in passing). Another, would be that I require them to visit or even post a question.
The clas...
@halirutan I already did that yesterday. I asked Anna. As I understand, it isn't easy, but can be done. The problem is that the behavior is weird, but not unlawful.
so, consuming developer's time is not completely justified
@belisarius it shows -10/+10 there. I have a couple of those and was wondering what they are. Seemingly people sometimes click where the arrow is without looking to see whether they clicked on it already!
@R.M I once did this experiment before in math.SE; the thing that would probably discourage it is that the "ninja edit" would be caught by somebody keeping a close eye on the front page, plus the current feature of updates now showing up in real-time (at least, in browsers that support it)...
@belisarius Maybe someone didn't like your sarcastic tones :) ? I've upvoted one question and one answer where you lost points, the other ones I had upvoted before.
@J.M. The strangest is one to @OleksandrR. He got an unupvote on Aug 20 for an answer written on February, that was never modified. And there are no upvotes around Aug20, so all the votes were locked then
@belisarius the "never modified" is peculiar. As I replied to R.M, a "ninja-edit" should have been easily caught by somebody monitoring the front page...
A few days ago I've read @Artes in chat "talking" about some inexplicable unupvoting activity going on.
Then, it happened today in some of my questions/answers, so I brought the issue into chat and there discovered this case, which I think is the most clear example . It comes from @OleksandrR....
Hmm, I tried it out on math.SE as an experiment when the real-time updating was first implemented... edited post showed up, but disappeared after refreshing.
@halirutan Yeah, bit me the first time MachinePrecision was introduced... :) The second construction used to produce a nice machine precision number in old versions...
@halirutan not sure what you're getting at... arbitrary-precision Exp for a number very close to 1 actually gains precision, whereas machine-precision Exp doesn't: Exp[x] /. x -> {a, b} // InputForm
@OleksandrR. I just asked whether someone is up for a brainteaser.
But yes, the answer is that although the two numbers have the same precision, one expression is evaluated with arbitrary-precision arithmetics which is more exact than the machine arithmetics.
Maybe one day I will post something comparing it with the analytical n th-order expansion of the covariance matrix (for n = {1, 2}, say), and the Monte Carlo result.
@halirutan nor me. But I am an experimentalist, and we try our best to understand our errors. :)
@belisarius if you're not Millikan, it's hard to understand how one could possibly get the right result. I never had any luck, although probably he built his equipment more carefully. IIRC there have been some allegations that he may have used unsound statistical practices and thus got an answer more accurate than was really warranted.
@OleksandrR. That is what I was talking about. You said we try our best to understand our errors.. Milikan went one step further ... he hand-picked the "best" drops :)
@belisarius since he didn't know the answer ahead of time, I would just say that he got lucky! I think it's ironic in a way since if he got an answer that was off by some noticeable amount everyone would have just said, "well, this isn't a very good way to do the experiment, but it was the best he could have done at the time". It's only because he chanced upon the right answer to such unexpected precision that it's remembered.
@belisarius the case of {4.0(1), 4.1(1), 3.9(1), 4.0(1), 5.0(1.7), 6.0(4.0)} might be nearer the real situation though. :) Discarding the last two measurements is not good practice but I would hesitate to call it malpractice as such.
@OleksandrR. That was exactly what I did when we did the experiment at uni (already knowing Milikan's procedure helped in soothing our conscience). The problem is that if you don't do that, the standard deviation is a shame.
This is an interesting question, I think. Not so much because of what it's about, but rather in that the OP got rather better answers than it would seem she deserved by jumping so naively into such a topic.
@belisarius I always liked using the adjective "pregnant" as the prototypical example: you can't really say "somewhat pregnant" or "partially pregnant"...
@belisarius I wasn't really saying that the answers are especially good, but I think that people seriously addressed this aspect at all given the poorly posed question is quite impressive.
Okay, here is a question of style. I want a version of Apply that operates on level 1 for lists and level 0 for non-lists. One may of course write:
apply[t_, p_List] := t @@@ p;
apply[t_, p_] := t @@ p;
Or, more compactly, we could have:
apply[t_, {p__} | p_] := t[p];
However, that generates a message (Pattern::patv), which is just a warning in this case and can safely be Quieted. If we want to avoid the warning altogether, we can try:
apply[t_, {p_, rest___} | p_] := t[p, rest];
But I consider this rather unclear. I personally prefer the message-generating one, but do not like the message. What would you use, or are there alternatives I've not considered yet?
(Incidentally, this cannot be a formal question because NC.)
@OleksandrR. How about: apply[t_, {p__} | q_] := t@Sequence[p, q]? This indirectly uses the puzzling behaviour that you noticed the other day — that a pattern that doesn't match gives nothing
@R.M well, I would use simply t[p, q] on the RHS, but that is indeed another possibility, which I had considered. I found it even less clear than the third version I mentioned, though.
@SjoerdC.deVries well, I downvoted one of them as I thought it showed such a lack of effort, but at the time it had not been answered so I didn't cast my close vote. The answers are heavily upvoted but the questions are very much not so (in relation to the number of views), so if [much] less than 5% of people thought them worthy of an upvote, I dispute how useful they really are.
I teach a college class on the basic use of Mathematica. I wonder what folks would think if I steered them towards this site? One possibility is that I just mention the site (which I've actually done in passing). Another, would be that I require them to visit or even post a question.
The clas...
A few days ago I've read @Artes in chat "talking" about some inexplicable unupvoting activity going on.
Then, it happened today in some of my questions/answers, so I brought the issue into chat and there discovered this case, which I think is the most clear example . It comes from @OleksandrR....
I have following code in Mathematica:
rbar = 0.006236
rt = r_bar
k = 0.95
sigmar = 0.002
betazr = -0.00014
sigmaz = 0.4
pi = 0.99
chi = 0.05
Cbar = -3.7
alpha1[n_] := alpha1[n] = alpha1[n - 1] + alpha2[n - 1]
alpha2[n_] := alpha2[n] = k (alpha2[n - 1])
sigma1sq[n_] :=
sigma1sq[n] = sigma2sq[n...
