« first day (1028 days earlier)      last day (2414 days later) » 

2:22 AM
I hate when a new user deletes his first question just when I got a good answer
 
They probably figured it out.
 
Just need to post somewhere :(
n = 5;
ListPlot[{u, v} /.
List@ToRules@
Reduce[ManhattanDistance[{1, 0}, {u, v}] <= n &&
Mod[ManhattanDistance[{1, 0}, {u, v}], 2] == Mod[n, 2], {u, v},
Integers], PlotStyle -> PointSize@Large]
I feel better now
 
2:37 AM
Reduce can do some pretty unexpected things ... pretty sure it's just brute forcing it, but still ...
 
@Szabolcs Well, integer programming ain't easy
 
@belisarius I want to use my own settings for some package function by default. Can you think of a way to do this without having to load the package in a special way? I don't think there's a robust solution for that.
Now I have a utility package that loads the main package then changes some things about it.
 
@Szabolcs I tried something like that on V8 and was heavily defeated. No, I only got convoluted code with leaks. More code added, more leaks. Sorry.
 
I'll just stick with the utility package.
 
@Szabolcs Seems a good decision to me.
 
2:43 AM
Was there any news about the 2014 tech conference here?
Did any of the SE regulars attend?
 
@Szabolcs Yup, there were some posts of Murta and PatoCriollo at least
@Szabolcs One of them is the last starred entry on my screen
 
Ah, so no news on v11. That's actually good. Want more bugfixes for 10. Many more.
 
@Szabolcs At least the essential ones. Yes
 
There's plenty of essentials ... I don't even tough some parts of the system these days because I don't have that extra day's time to figure out why it's not doing what it's supposed do.
Wanted to write some quick unit tests the other day but gave up because my last experience with v10's MUnit was awful (and didn't want to rely on Workbench 2 for the tests)
 
@Szabolcs Well, I decided to postpone migration completely. I get payed for testing software (usually)
3
 
2:53 AM
They're going to have a demonstration about v10's new features at my university tomorrow.
 
@Szabolcs They risk their lives going to a user community. But I think that the amount of bugs will not be better in next releases. Mma is now a very complex piece of software, and you know ...
 
I wanted to go and complain a bit, but I think it's no use. The job of the people who come is just to sell Mma, and any feedback through them probably won't reach the people who could do something about the problems.
 
3:10 AM
@Szabolcs Yeah, it's pointless.
 
hhh
3:31 AM
How can I do this?

{{a, i, 1}, {a, j, 2}, {c, k, 3}} ---> {{ai, 1}, {aj, 2}, {ck, 3}}?

I am importing a CSV file with header where unique identifier is specified by two lines.
Is the concatenation of values the best way to proceed (I think it is bad way, not big thing)?
 
hhh
3:52 AM
Come on, there must be some very easy CSV import thing for this kind of headers. I found this Q [here](http://mathematica.stackexchange.com/questions/3890/stream-csv-files) -- I hope someone picks up this problem because I want to get it solved to play with this data in Mathematica.

Perhaps the right question should focus on streams, CSV -- I don't know yet. Trying to understand why no activism in my question, worrying :(
 
 
2 hours later…
6:20 AM
@hhh are the a, i, etc. strings or variables? I would think strings, since they're coming from a CSV file, but you didn't quote them
 
hhh
Good question, I meant them to be strings, sorry about confusion.
But Look up at this, I got it nicely to Mathematica -- despite I cheated a little bit with AWK :P
.
 
Ah, well I was going to say Map[{#[[1]] <> #[[2]], #[[3 ;;]]} &, list]. There should be a nicer way to do it with pattern substitution though, anyway.
 
hhh
I cannot now remember how to choose only specific columns for plottings such as TIME, YEARS, K3_PRODGOPR?
 
That's a pretty simple task for Mathematica (the thing you originally asked)
 
hhh
?
@DavidZ I am always curious how to do things in Mathematica :D
3rd. ListPlot worked to plot everything but I should be very selective in plotting: plot only specific columns with ListPlot?
@DavidZ moved your comment to main here: I come back to it later, now want to get the plot nicely (3rd)
.
4th. Is there any way to get data by column name? data={{"colA","colB","colC"},{1,2,3},{4,5,6}}
someCommand(data,"colA") should return {1,4}.
5th. somePlotCommand(data,{"colB","colC"}) should plot only colB and colC.
(I can do this easily in *ix tools but I would like to learn how in Mathematica, hallo -- this must be easy to some guru here? Sad that I am always alone here, perhaps most sleeping -- hopefully someone notices the puzzles)
 
 
1 hour later…
7:50 AM
@kirma They posted about it on Facebook and Twitter a few days ago. They were factual and measured in their reporting of it :)
 
 
5 hours later…
12:37 PM
@Pickett Ah, those media unsuitable for proper content representation. (... which I try to avoid following.)
 
 
1 hour later…
hhh
1:41 PM
@Öskå I commented for you but you removed your comment, the Chris outputs different results to the intended results: mathematica.stackexchange.com/a/65378/2000
 
 
4 hours later…
5:54 PM
@Szabolcs it was mentioned during the conference (October 22nd) that version 10.0.2 would be released sometime in the coming weeks to address more bug fixes.
 
6:35 PM
@PatoCriollo Sounds great! I'm not expecting every important problem to be fixed by 10.0.2 but I hope it will be followed by 10.0.3, 10.0.4, etc., regular bugfixes until 11 comes out.
 
7:00 PM
I sense a great dash of optimism in that expectation.
 
 
2 hours later…
8:53 PM
Bug fixes would be fine, but improved speed and reduced memory usage would be welcome too.
 
 
2 hours later…
10:45 PM
@SjoerdC.deVries this was also mentioned :)
I think they followed more or less the same strategy used for the image functions
Lets first implement the functionality, and then improve the implementation
And yes, i have the feeling that there will be a 10.0.3, .4, etc
And that some of these, or at least the next 10.0.2 wil actualy bring a little more than just bug fixes
And for those sceptical on the way WR is going with all the new functionality, i did get the impression that they are hard at work on the speed matter, that is, that they are not trying to compensate anything...
I did not get any feeling in respect to the extent of this work. Will it be so fundamental that they will be able start programming more in WL than in C? I have no idea.
But i did got the idea that we will soon be presented with fundamental improvements. Since i've been using mathematica for some time now, i know that their time scale is probably not the same as we would like, and that "soon" doesn't typically mean what we expect... But we never know!
And i can add that the new functionalities that don't make a lot of sense for the old user of mathematica, actually make a lot of sense for the product line, since wr is more and more known as the wolframalpha company, and not the mathematica producers company (i'm not saying in terms of money, since i have no idea, but in terms of image). And so, i makes a lot of sense all this wolfram here and wolfram there... Would they have named wolframalpha a abcalpha, and i think that we would be
currently programming on the abc language ( this is me saying. I have no official information on this. It is just the feeling i got from the different discussions.)
(sorry for all the mistakes... The ipad is not the ideal interface for typing...)
And since money tends to have a positive feedback on product development, i welcome anything that might bring more users... If they arecoming from the realms of W|A, i'm ok with it !
 
11:14 PM
@P.Fonseca You were at the WTC, weren't you?
 
Yes ;)
But one thing i felt that was not clear in their mind... The industrial market.
 
@P.Fonseca That's how I feel too.
I have to work with data sets of several millions of records, up to say 500 million and things like Dataset just don't cut it then.
And calculating DateDifference on such a data set will take ages
The former lets my memory explode, the latter my patience
 
i thought that with mathcore acquisition they would be brought into this world, but i still don't see it... I thought that we would be able to use the wolfram engine like a headless player, with signed libraries, etc. and, although no one told me differently, i get the feeling that this is not what they have in mind...
@SjoerdC.deVries I got the feeling that, on the memory side, they are working on it.
 
Dataset was designed with obscure hierarchical datastructures in mind whereas most people simply use tabular data
 
... And i would consider the date difference problem more of a bug than a bad implementation
 
11:24 PM
@P.Fonseca well they forced the unnecessary use of units upon us here
It was slow in 8/9 and it got even slower in 10
 
On the dataset, true. But i got the feeling that they haven't finished it...
 
@P.Fonseca Indeed, I know
 
Yep... They always had a problem with dates... I remember that it was faster to transform it to absolute time, make the operations on it, and transform it back again, than to work on the "native" form...
 
An off-memory/on-disk implementation of Dataset was in the documentation earlier but didn't make the cut
but I hope that they come up with something to make Association more efficient for tabular data. Like implicit keys or so.
 
It was presented on the european conference 2013, and it is still discussed here and there. That why i say that .3 and .4 might bring up more than we are used to...
 
11:28 PM
By that I mean that you specify the keys once and don't repeat the for following lines of data.
 
I got it...
 
@P.Fonseca With 'It' you meant...
 
Implicit keys
 
two lines ago I mean
OK
chat is too asynchronous
 
Agree. Specially on tne ipad, where the screen keeps jumping around while we write:(
 
11:32 PM
I have that sometimes too. Never found out what caused that.
 
This things are really not for work... But agin, if one can make manoy out of it, seeling apps that get "used" after two hours, like if a cinema ticket it was, then, i'm all for it. I just can't figure out why i'm not in the same wagon...
Lack of vision i guess...
Not manoy, but money :)
I'm still waiting for a tabular interface. They keep saying they are working on it, and keep but this has already lasted for a couple of releases so...
@murta hi, award winning fellow
 

« first day (1028 days earlier)      last day (2414 days later) »