Re: [Jprogramming] locale error in Vocabulary/Locales

On Sat, May 21, 2022 at 1:30 AM Elijah Stone wrote: > On Fri, 20 May 2022, Raul Miller wrote: > >> Was that from intel? > > > > I guess (after reading below) that that depends on what you mean by "from". > > I want to know what _you_ mean by it. I meant that as a general rule, I see it being used

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, 20 May 2022, Raul Miller wrote: Was that from intel? I guess (after reading below) that that depends on what you mean by "from". I want to know what _you_ mean by it. there's going to be code which the compiler does not treat well. Like what? Repeatedly changing the nameclass of

Re: [Jprogramming] J904-beta-d

Can you try start J with an empty profile? eg jconsole.exe -jprofile jqt -jprofile On Sat, May 21, 2022, 5:05 AM 'Michael Day' via Programming < programm...@jsoftware.com> wrote: > Thanks. The upgrade was easy, but the result is unfortunate, at > least on this laptop running Windows 11. > >

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, May 20, 2022 at 6:02 PM Elijah Stone wrote: > Yes--that was one of the representations I was thinking of. But it makes > random access hard. One alternative is a pair of parallel vectors which can > be fed to i.. This has the advantage that, like with adjacency matrices, it > is easy to

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, 20 May 2022, Raul Miller wrote: a space efficient representation of a relatively sparse directed graph would be a two column matrix, with one column representing connections and the other column representing node ids. Yes--that was one of the representations I was thinking of. But it

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, May 20, 2022 at 5:08 PM Elijah Stone wrote: > On Fri, 20 May 2022, Elijah Stone wrote: > > I still don't know a more space-efficient representation for graphs than an > > adjacency matrix. > > Sorry--I know of lots. What I don't know is a time-efficient representation > which is more spac

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, May 20, 2022 at 5:04 PM Elijah Stone wrote: > Was that from intel? I guess (after reading below) that that depends on what you mean by "from". > My impression--pieced together piecemeal from various sources--is that > 'structure of arrays' was a design pattern and meme that arose in the

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, 20 May 2022, Elijah Stone wrote: rearranging axes such that common access patterns exhibit greater spatial locality wrt ravel order (See FFT for a case of this which applies to arrays which are long in multiple dimensions.)

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, 20 May 2022, Elijah Stone wrote: I still don't know a more space-efficient representation for graphs than an adjacency matrix. Sorry--I know of lots. What I don't know is a time-efficient representation which is more space-efficient than adjacency matrices. -

Re: [Jprogramming] J904-beta-d

Thanks.  The upgrade was easy,  but the result is unfortunate,  at least on this laptop running Windows 11. Attempting to open any of Jqt, JHS, Jconsole does establish a new instance of jqt.exe or jconsole.exe, according to the details display of Task Manager,  but it immediately shows its sta

Re: [Jprogramming] locale error in Vocabulary/Locales

Was that from intel? My impression--pieced together piecemeal from various sources--is that 'structure of arrays' was a design pattern and meme that arose in the late 90s to early 2000s in the world of game development, and which was itself ultimately a derivative of the earlier columnar store

Re: [Jprogramming] locale error in Vocabulary/Locales

I'll have to try it again. It was mostly for curiosity. On Fri, May 20, 2022 at 2:27 PM Henry Rich wrote: > A couple of releases ago I rewrote numbered-locale processing. I don't > know of any reason why it should be slow to delete them, unless you have an > awful lot of symbols in the locales.

Re: [Jprogramming] locale error in Vocabulary/Locales

A couple of releases ago I rewrote numbered-locale processing. I don't know of any reason why it should be slow to delete them, unless you have an awful lot of symbols in the locales. Henry Rich On Fri, May 20, 2022, 4:20 PM Devon McCormick wrote: > I wonder what use one would have for these d

Re: [Jprogramming] locale error in Vocabulary/Locales

I wonder what use one would have for these doubly-linked locales. To me, it has kind of a bad code smell but that may be because I'm unfamiliar with the sort of thing for which it's intended. On Fri, May 20, 2022 at 2:54 PM Don Guinn wrote: > Yes, I find the numbered locales quite useful, but b

[Jprogramming] New in beta-d: multithreaded primitives

If you create threads, beta-d can use them to speed up primitives. The first primitive to get this treatment is +/ . * (designed & coded by Elijah Stone). Say your machine has N cores. Set your startup script to execute 0 T. '' N-1 times, and your matrix products will be faster with no other a

Re: [Jprogramming] locale error in Vocabulary/Locales

Yes, I find the numbered locales quite useful, but building the double-linked list of locales might mean he might intended to have very many locales created and could be a performance problem to delete the locales. On Fri, May 20, 2022 at 11:46 AM Raul Miller wrote: > On Fri, May 20, 2022 at 1:2

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, May 20, 2022 at 1:29 PM Don Guinn wrote: > Just a thought. The management of numbered locales is not very efficient. A > while back I created several thousand locales as an object test. I went to > delete all of them and it took forever. Right. My understanding was that J's locales were

Re: [Jprogramming] locale error in Vocabulary/Locales

Just a thought. The management of numbered locales is not very efficient. A while back I created several thousand locales as an object test. I went to delete all of them and it took forever. On Fri, May 20, 2022 at 11:19 AM Raul Miller wrote: > On Fri, May 20, 2022 at 10:12 AM Robert Cyr wrote:

Re: [Jprogramming] locale error in Vocabulary/Locales

On Fri, May 20, 2022 at 10:12 AM Robert Cyr wrote: > I should add 2 small points here: > > 1.The one line correction. > My intent was to keep things simple, also retaining the focus on the ease > of using numbered licales. And it worked for me on my windows pc and my > Ipad. That is why I did no

Re: [Jprogramming] locale error in Vocabulary/Locales

Hi Robert, if you have update permission on the wiki, you don't need to request anyone's permission to change a page. If you change or add something substantial, you may want to let people know after the fact but that's up to you. Cheers, Devon On Fri, May 20, 2022 at 10:12 AM Robert Cyr wrot

[Jprogramming] J904-beta-d

J904-beta-d is available for windows/linux/macos. If you already run J904-beta, then upgrade is easy: load'pacman' 'upgrade'jpkg'jengine' If you have not yet installed J904-beta, time to get started! https://code.jsoftware.com/wiki/System/Installation/J904

Re: [Jprogramming] locale error in Vocabulary/Locales

I should add 2 small points here: 1.The one line correction. My intent was to keep things simple, also retaining the focus on the ease of using numbered licales. And it worked for me on my windows pc and my Ipad. That is why I did not add more code or warnings about its limitations But I like th