Wiki/Report of Meeting 2023-10-12

Report of Meeting 2023-09-21

Present: Ed Gottsman, Raul Miller, and Bob Therriault

Full transcripts of this meeting are now available on the following this synopsis.

1) Bob reports that he did not do much on the wiki while on vacation, but is heartened to see the contributions that Cameron Chandoke has been making, particularly in the are of Modifier trains https://code.jsoftware.com/wiki/Talk:Vocabulary/ModifierTrains.

2) Ed has been traveling as well, although he has been looking for a Linux box for testing. He is wondering if WSL2 is an acceptable testing venue for Linux on a Windows box. There was further investigations about ways to create User ID's to run J in a Linux sandbox. Raul feels that there are a lot of challenges in testing a UI application. This lead to questions to whether the J Wiki Viewer should be released for J9.5 as there are challenges J9.5's version of JQt with regard to isigraph. This may present issues with developing the viewer for new versions of JQt going forward. Bob mentioned that for Jig he uses an umbrella script to load different scripts based on the Users J version as there are challenges with tooltips with different versions of JQt.

3) Bob mentioned that he used the J Wiki Viewer very effectively to do research on Jose Mario Quintana's wicked toolkit https://www.jsoftware.com/pipermail/programming/2017-October/049263.html for the podcast episode on Modifier Trains. There have been other reports of users expressing surprise at the depth of the Wiki through using the J Wiki Viewer.

4)A deeper discussion followed on whether the J Wiki Viewer should be supported for the J9.5 beta. Ed was concerned that users might be discouraged with their first impression if it was J9.4. The application is solid in J9.4, but J9.5 is shaky. Bob feels that as a beta it should be expected that there will be bugs and that information for beta users should be that if they are using the viewer in a production environment then the J9.4 version would be preferred. 5)Ed has been doing updates to the wiki database every week instead of daily and wondered if this would be an issue as he is nearing the limits of free use of AWS. Bob felt that he is more concerned with historical information and uses the forums and the APL farm for breaking news, but that users should be informed about how current the information is used. Raul felt that a message on the UI regarding the last update and next update would be appropriate. Bob wondered if Jsoftware might be willing to pick up a minimal cost for AWS if the value of the J wiki view is worthwhile. Ed says that the primary benefit of the application is that its development keeps him sane.

For access to previous meeting reports https://code.jsoftware.com/wiki/Wiki_Development

If you would like to participate in the development of the J wiki please contact us on the general forum and we will get you an invitation to the next J wiki meeting held on Thursdays at 23:00 (UTC) Next meeting is October 19, 2023.

Transcript

- Let's get started.

- Okay.

So I was hoping to do more this week, but I've done almost nothing 'cause I've been working on other things.

But I'm heartened to see the amount of work that people are doing on the wiki, maybe not in the structural parts that I've been working on, but in terms of the information.

And I think, Raul, you've been part of it with Cameron Chandhok is another one.

Cameron's been putting some serious thought into some of these things.

- It's some really good stuff in there, especially around the modifiers and that stuff.

In fact, the podcast we just recorded on Tuesday, we were talking about Tasset and a big part of it was the stuff about the invisible modifiers and those kinds of things.

'Cause a lot of people think they're neat, but don't know what they're for.

And a lot of people think they're just insane, But he's doing a lot of work towards, and you've done a lot of work towards clarifying how they work.

Yeah.

So that's very nice to see.

I'm encouraged by seeing people come in and work in the Wiki.

I've got to do more on the structural end of things.

But I know that.

And that's probably what my next week is, because I'm an off week from the podcast week.

And I should be able to get some chance to do it.

So that's my report.

Well, I didn't do much in the last week because we were traveling, but I was online shopping for Linux boxes for testing purposes.

And I ran across something that I had learned about a few years ago, never paid any attention to and forgotten about, which is the Windows subsystem for Linux, WSL.

So it turns out Microsoft at some point decided there was some value to open source and put a Windows Linux subsystem into Windows.

So it's now up to WSL2.

It's apparently pretty good.

I installed it.

It does work.

You can get Ubuntu running, not just sort of next to Windows, but the Ubuntu apps run as Windows, Windows.

They're on the Alt tab gallery.

The task manager, it's very strange.

I don't know what Linux Window Manager it's using by default if it's using one.

>> It's using Windows as the Window Manager.

>> That was my suspicion, which makes me nervous.

This is not really a good test bed for what I'm trying to do here.

And that was the burden of my question for this evening.

And depending on the answer, I may have a couple of follow-up questions roll for you.

Is this a reasonable way of testing the JViewer under Linux?

Or is it sort of two windows adjacent to be a good test?

I think it's a good test.

I mean, as you've been thinking, there are different window managers.

So any test is going to have holes in it because it's going to be other things, other environments.

So you got to, at some point, you got to rely on other Linux users supplying some of their own debugging effort, and you got to support that to some degree.

But it's good for a start.

And I haven't, you know, from the Linux side of things, the big deal is that, you know, it's a hosted environment.

So you don't have all the processes there that you might otherwise expect to, but that's not a huge issue for me.

Especially not for Jay.

Maybe if I was building something that was a system component, I'd wanna have something different.

I don't think it should be, other than available fonts and system installation, dependencies and stuff like that, where you might wind up with different versions of something.

Personally, I'm running Debian under WSL2, instead of Ubuntu.

Network's fine.

Oh, alright.

So my next question is, I went ahead and installed J manually in user local, and I get directory permissions errors when I launch it, although it does ultimately-- not errors, warnings-- Although it does ultimately launch.

So where should I be installing Jay on?

On in in in in UWL partition or in you talking about?

I just went with the defaults.

I think I might have it stored in my my home account.

I I'd have to see the errors to begin to understand what the errors are about or the warnings that that there were about.

- So when you say you installed it in your home directory.

- Yeah, let me just double check that.

If I start up.

Just wanted to, I hadn't been running WSALS.

I'm gonna start.

Nope, I don't have it in my home.

Yes, I do.

Well, I have several instances of J installed on both Windows and under WSL, and I'm sharing my home directory between the two as well, which makes it a fair bit strange.

Let's see if I can even remember where I have.

And I've We've also got source instances of J that are built under WSL.

Right.

So I think that's probably my… You've got quite the menagerie going.

Yeah, I have a non-typical system.

And now that I think about it, I've got a JQT instance under a GitHub directory, and I have a JConsole instance under a JSoftware directory.

And those are my primary Linux instances of J.

I don't know if I've ever actually installed an official Linux version on this system.

But I don't get a warning either, so there's that.

[laughter] What did the warning say?

- Well, maybe this is my problem.

By default, it looks like WSL set me up as root.

So when I start Ubuntu, I'm root.

- Yeah, I would-- - I'm being kicked by my prompt.

Who am I?

- I have a, let's see.

I know how to administer Debian.

I know Ubuntu is derived from Debian, but I don't know all of the things Ubuntu does, but I have it set up in WSL so that I am user UID 1000, whose name is user when I get into WSL.

- Isn't that interesting?

The warning message that I get is around user ID 1000.

- Oh, really?

- Yeah.

- So if you type ID at the command line, what does it say?

UID equals zero root, GID, group ID, I guess that is, equals zero root, groups equals zero root.

- Okay, so yeah, you could, you could, you could ask you to, to, what, what, what's the user ID for 1000?

- The user ID for 1000?

Well, I can get the error message.

- If you say grep 1000/etc/passwd, what does that give you?

- Rep 1000, what was the rest of that please?

- Splash etc/passwd and spaces on both side of 1000.

- Yeah, I get nothing.

- Okay, so you don't have a 1000 user.

- Doesn't look like it.

- You can probably create one.

- All right, so that we won't go into the details on that because I can figure that out.

So the suggestion would be, and I was easing my way around to this anyway, that I not try to install as root, that I create a user ID and install in a user-- - Well, if you install in user local, it doesn't matter that you've installed as root, but you shouldn't, and generally in Linux, you often use root for installs, but-- - Oh, you use sudo.

- Yeah, typically you use sudo to do an install, but if you run it as, I mean, installs are an administrator actions and root is the administrator account.

So that's not where the worry is.

The thing is that when you're not doing administrative action, you want to be a different user so that you don't accidentally do administrative things.

Right.

Now I understand.

OK, so I'm still left with installing J as root makes sense.

Putting it in user local makes sense.

Actually, let me try to run JQT here and tell you what the warning message is that I get.

Yeah, if it's about the user ID, then creating a user account and running there will fix it.

And I know it's a different system, but on my Mac, I've got an administrator's account that I'm not in most of the time.

I install everything.

I'm actually in my user account and then it asks me, I need to put in administrator permissions to do it, but I do it all from my user account.

Yeah, Mac is all engineered very carefully to make sure that help people's hands for that issue.

All right, so here's, this is actually kind of a meaningful message.

Q standard paths colon runtime directory runs slash run slash user slash zero slash is not owned by UID zero, but a directory permissions Oh 700 owned by UID 1000 group ID 1000.

Okay.

Yeah, that's that's saying your root user and your non root account.

Did I say unplug myself?

Yeah, just for a second there.

Yeah, you're running as root, which means you're admin, and you're depending on a non-root account, and it's warning you about that.

OK.

Why didn't you say it's?

It sounds like the QT library.

Oh, QT library, yeah.

Could be.

This historically has been a problem that Root could get hijacked because if somebody else put something in its path and you relied on it and didn't realize you were relying on it.

How interesting.

Especially in university settings where you have a shared computer system and you can have students pulling pranks on other students and whatever else.

Yeah, yeah.

All right.

All right.

Well, let me-- I'll play with that.

Thank you.

And then, sort of before we left, so a week and a half ago, I was thinking, well, I'll get the Linux testing working, and then I'll go ahead and go for general release.

But I became, as we were away and I wasn't able to do anything, increasingly concerned about the behavior of parts of WD, which kicked in, seemed to kick in with the beta for 9.5.

And I'm now thinking, I mean, I've currently got code, I had to drop is a graph because as a graph was misbehaving in favor of as a draw, which just completely freaked me out.

And then I've now got code that checks the version 9.5 versus 9.4 to decide which maximization option to use when I do a P show in WD to launch the window.

And I'm freaked out by all of this.

And I think what I would like to do is wait until 9.5 releases, get rid of that check and not be trying to support 9.4 and 9.5.

And I want to open up this discussion.

The issue that I'm aware of there is testing a UI system is extremely hard because there's parts of it that are difficult to, parts of the tests that are difficult to automate.

And so people rely on bug reports and being especially careful working on UI stuff.

And in that context, the thing that I think you did one email and it was kind of a bug report and kind of a question, but I think it should be made clearer to, you know, like on the general list that there's this change between versions and a changing behavior between versions and that's probably undesirable for people that actually use the system, actually code up using the system.

I mean, if you're delivering an app using J, you could freeze the J version.

But if you're supporting something using J for other people, then you kind of want to.

.

.

I mean, your approach of freezing the old version out, that would work.

You force people to use the new version if they're going to use your app.

But it's also good, I think, to put a little emphasis on the bug report so that people can be aware that there's something going on that they've done.

>> Yeah, I'm a little reluctant to push on this any further in the general list.

I reported, I had one actual bug report where I included code that isolated bad behavior.

And I got one response, which was your code isolating the bad behavior is itself bad, which was to my mind a misreading, but I didn't push it.

The second attempt, which was a couple of weeks later, was when I had found workarounds for all of the issues.

So, dropping Isograph, checking the version of J before I do Pshow.

And my message was, and it was sincere, was, "Look, I think there are problems here "that somebody should care about.

"I myself have found workarounds for all of them, "but I think you should be aware "that there may be other things.

"These may be indicative of other problems "that have not manifested themselves yet.

" And I don't particularly wanna repeat myself, if you see what I mean.

- One thing to be aware of there is, Issygraph is itself relying on a QT feature, which is problematic.

- Oh, really?

- Yeah.

There's only things that have to happen inside the paint event, and it's relying on a thing called a Qt Painter widget, which has all sorts of runtime constraints that, of course, bubble up into J because J is affecting runtime behavior.

So I'm not surprised by problems with Issygraph, just because of that, the underlying foundation is that quirky.

And Qt has some other quirks too, probably we should think about rebuilding all of Qt to better address some of these.

Or at least the JQT, our implementation on top of Qt, but replacing Qt with a different framework might be a worthy choice too.

It's-- there's some troubling things going on.

So thank you.

>>All right.

Well, I think my takeaway here is tough.

You're just going to have to rely on bug reports from users and do as well as you can.

It sounds like the foundation is, um, creaky.

I think it's probably better.

I mean, QT was very carefully engineered and then they found out that the design was actually not supported by, or they, they couldn't figure out how to, how to implement the design.

And, and I think we're now up to QT6 where JQT was originally done under, under QT4 and most of it's Qt 5.

And the problem with Qt 6 is it hasn't been there's some environments that were j runs more Qt 6 isn't available.

But it's better.

It's a better it's a better engineered system than Qt 5 and 5 is a better engineered system than 4.

Okay, well, I did talk to you about, you know, being freaked out by having to do different versions for different versions of j that that you do our JQT that you're working.

That's exactly what I had to do with Jig is that I've got a version for 807, I've got a version for 901, a version for 903 and a version for 9.4.

And so when I do that as an add-on, when you go to the add-on, it queries what your system is, what version your system is, and it loads up the appropriate add-on.

So the actual add-on file that you're looking at is a distributor to the other files that are sitting in the manifest.

And I'll pull the one that you're working under.

I'm probably not bright enough to be freaked out by that, but that was my way around it, is that I was working with the WebGL, not WebGL, whatever their web viewer was.

And it's changed a couple of times.

and each time there's a specific thing I'm using tooltips and each time that's where it shows up.

So if I don't make a change, the tooltips all show up black.

And so I figured out what the change was and it was a very strange, had to do with non-breaking spaces.

And when I put the appropriate number of non-breaking spaces and suddenly my tooltips worked again.

And in some versions you can do line feeds to go down steps of a tool tip.

And in others you can't, you've got to rely on breaking spaces out to get to a, you know, a wraparound.

So that was my way around it.

But again, I was using JQT as a way to be the stable part of the platform.

- Oh, yes.

- Yeah, yeah.

- And the thing that's driving this, I mean, fundamentally the thing that's driving this is each person on a project or as an employee or whatever has a limited time span.

And then they-- [INTERPOSING VOICES] And that grades all the way down to the hardware level, where different machines are subtly incompatible at the hardware level because of the turnover issues.

But also, it's every level of the stack.

People have to rebuild stuff because somebody's gone and they didn't understand it.

So anyways, it's definitely a disturbing thing, but we've been managing to cope so far, I guess.

- All right.

- And I think the new version will usually land in December or sometimes January.

That's your timing.

- Right.

All right, but it sounds like waiting around for that is not gonna solve my problems.

this is a more fundamental problem than that.

- Yeah, usually what you try and do is encaps, you know, put, come up with the fixes, put a wrapper around them and have a stable interface to your own internal thing and live with the overhead of that.

That's the general approach, although at some point that becomes, you get too many of those and your system starts slowing down and then the whole thing gets replaced.

- Yeah.

All right.

Well, I'll work on getting WSL, the WSL install of J to work properly.

And I'll check and make sure it works and take a checkpoint and see how I feel.

And we can talk again next week, I guess.

Thank you very much for all this discussion.

I very much appreciate it.

And on a very positive thing, getting ready for my Invisible Modifiers episode, having to brush up on my old knowledge of the modifiers and stuff, I realized I wanted to do a search for an email that Jose Mario Quintana had sent out with his wicked.

.

.

Here's what I found.

Oh, thanks, Siri.

I wasn't asking you.

But um.

.

.

Wicked Toolkit?

Yeah, the Wicked Toolkit.

And I used the browser, the search engine to do that.

And it was amazing.

Like, it's just, that's amazing.

Like, I went back, I not only found the one I thought I was looking for, I found others that were more important.

And all I did is did a search on wicked, because that was a term that he used.

And it kept coming up as he would do tacit stuff.

And it was just, well, I guess it's the usual thing.

You find the right search term, but with that, the door just opens.

And I just had to do that scroll back, you know, run back in time because it was about 2017.

And yeah, it was, I just find that so useful.

- I suppose you had a comment that he was surprised by how rich the archive is.

And he didn't give me any details about what in particular he'd found or that had particularly struck him, but he was clearly surprised and pleased by the depth and richness of the archive and found the tool a way to see it, experience it in ways he'd never been able to before.

So I think there is something to that.

Yeah, I'm, well, and I don't know whether, I guess we got that feedback from Ian Clark a little bit, because when he got around the interface, that was one of the things he said, like he made a change for him, and he said, suddenly it became immersive.

Oh, I missed that.

That was one of his emails back to you.

I mean, I'm so glad you're working with him because he's so good at this stuff.

I mean, he and Henry were the people mainly behind NUVOC.

And Henry credits Ian with doing most of the writing because he was blown away by how well Ian could communicate things.

Right.

No, Ian is a particularly good communicator.

It's true.

There's no question about that.

Yeah, yeah.

No, actually, I do remember that now.

Yeah, that's right.

And it was funny because when he said that, I just went, "Oh, there's the nickel dropping," right?

We've all been saying, "You know, this is really cool.

This is really cool.

I don't like the way this works.

I don't like the way this works.

" And you've changed, you tweak something, and you go, "Oh, oh, this is really cool.

" Yeah, so I haven't seen exchanges back and forth between you guys.

you know, for a while, but.

.

.

It's been silent.

Yeah.

Yeah.

He, in his last exchange, he said, "I could write a lot more.

I may at some point.

" And so I've sort of been waiting for that to happen.

I don't want to be clingy, I guess, is how I feel.

Yeah.

Well, I think to some extent, you got him to the point where he's crossed the boundary into the new area and now he can explore.

And so, yeah, do that for a bit.

There is a school of thought that says that once you've made the sale, you shut up.

- Yeah, that's right.

(laughing) Yeah, I've watched salesmen talk themselves out of sales before.

- It's horrible.

Okay, I guess, is that all that I've got?

Yeah, I think that's all I wanted to talk about.

So that's everything I have.

- So going forward, what are you thinking now?

Would you develop it for J9.4, and then when that's locked down, move on to 9.5 and hopefully get to that before they release, or what do you think?

- No, I mean, the problem is that J9.5 is already in the wild.

There are people using it.

- But it's a beta.

- But they're using it.

- Yeah, but it's a beta.

You don't worry about somebody being in a beta and something breaking.

- I think you do.

I mean, the whole point of having the beta out there is so that you can do work on it in advance of its release and get rid of any bugs that there might be.

So that there are beta users out there is an opportunity for you as a developer on top of the system.

- Yeah, but I'm using the beta from time to time too, just for that reason to get used to it and to find things in it.

But I've also got 9.4 in my system and I always can go back to 9.4.

And I do, I mean, there's times and sometimes I'll be working in an area and I'll actually work in that area 9.5 and when I wanna work on something else, I'm back to J9.4.

- I think there are probably, and I myself am one of these people, enough folks who have installed and generally use 9.5 beta right now that their first experience of the add-on would be to try to run it under 9.5 beta.

Oh, cool.

All right, I'll just try this.

Yeah, yeah.

And if it blows up for them, I may lose them.

Yeah.

See, there was one version and I think it was 901 that I was just cruising along with Jig right up until the second to last version of the beta and it blew up then.

And so I had like about three weeks to fix it.

But in that three weeks, I got it all fixed and everything.

So when it actually launched, it was working again.

But if I had worried, if I'd said, Well, and you may find this because they may change something in the beta that affects it towards the end.

Yeah, but I mean, I've got it, I've got it working at least Mac and Windows in the beta now.

Yeah, yeah.

I mean, I can, I can launch with the, I guess I could launch with the information that it works in 9.4 and in 9.5 beta as the beta is currently constituted.

that's not guaranteed always to be true, but I'll do my best.

Yeah, I don't want to freeze out people who have switched over and the beta is quite mature.

I don't want to freeze out people who switched over to the beta, I don't think.

No, I wasn't thinking of freezing them out.

I was just thinking develop for the for 9.4.

So that's solid.

And I have been it is.

Yeah, yeah.

Lack of solidity while dropping is a graph obviously was emotionally difficult.

(laughing) But the P-show thing is having to check versions to determine which P-show option to use.

That's quite strange, but that works now between two versions across Windows and Mac.

So that's okay.

No, I think I wanna try to support, I want to try to stick with 9.5 beta.

And if a problem comes up, fix it.

Rather than say, you must revert to 9.4, you adventurous pioneers.

- Yeah, no, I'm not saying you don't fix it.

I'm saying, if a person says, I've got this problem in 9.5, you can say, you know what?

If you really want to use this, 9.4 is solid.

- Yeah.

- And thanks for letting me know I'm working on, tell me how.

- Yeah, I'm working on it.

- Yeah, I'm working on it, it's beta.

- No, that's absolutely right.

Yes, thank you.

- Yeah.

So I don't think anybody's expecting the beta version to be without bugs 'cause it's a beta.

- Right, right.

- And if you've got something to refer back to, if you say, if you got 9.4, run that and it'll be solid.

Because think about what you're, it's a viewer on this corpus, right?

- Yeah.

- And the corpus isn't changing.

- No, well.

you know, it doesn't change between 9.5 and 9.4.

- Right, the corpus doesn't.

- The corpus doesn't.

So if you need to get that information, it's solid on 9.4, you can go for that.

- Go ahead, I'm sorry.

- I was gonna say, there is an argument that says, 9.4 would be something I would use to know that I can use the JWebViewer on that, or the WikiViewer on that, 'cause I know it's solid.

So there's an example of, I would leave that at 9.4, And I could try it on 9.5, but if it wasn't, I'd always be able to go back to 9.4.

- Right, yeah.

The thing is, this is one of those tools whose point is to be ready to hand.

And if you're working in 9.5, it should be ready to hand in 9.5.

It's not a-- - I have another question for you.

- I'll revert to 9.4 when I wanna look something up, that's no good.

- If 9.5 changes, so that P-SHO is consistent 9.4, would that be a problem for you?

Yes, I'd have to go get rid of that code.

Yeah, absolutely.

That's the thing.

That's the sort of instability I guess you got to think about.

Well, but from what you said, and I think what you said was exactly right and helpful, that kind of instability is what I have bought into.

That's the nature of the beast these days.

And what's your decision on on which piece show is it based it's based on the operating system?

The version.

Okay.

I see.

So if it reverts to to what 9.4 is doing.

Then you work around is a problem.

Yeah, exactly.

Work around this.

There's no there's no way of querying, whether the the general thing with UI stuff is it, you have to have the user in the loop to find out whether the UI code has changed.

There's no way of querying UI code to see which behavior it's exhibiting other than a heuristic like checking the version.

>>Yeah, exactly.

>>But if it's a change to the version, then that's going to affect everybody using that version.

- So you would just change the code back to what J9- - Yeah, once you detect the problem, it's easily fixed.

I mean, there may be a brief period of panic, but yeah, it's easily fixed.

- Or sometimes two to three weeks while you're waiting for somebody to launch their beta.

- Hey, I have another question.

I am relying on Amazon Web Services free tier for file delivery.

So Amazon is my AWS is my content delivery network.

And I've been running updates of the corpus daily and Amazon is a little vague.

I even now haven't gotten started to get warning emails about me bumping up to the limits, excuse me, bumping up against the limits of the free tier.

I'm a little unclear about how nominal charges or virtual charges are being incurred, but I am running up against limits somehow.

So I have switched over to weekly updates of the corpus.

And as a practical matter, that means that any changes, so changes include updates to GitHub repos, changes include new posts to the forums, Changes include updates and new wiki pages.

Those will accumulate not for 24 hours before being reflected in the corpus, but for a week before being reflected in the corpus.

Problem, nobody cares.

What's your thought?

- I haven't been relying on it for, you know, up-to-date content.

I generally that's, I'm not involved in, Either I am involved in which case I've seen it through some other channel or I'm not involved in which case I rarely I rarely am aware of it.

Um, I know that that Amazon has schools for looking at some of those those things, but I guess they want to keep it vague for the free tier to make people that people into buying stuff.

So, I think there may be some of that at work, or it may just be a big complicated beast and it's not the result of any malice on their part.

It's just the nature of the grand.

.

.

Kurt Vonnegut came up with the term 'grand faloon' for these giant organizations that have been around long enough and accumulated enough rules and traditions that they no longer make sense to the outside observer and it's possible AWS is just something like that.

- I threw a URL into chat.

I'm wondering if that's one that you've visited in your exploration of this issue.

- Interesting.

I will look at that later.

No, I don't think I've seen that.

Thank you.

Thank you.

Bob, any thoughts on that?

Yeah, you see, this is a case where my use of it tends to be historical.

All right, good.

So I, you know, and, and I think I'm in the same situation with Raul when I'm, I'm, I'm looking at the, you know, the mailing mailing lists and everything.

Every, every day.

And I'm, I'm looking at the APL farm for the J and the various other discord channels every day.

So that's where I'm getting my news, my nightly news or daily news is all coming through that.

So I guess the only thing that the Discord I would see in that case would be, I've seen something recently and I go to look for it and it's not there.

- Yeah.

- Yeah, I might notice that maybe, I haven't yet.

- Right.

- But I think the key thing would be to let people know that it's being updated on a weekly basis.

Yeah, of course.

I've already updated the wiki page to say that and I'll make it clear in the introductory post to JChat.

Would it be possible to make the changes to the hostings more often and it would be a lesser amount?

Maybe it would be, it would be a lesser amount.

I'm just thinking as opposed to updating the wiki.

Yeah, that's really how large databases and APL-like languages work, because you have a small section for frequent updates and a larger section for the.

.

.

Yeah, we did go through this.

It becomes very fiddly indeed to try to do that.

Especially with something that's not an APL backend or not a J backend.

You know, it's designed to be with a opaque interface to things like search relevance and stuff like that.

It's not meant to be composed.

- You also get into problems where you start thinking, well, maybe the end user application can retrieve the data and do the index updates.

And down that path lies horrible things to my mind.

>>Yeah, no, I know that was a discussion we had before, and it's not the good direction to take it.

>>But it might be worthwhile having somewhere in the UI updated as of and maybe a hint there as to about weekly or something like that.

Next planned update.

>>Yeah, let me think about that.

>>I don't know if there's space in Screen Real Estate for that, especially on a small screen like mine.

And what would we be looking at in terms of the cost if we were paying for AWS?

Well, I don't actually know.

I call it a budgeting support.

Sorry?

But AWS definitely has budgeting support for people that want to plan out that stuff because everybody in business likes to make plans.

Yeah.

Yeah.

Well, my plan was to try to keep from ever paying for it, which And that is not something for which they provide budget and support, I suspect.

But they do have a point.

I'm sorry, go ahead.

Did you have a they have they have a 12 month evaluation free, and they got a lesser always free.

They both they both options are provided.

Well, I will review the document that you sent me a link to for which again, thank you.

I don't, Bob, I don't think, I mean, I've got a certain J budget that shows up in an annual contribution.

Not talking about you making the contribution, you're making enough.

I'm talking about if we had a number to give to Eric and to say, you know, this is valuable and would J software support this?

Let me tell you why this is valuable.

Let me tell you that by far, the most important value this delivers to any individual in the community.

It keeps me distracted.

I know, but that's not the point.

You can go on being distracted if somebody else is paying for the ride, you know.

Well, okay, we can have that conversation if we have to.

No, but the reason I ask that is because I was involved with television, and we had this long discussion at the television station about what we could do, what we could do, we're having this problem and it was you know we could change this and change this and somebody said the back you know what if we spent 150 bucks on a piece of hardware it's fixed we all went do it because we were bending ourselves into positions to make something work that could be solved with 150 dollars yeah yeah and so that's why i asked the question is what are we really up against if it's not a big thing, if J saw the value and they wanted J software saw the value and they were to support it, does this all become a lot easier to do?

I don't know.

Yeah, I see.

I see.

All right.

Well, let me again, review the document and try to get some clarity around things that I've been deliberately vague about up until now, costs in particular.

Okay.

We can talk about it next week.

Yeah.

There's probably other better documents than the one I forward for that.

That's the start of the string I can pull on and I appreciate that.

And when are you back from Dublin?

Are you back for the winter?

Yeah, we're coming back in October 25th.

So less than two weeks now.

We'll be back in Texas.

Okay, well, so a week from now we'll be Dublin and then two weeks from now we'll be back in Texas.

Yeah, I won't have to stay up as late.

You won't have to stay up as late.

It'll be middle of the afternoon for you.

I'll be interested to hear whether you think I'm any more or less coherent as a result.

I'm always curious about that.

I definitely feel impaired at this time of night.

Yeah, we see if we thought you were more impaired, I think we'd blame it on Texas.

Oh, all right.

Fair enough.

You've lost the culture of Dublin.

Yeah.

Yeah, Texas has its problems, there's no question.

Well, yeah, I think it's a fine place for people who live there.

Yeah, they've been living there all their lives and they never leave.

Yeah, yeah.

It's different.

Okay.

Was there anything else?

anything else?

No, let you go to bed.

Okay.

Well, Bob, thanks so much as always.

Take care.

Thank you and take care as well.

See you, Raul.

See ya.