Wiki/Report of Meeting 2023-04-20
Report of Meeting 2023-04-20
Present: Ed Gottsman, Devon McCormick, Raul Miller, and Bob Therriault
Full transcript of this meeting follows this synopsis
- We started off with Ed's seventh demo of the new J wiki browser. Ed mentioned that Bob had created a prototype that allowed child controls to change their visibility and size which makes the appearance of the browser more flexible to respond to Search requirements and NuVoc screen space. Ed had also created the an animation of the centre pane sliding over the table of contents. Raul found that this was a bit distracting and Ed feels that it is probably not worth the distraction for what the gains are. Bob wondered if fading colour might allow the background to be less distracting. Ed also reduced the font size, as had been advised by Stephen Therriault, and taken the palette of NuVoc. Ed chose not to use the palette of the wiki and Bob agreed that there should be a distinction between the viewer and the web pages. Ed suggested that it was an open question.
- There was some discussion about the hierarchy of the wiki structure and how it is represented by indentation in the table of contents. Ed raised the question about whether there was a different way of organizing the wiki based on popularity of the pages as opposed to the wiki structure that emerged as it grew. Bob felt that the table of contents does not have to be an either or and that two category trees could be created. The issue with that would be that we would be dividing our resources for indexing. Popularity might be a particularly difficult indexing challenge as it might have more evolution over time. Other options would be to have particular interests reflected in the table of contents such as mathematician, engineer, computer science etc.
- Ed has found that search has been a challenge because the search field below tends to hide information as the cursor moves over other pages. Bob suggested that the search bar might be put above, but that would require that Ed split the table of contents and the centre pane into separate isigraphs. Devon wondered if the table of contents could be made into a tab to get it out of the way.
- Ed discussed making the font size smaller, since the larger the font, the less whitespace is available. Stephen had also mentioned that there was a lot of information on the screen and felt that reducing the font size while retaining the target size allows quicker scanning while maintaining same target size. Ed tried out a smaller font live and it seemed to address some of the issues.
- Ed talked about issues that he had with the search results being displayed in the table of contents. Raul wonder if using colour to separate the search terms from the results. Bob wondered if there was an opportunity to put the results in the centre pane under the different locations. This was similar to earlier versions, but there were issues with the number of search results fitting into the centre pane.
- Raul showed an issue that he has seen with Forums and the way that his dates displayed. There seems to be an issue with interference on the forum dates. Also, some of the search results are not showing up and Ed thinks that it might be an http https issue. It may be an issue with the ampersands not being recognized properly by the browser. Ed's feeling is that he will need to keep focus on the smaller displays because they seem to be the source of many of the issues that have shown up. Ed is exploring the option of shrinking the browser window until you want to display it and then having the window expand when the mouse wheel event happens on the webview. This will be solved when webview can view mouse move events, which is apparently coming in the next iteration of jQt. Mouse events are sampled so that it is hard to know exact position of the mouse. Bob wonders if there is a setting that we are missing on the webview window that allows us to use scroll bar.
- Ed has added a black border that will lock the screen for 3 seconds while you move across the screen and that seems to be a good interim fix for the inadvertent hover. Raul wondered about whether there is a way to convey the structure of the program, so that he might be able to better trouble shoot some of the issues that he has found.
- Bob mentioned again the popularity index to structure the tree and that it should be considered for a while before any commitment. Ed agrees that there is work that will always be done in curation. Bob suggests that NuVoc might provide a portal into the wiki information with links to J Playground for examples. This could be provided through the category tree with links to the primitives.
- Bob also noted that Wiki Hints has moved over to the side bar as it is more maintenance than it is information about the J language. https://code.jsoftware.com/wiki/Category:Home He also showed the Delete Update and FixMe pages that could have links for the pages that require attention https://code.jsoftware.com/wiki/Category:Delete We may have regular review in the wiki meetings to go over some of these pages. Devon mentioned that this should be in the Contributing to wiki page and the How to use this wiki page. Devon invited Ed to show his work at an NYCJUG meeting and Ed planned to do a presentation at the May 9 meeting. https://code.jsoftware.com/wiki/NYCJUG/2023-05-09
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 April 27th, 2023.
Transcript
There we go. And I guess to start off with Ed, I'll turn it over to you to show us what you got.
Right, so it's been a good week, I think. So of course, Starters very kindly put
together a working demo of what I think of as a malleable user interface, which
is not a great term because most user interfaces are malleable, if you put it to a better term.
The notion he came up with was to change the dimensions and even the existence of child controls, depending on what part of the table of contents you were looking at.
So in the normal course of things, which is what we're seeing here as we're looking at forums, going into tags and so on, there are no surprises.
close observer.
Go ahead.
You're not sharing screen yet, I don't think.
From your description it sounded like you thought you were, so...
Very nice, that's very much.
How...
Where are we here?
There we go.
Is that it?
Yep.
Alright.
But...
Turns out when you move to something that needs it, like say search, it needs a couple of additional child controls.
that means a clear searches button and a search field.
In the lower left corner of the screen, those elements appear when you hover on searches.
If you hover on something else, they disappear.
Similarly, if you hover on Newbook, the interesting thing about Newbook is that it needs extra space if you've got a small window.
So hovering on nuboke expands the area dedicated to detail and shrinks the browser.
So the ancillary pages, that's reversed and we're back to 50/50 for the detail area in the browser.
I like that what I had done was
To make the table of contents in the left margin
disappear or shift like a curtain in order to make room for new book and
If we if we take the approach of having the child controls
Rearrange reapportion their real estate. We don't need to do this
So I will probably take that out. Although it is kind of nice if we know that you're not actually looking at the table of contents at the moment, that you're focused on something else.
It is kind of nice, I think, to have the table of contents get out of the way, even for something like forums, for example. So I'm not actually sure.
that's fair. All right, it's easy enough to take out, much easier than it was to put in, so I'll just drop it.
The other possibility is you don't have to go quite... you don't have to reclaim quite as much real estate.
So if instead of sliding over that far, it slid over, you know, like about where the the years line up, that might be enough to give it some, you know, more presence.
Right, but it is to Roel's point distracting. And if you're not, if you're only going to recover, if you're recovering, you get even less real estate than I'm recovering now. I don't know if it's necessarily worth the distraction. I don't know.
Yeah, go ahead.
Oh, I was going to say when I was talking to Stephen, he had an interesting take on
how to increase focus and sort of put stuff into the background.
If you grade out that table of contents when it slides over, as opposed to keeping it quite
as sharp, I think you'd find it dropped into the background pretty quick.
I think the fact that you can't read anything is a pretty good drop into the background.
Yeah.
On top of that, maybe the gray doesn't strike me as like it's going to add much. But maybe. I don't know. I think it's sort of cute, but maybe...
the um...
uh...
[typing]
...heater than it is useful.
Um...
Whereas I think the
way the browser rearranges itself
and, Ronald, correct me if I'm wrong
on the other side of yourself yet,
this strikes me as less
distracting, less cute,
less of a half.
Um...
So I think what I will do
do is take it out for now but it'll be easy enough to put back in again if we
discover that we for whatever reason need the real estate or that we like the
mechanism after all. The other thing I did, and this was the session, this was from the
session that Bob had with Sam Stephen, I reduced the font size in order to put
more space in between elements on the screen where all the elements are also
the elements are textual. I also took the palette from Newbook, so I used some of
the colors from Newbook as a way of trying to rhyme a little bit with the J
website or websites. I'm a little nervous. Visually it needs a lot of help, there's
no question about it. I'm a little nervous about trying to take the color
scheme of the web over here and apply it over directly over here so that would be
for example turning all these links blue only because the interaction is a little
different so it strikes me that the interaction on the left is a little
different from the interaction on the right so it strikes me that it shouldn't look
the same, but I'm open on that, I'm not sure.
Well, you and I are on exactly the same page on that one,
and I imagine if when you saw the session
I had with Stephen, the first thing he asked me was, "Have you thought about
trying to match up the right to the left?" And I said,
"I think the interaction is different enough that I'd rather them look a little bit different."
So I absolutely agree with you.
But it's an open question whether the interaction should be different.
I mean, I think in principle there's no reason why we couldn't support clicking,
rather than hovering, and that may in fact ultimately be where we wind up.
One of the things I get, another thing I get is, by the way, you'll notice there's no scroll stripe here,
and it's not scrolling.
As I am over on the left side, and it's not scrolling, to scroll you use the mouse wheel.
What I found was that with this mode, when you came back, you had a tendency, or it was possible,
to wind up inadvertently scrolling the table of contents on the left.
And that was not working for me as an interaction.
And now that I've got the mouse wheel working, I'm thinking, "Gosh, I think the app will do what they were doing with the mouse wheel. This is pretty cool."
Not to have to worry about inadvertently scrolling.
The other thing that's nice about it is that removing the scroll stripe, the vertical scroll stripe, means that the indentation in the table of contents is a lot more obvious.
visually it's not being fought with by the scroll stripe, which is a point that
Seaton made. He had an interesting way of expressing it. He said that the
indention, the fact that it was an outline, did not read very well in the
display as it was, and I think he's absolutely right. So taking it out, I think
it's the fact that it is an outline that is a hierarchy, visually a little more, a
little more obvious
And we can play around with it, but the other things he was talking about was actually changing
Maybe a slightly lighter
grayscale on the lower
Hierarchy the lower parts
So that the higher the the darker it is the higher up the hierarchy you are
Because I was saying to him on that on the table of contents. You've really only got three levels ever
Yes, and it's not a team over the next one. Yeah
I think it was a point that I made to you all in an email that um
There there's already sort of an implicit priority here
That new book is on the top, and that's the thing that's perhaps most often referred to yeah
And I don't know that this is possible, but in principle there could be a prioritization as you move down the hierarchy,
in that the further down you go, the older or/
less immediately useful or
I'm not sure, or more out of date,
the material might, in theory, become... it's hard to achieve, I think, but it's a curatorial problem, not really a presentation.
problem. But I think the fact that parts of the hierarchy are much lower than
other parts is already pretty apparent. I don't think we need to turn parts of it
gray, to make that clearer. Maybe the turn parts of it gray indicate that, hey, this
is wrong. In other words, there's a lot of material in here that's simply no longer
accurate because the environment, the language have been gone for succeeding decades.
But I don't think using color to indicate that you're further down the hierarchy is necessarily helpful.
Yeah, actually I had put down as one of the agenda items is talking about the organization of category trees,
whether we do it popular versus the Wiki structure.
'Cause right now what we've got essentially is a mirror
of the way the Wiki was structured
and the way it grew organically.
So I haven't gone too far off that process.
Although I have introduced newcomers and developers
and referenced sort of as separate categories
that wasn't done before, but within that,
the categorization is pretty close
to the way the Wiki was structured.
And there was gonna be a discussion
about whether, and I don't think it's an either or,
that's the other thing I was thinking.
If you could come up with a way to structure it
in terms of interest or a heat map,
you can always build another category tree
that goes that way and you could just go to a different one
if you didn't like this one.
- That's a significant undertaking of course.
I mean it's a couple of thousand.
Yeah, it is. And the other challenge to it, so this one we're talking about it now, the
other challenge to that is it's probably a bifurcation of effort. If you get people going
two different directions, you know, depending on what ones you're working on, I think you
could end up people working in one area and another area being neglected or vice versa,
and I'm not sure you get the same...
We're not overwhelmed with indexing resources.
No.
No.
But I think if we were to say that we had a way to do the heat map, for lack of a better, the popularity way to do it,
then I think I would leave the existing one pretty much as it is and we would focus all our attention on the heat map
and I think the heat map is actually quite a bit more challenging to do.
Now do you think it's challenging to do from a curatorial perspective or from a presentation
perspective? I think it's from a curatorial perspective for sure.
Yeah, fine. I'm sure you could come up with a way to do it in this presentation. I don't doubt that,
but to figure out what the most popular things are,
what people are looking for,
you're looking at such a wide audience.
Unless you did a, you know, we actually focused down to,
you know, sort of micromanage it.
So if you're an engineer, these are your resources.
If you're a student, these are your resources.
If you're a mathematician, these are your resources.
And do it that way, 'cause you could probably get
the most popular things that way.
You could break it up that way, I guess.
- I think you can do that in this context.
You've already got multiple indexing.
You've already got particular pages appearing
under several categories independently.
- Yeah.
- There's nothing stopping you from adding
half a dozen categories immediately below tag,
for example, or even below forums or search,
where they would reflect your judgments about popularity or utility or whatever to various sub-audiences of the J-programming community.
I absolutely agree. This is absolutely flexible enough to accommodate that for sure.
I want to return to a point I let go, which is the notion of the fact that when you hover on search, you get a search field where you can...
I don't think I'm resonating to that only because it is too easy.
I'm sort of resigned for now to the fact that it's easy to inadvertently hover on something you don't mean to and select it,
and thereby have the informational landscape change on you significantly.
I'm a little more nervous about, here I am on search, and I decide I want to clear searches,
and I go down to do that, and in so doing I lose my clear searching button.
The idea that user interface elements are inadvertently disappearing on me without my wanting to,
that I'm doing it inadvertently, I'm very nervous about that.
What about if we position search above Nuvok?
And then put the search field above?
Well, you can't put it above, I don't think.
Although I guess that would
solve the problem that I'm looking at.
The problem with that is
if you put it above, when you hover on search,
the entire
left side, that that child element, that is a graph, will drop down a couple of lines to accommodate the search field and the Clear Searches button,
and that means that you will no longer be hovering on search. You will now be hovering on ancillary pages or any font,
which means the search field and the Clear Searches button will disappear, are you?
So the only way I could think of to make this work was to put those two changeable fields on the bottom
where they wouldn't result in the table of contents shifting down underneath the mouse. I don't know if that's clear.
Yeah, I think when I gave you my example, what I had done is I'd actually broken, for lack of a better term,
term the line between the table of contents and the display area is two
different of is a graph side essentially right and I think you're doing it as one
right I'm not sure
that
than
book
that
This mechanism is very much where the browser shrinks, the WebView shrinks, and that is enamored of this mechanism.
Where we actually use our game controls. That's all I'm saying.
But I guess your interface is much more complex than the demo that I did.
I did. But if you were able to split at that line of Table of Contents, then you have more
freedom into what that next section does, that middle pane does.
So I'm not tracking.
So if you just load that right now, you should be able to see the...
Yeah, I'm not able to.
[inaudible]
[inaudible]
[inaudible]
[inaudible]
[inaudible]
That's what it's claiming. How did I do this before?
[inaudible]
[inaudible]
[inaudible]
[inaudible]
It says J WikiFlix, not WikiFlix.
The name begins with a J.
[inaudible]
I think it okay
Okay, you set it up so that um
The control
Surfaces the green the red and the purple yeah, not we're not affected by the appearance of the search field
That's great if you can do it, but with our interface, I'm not sure I can do it.
Well, at least as I said, you've got a far more complicated thing going on there.
But when I was thinking about it, if you split that next level of search away from the table of contents...
Oh, I forgot to repeat that.
Yeah.
Interesting.
And then you've got... I believe you've got a lot more options with that middle pain if you do that.
Alright, let's think about that.
[silence]
OK.
I will look at that.
All right.
Could the table of contents shrink,
dad, to just be like a little tab with TOC on it?
When?
What happened?
Well, when you shipped over and you get more detail
on the things you're drilling into, like there.
Oh, so visually manage the disappearance
of the table of contents as a tab rather than sort of as truncation.
Yeah, I'm just saying so that I maybe would save you probably about be about the same space wise, actually.
Yeah, visually, it would look better. Just sort of leaving partial screens up is probably right.
Because you can't really, you have to remember what that is, where if it just said TOC.
Right, exactly. I'm really thinking I'm just going to take this out. It's cute, but Torello's point is kind of distracting.
And Bob's mechanism of recovering realistic by shrinking the browser, I think, is pretty good.
What else did I do with anything?
Right, I used the new vocal palette, took out the scroll stripe, produced the font size.
Not sure about search disappearing, talked about that.
You got font size in there?
Yeah, I took the font size down. So that was the result of Bob's session with Steven.
I'm not a visual designer, Steven had to point out that
that, you know, in a typical web page, the font size is actually smaller than what I
had before. I was using the 16-point type. And what happens when you do that, what was
happening when I did that, and the line height that I chose, was that it was very crowded.
Stephen also objects to this, I think he objects, I'm not 100% certain about this,
the sheer amount of information that's getting shown.
He mentioned the idea of progressive disclosure as a way of sort of hiding a lot of the information
until he indicated he were interested.
I get that philosophy, and for a certain class of user it's necessary, there's no question about it.
about it, I think that for the audience that we're targeting, it's less necessary.
So I did take the font size down, but I kept the line height to say there's no more white
space or we can't hold the screen.
Yeah, I disagree about the hiding thing.
Personally, I want to see as much there as I can.
See, we're different, right?
Steven is not targeting us typically, I think, with these works.
'Cause if you've ever had to track something down
in like a sparse directory tree, you know, it's a pain.
- Oh yeah.
- Yeah, and I think that when he was talking about that,
when I was talking to him about that,
he was just regarding that as that is a challenge.
Like there's an awful lot of information there
that is a challenge.
One of his ways of doing it was disclosure,
but the other way is exactly what Ed's done,
is you keep the target sizes the same
and you just reduce the font size and you get yourself more white space that way.
And as he pointed out when we were talking, he was saying,
you're really looking for speed and that's probably the...
if you give more white space, you actually can increase the speed people can scan,
and you increase the speed in which people can track and hit targets.
Because the target is just as large.
It is effectively just as large. The targets are built around line height, not around font height.
I can definitely see how visually it would be easier to scan if the text is not as crowded.
So yeah.
And I think you could actually go even smaller with your font size when you compare it to what you're seeing on the...
Yeah.
This is me hotdogging here.
Yeah, yeah, no worries.
Those are a little of a demo.
Well, this isn't so much of a demo, is it?
You know, it's, you know.
Are you going to start doing Twitch, Ed?
You're going to do this stuff real time?
[LAUGHTER]
Yeah, that can get pretty small.
That's a cool point.
Yeah.
Yeah, if I was like, I'm just looking at the sidebar in the wiki and it's probably maybe 13 you could go to.
But the difference in white space is significant in those two levels.
Well, visually, I take your son's point on that. No question about it.
Yeah, it might be eventually be a tweakable thing. I've got something to show you that a problem that you haven't seen that related.
So,
some point. Oh,
better. I express my embarrassment for introducing infinite loop and inflicting it on you. I apologize.
I've done worse to myself.
I don't mind doing it to myself. It's doing it to others.
I get that.
We signed up as beta testers. Come on, you know, we knew it.
the whole thing. I'm not sure what you're getting into. Oh, you're embarrassed that I had to have Bill find it, find the problem and find it myself.
All right, yeah, so that's 13 point type.
Maybe Raul's point is right, maybe it should be tweakable.
That's certainly easy enough to do.
Right, well that is all that I've got at this point.
So my to-dos are to take out the hiding TOC, the shy TOC, table of contents.
I guess other than that, I'll still have it on the agenda.
I'm going to declare victory on the color scheme for now.
One of the things I would like to fix is the search results display.
So it just uses the standard layout, which is a hierarchy.
but it's not at all obvious what's going on here. You've got biadic transpose,
you've got random fixed seed, and for each one you've got some results underneath.
But partly because of the bar chart, the indentation is not obvious, and there's no
visual distinction between the keyword you typed in and the result categories
that came out, if I would like to think about it.
And that is...
I was saying you could use colored text for that.
Have the sort of result...
Yeah.
Or you could draw a border around it,
you know, underline it or italicize it or...
Yeah.
Yeah. I don't think it would be hard, but I do need to do it.
And the only reason I hesitate a little bit is that it's...
I wouldn't need to go in and make a special case that right now it's very general, it's just showing a category hierarchy, no problem at all.
So it's just a matter of...
What about only having the search term in that column, and then when you get the results it's split out the different ways?
So that there's a section of wiki...
Yeah, I think that works.
until you start to have a lot of results.
And then you'll be happy that the results are limited to a single column,
that all of your results, sorry, all of your result categorization is limited to a single column,
because this list can get fairly long, particularly if you've got a small window.
So I think I like the idea of leaving it all in one place, but you're right, we could
dedicate another chunk of real estate, another column, solely to the categories.
No, I was actually thinking that in your display of what you see under Wiki there,
the essays down to Wiki report at the bottom, put that as... that's displayed when you go to
dyadic transpose. But then underneath it is a section that's called J-programming,
and it has its results shown.
So that all you got...
JProgramming is a category of Wiki.
Right, I think when you originally had your search,
the first time I'd seen your search,
the search results came out and they were split,
so you could see some part of it was Wiki, some part of it was JProgramming, some part of it was JChat.
And the actual results looked that way.
Well, what would happen...
and actually, it wouldn't happen any more...
what would happen is, if there were few enough results in total for the search term,
they would all get shown on the road at once and the major categories would be highlighted.
The major categories would show up in bold red.
And then if there were too many items to show that way,
you drop back into this mechanism where you have the categories on the left,
you have to pick the category you're interested in.
Yeah.
What I decided to do is default to that.
So I can certainly go back to trying to show everything,
but I need to fall back to showing individual categorizations of results.
In the event that there were too many results.
And looking at it, unless you had a lot of different search terms that you wanted to keep on record and go through,
that's mostly an empty column most of the time.
Right?
Well, depending on how big your window is, yeah, that's true.
Yeah, so I think you're...
The way I'm thinking of it is you actually may be better to differentiate between the search term
and the different locations that you'd find stuff and then use up that space
anyway because I don't think that space is gonna if you just went with search
terms alone you're you're really not going to fill that space most of the time
I think a good point um I guess the main message is I'm I'm definitely I've
definitely got capacity to do additional stuff and with that in mind, Raul I'd like to invite
you to show the--
>> Oh, you might, sure. Let me--you'll have to stop, I think, before I can--
>> Yes, I will do that. Go ahead.
>> Okay, so I'm going to share my screen and I get a--I think it's this. I want to share
Now, one issue of course, I don't know how it's going to show up in the thing,
at the bottom of the screen, this is just a little bit bigger than my screen.
The big one for me though is the forums.
I go to the forums, you can see that my dates are displayed a little bit oddly.
And the secondary one, I haven't figured out what that is.
I haven't tried any debugging here.
A secondary thing though is if I go to searching, the search results don't really seem to be,
there's something wrong with my machine, I guess, that's not giving me real search results.
That's all.
- When was the font thing?
- The only thing I can think of on the search results role
is that I'm having trouble with get HTTP.
So get HTTP works great up to two headers,
but after that it seems to flip over to some other mode.
And this is me guessing at this point.
So I'm actually spawning a curl.
And that may not work on whatever machine you're running on. It works fine on Macintosh's, but may not work for you.
For me, HTTP is a curl.
Oh, it is?
Yeah.
Alright, so that area that I learned in Confident Expression that I just made, just ignore that.
[laughs]
Because clearly I'm wrong.
I mean it's curl.exe says Windows, but it is a curl.
OK, then I assume that... I think I'm using 2 bang colon 0.
I assume that should just work.
Yeah, 2 bang colon 0 probably won't get curled because curl is not in the path on my machine.
Oh, then maybe that's what's going on.
I
Mean it's on my list to get HTTP working. I just need to post a question. I'll do that
Like just need to isolate it down to something like if you look at HTTP
It should be calling either curl a curl or wget
On your machine is probably fair
Okay. Yeah, I'm sure it's curl
The problem is that when I have what I've discovered experimenting with getting HTTP is when I have more than two headers
it seems no longer simply to return the HTML that it recovered.
What do you mean by more than two headers?
Parameters, I should say, empress and parameters, or -h headers.
That may be a bug that needs to be fixed.
Well, I will post something that will let the appropriate person track it down.
But what seems to happen is that no longer simply synchronously returns the text in its retreat,
but rather it asynchronously writes to an output file somewhere.
Right, it sounds like you're getting background processing. It needs to be quoted somehow.
Yeah, I tried a bunch of stuff, I wasn't able to. But I haven't taken the next step of exposing myself,
exposing my ignorance of the community, because obviously the next thing I have to do
But more interestingly, we need to figure out how to handle even smaller screen sizes than I had originally imagined.
Because that pretty clearly is what's happening to you.
Yep, I've got a tiny laptop.
But it was cheap!
Well, you know, we have a saying, "Cheap is cheap."
There's nothing wrong with tiny. Tiny is okay.
When you create your application, you've got access to the dimensions of the user screen, right?
Yeah.
Physical screen, yeah.
So is it possible to make it, like I think I did this when I was working with Jig, I
would I think create a window that was about, I think about a third of the size of the user
screen.
That was my default.
They could make it bigger or smaller after that.
But by doing that, like I just took their screen size and said I'm going to make it
about this size.
Now the thing with your display is you'd need to adjust fonts as well to be able to fit things in. That's where the real challenge comes.
Well, maybe. That's one thing you could try to do. The other thing you could do is just to keep shrinking the browser window until it disappears down to nothing.
Yeah.
Not a great solution. Maybe.
One of the notions that I...
I thought you had 500 pixels high, which should have been...
my screen should have been big enough for that, so I don't know quite what was happening there.
Yeah, no, it is very odd. I would have expected overlap.
No, I wouldn't even have expected overlap, because I'm not...
Well, you're right. That is mysterious.
I've got 768 pixels. You're asking for 500 high.
And even with other decorations on it, that's about at least, almost 200 pixels of excess above that minimum height to account for.
I didn't see that much going on.
What resolution are you looking at your screen at? Are you looking at full resolution?
Yeah. It's a laptop. There's no scaling involved.
Well, okay, I guess maybe with the PC, but with a Mac, I can revert to a lower resolution on my screen.
Right. Yeah, there's no doubling or whatever the term for it.
That is mysterious because I don't adjust anything in response to the physical dimensions of the screen.
So the forum years and months should not be getting messed up in the way that Raul was able to show them getting messed up.
Yeah, I should probably spend some time hooking into that, but...
I
Please only if you're really really curious
But I really do it as my responsibility and I will do that um. I think maybe it's time to try to
Support even smaller resolutions that I'm supporting that I don't remember see
and right now
1353 by 836 is the smallest
Resolution I'm willing to support
1953 by '36.
And I think maybe I need to be willing to support a much smaller resolution than that.
One of the thoughts I had was to
dramatically shrink the browser window.
But then when you mouse over it, which we can't detect unfortunately, but if we could,
When you mouse over it, it would expand in the same way that it shrinks for nouveau with Bob's approach.
So you're mousing around on the left side of the screen doing your navigation,
but when you get to something you actually want to read, which we know because you move the mouse to the right over the browser
pane, WebView pane,
we expand the width of the WebView pane
appropriately, and it maybe
significantly
occludes the detail pane on the left.
And we can't do it with mouse movement, but we could do it with mouse wheel, with scrolling,
because that is reported by the WebView.
So when you start scrolling in the web, on the web side of the interface, the WebView
could expand its width significantly.
You were saying mouse movements can't be tracked on the web view right now.
It's a bug.
I think Bill acknowledged that and said it'll be fixed in the next iteration, whenever that
is.
I'm not sure how often they update JQT.
But what I'm thinking is, you do know when the mouse leaves the application, like the
the wall of the wall?
the house the house is the house has moved the house is moved the house is moved and
at some point I stop hearing and really all I can assume is the mouse is hovering at the
last reported location. But it's not. It's left the...is it...is it...is it...is it...is
a graph entirely and is clicking on links over in the web view. But I have no way of
doing that.
What about putting a thin, we sort of played with this before, so that when you head towards
that browser screen, you're going to cross that little barrier. At some point the mouse
is going to go, "I was here." As soon as you see it on there, you expand the web view.
No?
the world but most reporting is not continuous it's sampled and if you go fast enough which is not harder to do no mouse event may be dropped on that barrier. The solution to that is to make the barrier very wide and then you're in a good place.
And we can't overlap components. That's the other challenge. If you could, then you could sort of get around it that way.
Maybe. Maybe.
But I haven't seen a way that you can overlap components.
No.
I think maybe scrolling might be a pretty good second best event to detect.
So if you actually start scrolling, and you will start scrolling because the browser window is so narrow,
you'll have to scroll in order to see anything at all.
And if as soon as you start scrolling, you expand the width of the browser window, that might be good enough.
Possibly even better, potentially, I don't know. But good enough.
I don't know, what do you think?
I think the best solution is to be able to track mouse position through the web view.
And this might be a stopgap until we get that.
Exactly.
I wouldn't stay with that, but we could try it and see what people think.
Alright, well let me see if I can dramatically reduce the minimum size of the window to the point where it will fit on Raul's screen.
And maybe if I do that, the other font problem, the text rendering problem that Raul demonstrated will evaporate. I don't know, but it's possible.
Well the other thing that happens with the WebView is if you make it smaller than the content, you get the scroll bars, don't you?
I haven't seen them on yours, so I'm wondering whether there's something, a setting, because
when I was playing around with Jig, I get scroll bars at a certain point, but I'm using
SVG as my source. But there might be a setting.
When I'm creating my images with Jig, I basically take an SVG display, so I can see the image
so I can zoom in and out on it, and I wrap it in HTML and fire it at a web view and it works fine.
But when I...
No, I don't think I've ever seen Spelbow, but that's the Mac for you.
Yeah, I'm thinking it might be a setting somewhere in JQT, like it's one of those max/min things,
that it will actually, if you set your constraints on the cropping is too intense,
it'll give you those bars that you can slide around on the screen to see what you're seeing.
And that would encourage use of a scroll.
Oh, I think just not being able to see much of anything would encourage the use of a scroll.
I feel okay about that.
No. There is one other thing I'm doing that I'm not entirely happy about, which is, and I had this in at one point, took it out, and I'm putting it back in now, which is, you know, if you're looking at a page on the right and you want to interact with it, it's a problem to move over because you'll wind up loading some other page.
So the hack, and it is a hack, is if you click, it gets a black warning ring around it.
And that locks up.
It's interesting.
We're not seeing your screen right now, right?
Oh right, of course.
I am so used to being the center of attention.
Ha ha!
Gotcha.
Yeah.
the theory is that if you've got the black border, you can safely move around without, you can navigate, but you're not loading any pages.
I think that's fantastic. That's great.
Well, it's sort of fixing a problem that you could argue you shouldn't have in the first place.
But yeah, I think it might be progress.
I don't know, let's see.
And I think that's all I've got at this point in terms of demo or discussion.
So if there are other points people want to bring up, I think that would be great.
Just a general comment now that you're in beta.
If you can give us tools to look at specific parts of the system, or even just, you know, a little paragraph of notes or something, we can actually...
I'm thinking that in terms of the problems that I have since they're specific to my machine, you're gonna have a terrible time trying to investigate them.
And if I knew where to look, I could maybe be a bit more helpful.
And I appreciate the offer and I appreciate my own responsibilities in that regard, so thank you.
I think that for the particular problems that you're having, if I can just make a window that's small enough that it will fit on your physical display,
That will be progress right there, and I will undertake to do that next
In terms of
Ralph so makes a good point if you're going to deploy this there's a whole deployment thing that you need to consider at some point
Oh, and I'm I'm not yet to the point where I'm thinking about deployment
I'm just sort of sharing among a very small group. Okay folks
Precisely because I haven't thought through deployment
But, Roel, I wonder if you could peel back a layer of detail on what you mean by what I could do to make it easier for you to investigate.
Well, in my case, I was thinking about looking at what you're doing with the forums,
but I don't know what the naming conventions are. I don't know. I see a whole bunch of references
of forums. I don't know where I would need to spend a good half hour on that just to figure out
what data structures I should be looking at to see what is, to even know what I
Before I can start making progress, that's meaningful, I guess.
And I don't expect necessarily it should be hard for you, but if you just say this is a routine that does that,
and then I could maybe change local assignments to global assignments or something to expose it to me.
I'll just mention that there's nothing, programmers always say this, they're often mostly, I've
said it repeatedly, it's been a lie over the years, there's nothing terribly complicated
going on.
Basically I've got a SQLite database and one of the things that might be helpful to you
if you're interested in understanding the data structures better would be to get the
SQL-like browser for your Windows platform and just open up that database.
There's six tables in it and they cover categories and pages along the history
menu and there are a couple of others. And essentially all that happens for the
most part is when I go to show a forum for example, I just retrieve the entire
contents of that forum out of the database and keep it in a local memory
cache and then I just do indexing to pick up whatever the appropriate forum posts
are as the user interacts. And that pattern of retrieve a bunch of data from
the database, keep a local cache of it and navigate it is repeated for all the
different aspects of the interface. So a good first step would be to be able to
look at that database. Let me put some notes together on the forum, the forum for navigation
part of, I gotta write this down because it's, you know, two in the morning here. I'm just
not going to forget. I am going to forget. Yeah, I forget quite often too.
I've got a couple things we did talk about the structure,
the organization of category trees,
whether we reflect the structure,
the underlying structure of the Wiki as it was built,
or whether we go to popular.
I think that's something to think about.
I don't think I'm comfortable making a decision
one way or another to go that direction
'cause I think it's a significant thing.
But if people can think that there's enough reason to do it,
and it probably is worth me investigating,
the thing is it is an option
Because you're building category trees.
Yeah, I mean you already alluded to this a couple of weeks ago,
where you pointed out, or maybe it was an email,
that now that you've got, now that ancillary pages are so accessible,
the category is sitting right under New Book,
you could have additional pages, additional categories,
under New Vogue, like indexing, like special combinations, like other stuff.
And just that would be a way of putting more popular or more commonly used or more useful material
in front of users, from the center, without fundamentally changing your approach to curation.
So I don't think it's an either/or. I think you can do everything you're doing now and do some popularity or utility-based organization.
And I think this is another piece of the presentation, creating opportunities for curation that might not have existed before.
I've thought about using NuVoke as more of a portal,
and the same way as you're talking about linking out of NuVoke to category trees that would group things,
also linking to the playground so you could actually play with examples of the same way.
And that way you're focusing the user to look at NuVoke,
and now NuVoke is a much more powerful way to explore the language.
So you're almost not looking at the wiki anymore.
Well, but in terms of... I wouldn't say that. As a presentation question, if you were strictly on the web and you didn't have what I'm thinking of as the Jane Wiki viewer,
If you were just on the web, yeah, you'd link out of Newboat the way ancillary pages do now.
Although that's kind of unpleasant because they're not actually visible when you're looking at Newboat.
You have to scroll down to find them.
And as you augmented that list with indexing, special combinations, and so on, it would become even less accessible.
But if you do it in the context of the JWiki viewer, if you do it as a subcategory under the new vote category,
so you've got the ancillary pages subcategory, you could have an indexing subcategory, you could have a special combination subcategory,
you could have a, I don't know, mathematics subcategory, what have you.
So the top ten items in the table of contents could be the most commonly used, most interesting, most useful collections of stuff you could identify.
And that wouldn't require a fundamental rethinking of the wiki structure in any way.
You could also put some of those links back into the actual primitive descriptions.
So that you're not looking at ancillary pages, you'd actually see something that said mathematical,
mathematic examples, you know, and link you to other, this is one of the selection, you
know, primitives, and then you click on that, it'll take you to a page that just tells you
about the different ways to select and the different primitives.
Again you're developing new, which is, you know.
Which is good. There are two ways to develop Do-Book. You can develop Do-Book by putting links on a Do-Book page, which is how ancillary pages work in today's wiki.
And I'm not happy with that for the same reason Raul found the out of sight, out of mind notion to apply to what happened when the table of contents disappears.
is not visible, but if you put those pages under subcategories under the
new vote category in the table of contents, they'd be much more visible.
Covering, which is the most casual, easy thing you could possibly do, would bring up this whole plethora
of useful pages on special combinations, for example, or on indexing and any
I think that, I just have a bias for externalizing structure into the left side of the interface, rather than burying it in pages in the right side of the interface.
I guess that's where I come out.
Other thing I'll show you, just to share my screen.
I think this is going to come up.
There we go.
So we're back at the homepage.
You might have noticed a few things that I've changed.
There's no more Wiki link anymore down here because I've put Wiki hints over here.
So it's on the sidebar now.
I also cleaned up all the references to fix me and stuff.
And now what I've done is, and we can talk about whether in future meetings this is something
we can take a look at.
If I click on Fix Me, there's no pages I've got right now that have got a Fix Me link.
So that's not a good example.
Let's go back to, I'm pretty sure there's a delete.
There we go.
So on the delete page, if you see a page that needs to be deleted, you'd go to the wiki
hints, you'd go to delete, and then you'd enter, you'd have to be in edit mode, but
you enter the link and then maybe the reason it should be deleted. And then this would
be something that we could review at a meeting. So we could look at all the pages up for deletion
and say yeah that's gone, that's gone, that's gone and then that next week we could take
them out. And similarly for update, again there's a bunch of links I found that they
need updating and we can decide whether they're worth updating or maybe they get moved over
deletion. But that's sort of my process for how we would use FixMeUpdate and deletion,
is it's not so much taking the category and putting it on the page, but using the content
in that category and dropping a link on that page so we can go to one spot and see what's
currently up there. And it's all available through wiki.
>> Right. I like that just because it doesn't introduce category structure,
for an mysterious category structure.
You're not actually making the pages that you want to edit subcategories of the FixMe category.
Nope, I'm just using the category structure for links and attaching it to this
this particular WikiHints page, which is a category, but that way I can play with
categories and not get sidetracked with pages on the WikiHints page.
But that's what I've done. So that might be something in the future that we start looking at.
It might be, I'm sure it would be worth a video to educate people if you see a page you should
be deleted or updated. Click on this, drop it in this page and then we'll be looking at those pages.
The other way to do it is just put information into the talk but I don't think that's as effective
is this because it gives us one place that we're looking for the attention.
Well there is, well I assume you have in the contributing to, right?
Contributing to?
That's where you need to know that there are these special categories that you use.
Yes, absolutely. It's something that I absolutely would put in contributing to.
- Yeah, so it'll be somewhere on this page, for sure.
Yeah, yeah.
But again, I'm not relying on people
to go through all of this.
Probably put it up near the top, I think, yeah.
- Well, I think mention it as a safety thing.
So if you're a new contributor
and you're not sure whether a page should be deleted,
this saves you the trouble of you having to decide.
- I think there's a strong argument
to put it into this page too,
which is I think a new user page.
- Yeah.
- Yeah, mission in both places, yeah.
- Yeah.
Anyway, that's about all I got.
And we're at our hour, so.
- Great.
- We can put more information to the next meeting
if you wanna throw something up now,
or just email me and we can throw it in the agenda.
Well, one thing is, Ed, when you're at a good plateau with what you're working on,
I'd love for you to be in a NYC JUG meeting and show it.
You know, I intended to come to the last one, and I slept through it.
No offense.
Oh, no, well, you're in a faraway land, so...
I am in a faraway land.
I would be very happy to show it in the next one, I would even undertake to stay up.
Okay, well, I mean, there's no rush on it.
at some point where you're comfortable with it.
- Oh, I'm happy to show it anytime.
I'm really at a stage where I'm interested
in getting as much feedback as I can.
- Okay, well, Jay, why don't I pencil you in
for next meeting then?
It should be May 11th.
- That'd be great.
Marvelous, I'll be there.
- And I think they started a little earlier than this.
They're 3.30, aren't they?
Well, my time, I guess, but 11.30 Ed's time.
- I start half an hour before this one started.
- Yeah. - I started at 6.30.
Yeah.
Yeah. So that'll be, I'll be 30 minutes more conscious.
But you'll be in Dublin.
Whatever that means.
I can even move it up. It's moving up half an hour to help.
It's fine. It really is.
I've looked back at my body clock. It really doesn't matter.
Would you have people from Australia? And if I move it up,
it starts getting early for them.
Right, totally cool.
Okay, great.
Bob, Earl, Devon, thank you so much.
Take care everybody.
My cat says bye.
Bye bye.
(laughs)