Somebody wasn't listening.

All flavors welcome.
Forum rules
Be kind.
Post Reply
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Somebody wasn't listening.

Post by richmond62 »

https://opensource.com/education/13/2/livecode

"I want LiveCode in my classroom. More importantly, I want the tool to be free, and to ship on every single copy of Ubuntu and Fedora that goes out the door. Countless millions of kids have access to computers, but the fact is, the programming environments they have access to are difficult from a teaching and learning perspective." [my emphasis]

"Please don't try and tell me that IDLE is amazing, or that PyGame is great for teaching people how to program: I do it for a living. And don't waste your time telling me how Processing or the Arduino environments are the bee's knees: Java and C are extremely difficult languages to teach to first-time programmers. Scratch is the bomb, but it only takes you so far. LiveCode is an environment that students could graduate to, from Scratch, and go on to do some really amazing things."

But then they never did . . . as I have been banging on for years and years (well, about 20).
https://richmondmathewson.owlstown.net/
FourthWorld
Posts: 280
Joined: Sat Sep 11, 2021 4:37 pm
Contact:

Re: Somebody wasn't listening.

Post by FourthWorld »

Do you have any idea what it takes to add half a GB to the OS disk image download?

I had more than a few conversations about this with the Ubuntu community team back in the day. It's not an easy sell, and for good reason. Distro maintainers need to keep the download as lean as practical.
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Re: Somebody wasn't listening.

Post by richmond62 »

No, I don't, but surely the point is that a version of LiveCode could be present in the software repository of every distro, tweaked in such a way that it would work without caveats on installation.

AND one of the other messages in this article is the one that has been overlooked, and comes in 2 parts:

1. It IS the 'missing link' between SCRATCH-like setups and the steep learnig curve brigade like Python, C#, and so on.

2. Grab a kid at 7-8-9 and you've got them for life.
https://richmondmathewson.owlstown.net/
User avatar
OpenXTalkPaul
Posts: 1486
Joined: Sat Sep 11, 2021 4:19 pm
Contact:

Re: Somebody wasn't listening.

Post by OpenXTalkPaul »

FourthWorld wrote: Mon May 02, 2022 8:44 pm Do you have any idea what it takes to add half a GB to the OS disk image download?

I had more than a few conversations about this with the Ubuntu community team back in the day. It's not an easy sell, and for good reason. Distro maintainers need to keep the download as lean as practical.
It's not back in the day, a half a gig is certainly not as much as it used to be, I've downloaded plenty of 3-4 GB distros, and regularly pull in hundreds of megs of updates, all 10000x faster than back in the day.
But there's two other things to consider here:
1) Why is it that large, does it really need to be? Can some of the fat be trimmed out?
2) What Richmond is getting at, it should at least show up on Ubuntu's software repos, and/or ANY of the other Linux software repos would be a good start.
Nowadays there's lots more Software distribution methods available on Linux, such as "Snaps" or my favorite, AppImage(s).
Even if they had just made it a regular .deb or .rpm packages (the way MaxV did) it might help with dependency resolution, instead they have a proprietary installer GUI app, that has an embedded user registration stack, and then installs a redundant copy of itself into the IDE's folder for some reason wasting about half a gig of disk space.
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Re: Somebody wasn't listening.

Post by richmond62 »

Those software installer apps all seem dysfunctional.

With Xubuntu I stick to .deb files and app get.

And your app-image seems the best way to do things.
https://richmondmathewson.owlstown.net/
FourthWorld
Posts: 280
Joined: Sat Sep 11, 2021 4:37 pm
Contact:

Re: Somebody wasn't listening.

Post by FourthWorld »

OpenXTalkPaul wrote: Tue May 03, 2022 1:30 am 2) What Richmond is getting at, it should at least show up on Ubuntu's software repos, and/or ANY of the other Linux software repos would be a good start.
What Richmond wrote was (in red, presumably because he felt it was important): "to ship on every single copy of Ubuntu and Fedora".

I was addressing only what he wrote.

If we skip what he wrote to transform this to a different discussion about packaging in general, yes, there are many packaging formats available that better meet Linux users' expectations for executable downloads.

But the extra steps to produce the packaging format(s) is just the first step.

I don't know the details for Fedora's repositories, but packages for the Ubuntu repos (and thus their GUI app store) should have a PPA set up for them. Those requirements are not as onerous as Apple's, but will mean extra steps in deployment to keep the package current, something to keep in mind when budgeting time for the project.
User avatar
OpenXTalkPaul
Posts: 1486
Joined: Sat Sep 11, 2021 4:19 pm
Contact:

Re: Somebody wasn't listening.

Post by OpenXTalkPaul »

FourthWorld wrote: Tue May 03, 2022 5:35 pm
OpenXTalkPaul wrote: Tue May 03, 2022 1:30 am 2) What Richmond is getting at, it should at least show up on Ubuntu's software repos, and/or ANY of the other Linux software repos would be a good start.
What Richmond wrote was (in red, presumably because he felt it was important): "to ship on every single copy of Ubuntu and Fedora".

I was addressing only what he wrote.
OK cool.
But I could also see some integrated install on some distro, maybe not a big one but perhaps one such as helloSystem or ravynOS (airyxOS) where I could see an xTalk (not necessarily the current OXT, could be OpenXION or something else) being a PERFECT fit as an open-source alternative to AppleScript for inter-app / automation stuff. I've played around a bit with something in a similar vein called "Hammerspoon", which can be used for automating things on macOS but using Lua instead of AppleScript or JSX. https://www.hammerspoon.org/docs/index.html

Could maybe get into something like Distro "Spin"(Fedora) perhaps that concentrates on early computer learning or something like that.
FourthWorld wrote: Tue May 03, 2022 5:35 pm I don't know the details for Fedora's repositories, but packages for the Ubuntu repos (and thus their GUI app store) should have a PPA set up for them. Those requirements are not as onerous as Apple's, but will mean extra steps in deployment to keep the package current, something to keep in mind when budgeting time for the project.
Personally, I love the one app = one file paradigm of .AppImage, which I believe was developed by the same guy that's working on a FOSS FreeBSD macOS clone called helloSystem. AppImage obviously inspired by macOS/NeXT bundles (and he's stated as much).

The Buntu's...if not a main repo of some major distro then at least a PPA would certainly be something to consider, as well as any other alternative packaging, Snap for instance (as well as distribution methods such as distributed files hosting with BitTorrent).
I wonder was distribution in Repos not considered by the ex-Mothership? Was it deemed too time consuming? What?

I really think there's been some huge missed opportunity in the Linux desktop apps sphere. Imagine if people started pumping out useful little Linux tools, or little niché tools, Some GUIs for CLIs, seems like that could be a good xTalk promotional mechanism because what's better than useful working example that you can also take apart while it's running to get to know the language in action while also being fully customizable. Example: I could not find a decent editor for freedesktop.org .desktop files, seems like something pretty important for DEs and something that people would want/need to use. I really want Linux to be a full-on first class distro of the OXT IDE. I'm planning to add a library of Linux-specific handlers to OXT soon that will hopefully grow over time. I'm also finally updating/testing some of Builder Extensions for Linux(64). I will likely add in LIBHIDAPI (for using USB / Bluetooth devices like joysticks) as well as Hunspell, which I tested on Linux this afternoon and it worked fine (I've also built a basic xTalk-HunSpell dictionary file, which might be useful implementing an syntax autocomplete)
FourthWorld
Posts: 280
Joined: Sat Sep 11, 2021 4:37 pm
Contact:

Re: Somebody wasn't listening.

Post by FourthWorld »

OpenXTalkPaul wrote: Tue May 03, 2022 10:34 pm Personally, I love the one app = one file paradigm of .AppImage, which I believe was developed by the same guy that's working on a FOSS FreeBSD macOS clone called helloSystem. AppImage obviously inspired by macOS/NeXT bundles (and he's stated as much).
First time I saw AppImage was with Nextcloud. Nice implementation. I hope it solves the Flatpak vs Snap war. I'm tired of that war; it's going nowhere, and slowing things down. AppImage is tidy enough to replace both well.
I wonder was distribution in Repos not considered by the ex-Mothership? Was it deemed too time consuming? What?
Considered, but difficult to justify the time for. Their spend on Linux has always been favorably disproportionate to revenue, and extra steps for the platform that brings in the least amount of money to keep the lights on was easily a low priority. They would have liked to do it, but it would have some at a time cost to other activities that produce revenue.
I really think there's been some huge missed opportunity in the Linux desktop apps sphere. Imagine if people started pumping out useful little Linux tools, or little niché tools, Some GUIs for CLIs, seems like that could be a good xTalk promotional mechanism because what's better than useful working example that you can also take apart while it's running to get to know the language in action while also being fully customizable.
I used to be extremely enthusiastic about the growth potential for consumer apps on Linux. But over time I've mellowed. Computers generally come with an OS preinstalled, and for most folks that's good enough. It doesn't make sense to the average person to go to the trouble to replace the OS their computer was designed for with something they downloaded from the Internet. The strongest benefit Linux offers over the Mac+Win duopoly is flexibility. But 80% of users never even change Prefs defaults in app, so the freedom to tinker with an OS just doesn't speak to them.

I love it, of course, and still prefer just about any Ubuntu box to my new M1 Mac, sweet as it is; Linux feels like an OS made by friends. But I'm a tech enthusiast. My tastes aren't reflective of the buying public, for whom source availability is completely meaningless, and app availability outside of devops is slimmer than they had with the OS that came with their computer.
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Re: Somebody wasn't listening.

Post by richmond62 »

Just for the sake of argument:

LC Community 7.1.4 comes to 146 MB.

LC Community 8.1.10 comes to 277 MB.

A modded build of 7.1.4 called "OpenXTalk" basic could be included with distro install images
with an update screen to install a "9" version that could then be downloaded and installed.
But I'm a tech enthusiast.
Indeed.

I must say I do like those apps packaged in a Java wrapper so, as long as you have Java installed on your
operating system (whatever it may be), you can run them.
https://richmondmathewson.owlstown.net/
User avatar
OpenXTalkPaul
Posts: 1486
Joined: Sat Sep 11, 2021 4:19 pm
Contact:

Re: Somebody wasn't listening.

Post by OpenXTalkPaul »

FourthWorld wrote: Wed May 04, 2022 5:52 am I used to be extremely enthusiastic about the growth potential for consumer apps on Linux. But over time I've mellowed. Computers generally come with an OS preinstalled, and for most folks that's good enough. It doesn't make sense to the average person to go to the trouble to replace the OS their computer was designed for with something they downloaded from the Internet. The strongest benefit Linux offers over the Mac+Win duopoly is flexibility. But 80% of users never even change Prefs defaults in app, so the freedom to tinker with an OS just doesn't speak to them.

I love it, of course, and still prefer just about any Ubuntu box to my new M1 Mac, sweet as it is; Linux feels like an OS made by friends. But I'm a tech enthusiast. My tastes aren't reflective of the buying public, for whom source availability is completely meaningless, and app availability outside of devops is slimmer than they had with the OS that came with their computer.
Sure as a business you have to prioritize actual business, they have creditors expecting payments, and employees expecting pay checks, and their customers with their own business interests... I get that part. However if your business is built around a still-niché programming language, if you ever wanted to actually grow that business, you'd want to get that to be a lot less of a niché language, I would think. Get it into the hands of the tech enthusiasts, tinkerers, coders, etc. many of whom use Linux OS and I'd wager that percentage of tinkerers is significantly higher for Linux users than it is for macOS and Windows users.

Python has been preinstalled in macOS X since forever, I don't know how much it help Pythons popularity, but I'm sure it didn't hurt. LiveCode does not even show up in any form in a search for LiveCode on Apple's AppStore. If you're after mac users that are looking for software you'd probably at least want the manual to be on there as an eBook. Or it could even just be demo app, like a slide presentation, perhaps with some sample cards that do things, SOMETHING, just get it on there so regular people can actually find it!

Then there's MS Windows store... That platform I believe is still 80% or more of all desktop computers. So if that was important... ?

IMO they still need to seriously, heavily invest in marketing. I mean if I was running a business, and if was convinced that I had a product to sell, that people would want, at a price where people would surely want to buy it, I'd sure as hell be paying out some significant monies to Facebook, Google, etc. so that anyone that searches for anything remotely "easy coding", or "low-code" or "natural language" or "English-like programming" related, would see my ad in their feed for days... but what do I know, I'm not running a business. Maybe some of these new people they hired will help fix that.

It seems they are finally concentrating on the IDE (why didn't anyone think of that!) somewhat, my guess is that's motivated by WebASM / running inside browser stuff, maybe they'll do a mobile or browser based IDE? That would be cool. As has been discussed here in the past, for desktop apps I'm personally not a fan of the single window interface thing, however on mobile / web you are basically limited to that, so working on single-window would make sense if making mobile or web app IDE.

Personally, I see OXT being a FOSS project that is no longer tethered to commercial business interest as a big plus.
I could spend all my time working on the Linux side of OXT, for no other reward than my own satisfaction, if I want to. I am become somewhat of a regular Linux (and FreeBSD) users myself and eventually I plan to ditch macOS entirely when I no longer need it for my own business interests (so probably not until after I retire in maybe 18yrs).
User avatar
OpenXTalkPaul
Posts: 1486
Joined: Sat Sep 11, 2021 4:19 pm
Contact:

Re: Somebody wasn't listening.

Post by OpenXTalkPaul »

richmond62 wrote: Wed May 04, 2022 6:42 am Just for the sake of argument:

LC Community 7.1.4 comes to 146 MB.

LC Community 8.1.10 comes to 277 MB.

A modded build of 7.1.4 called "OpenXTalk" basic could be included with distro install images
with an update screen to install a "9" version that could then be downloaded and installed.
But I'm a tech enthusiast.
Indeed.

I must say I do like those apps packaged in a Java wrapper so, as long as you have Java installed on your
operating system (whatever it may be), you can run them.
Someday I may take a close look at what can be back ported. v9.x can't run on anything older than macOS 10.9, and PowerPC support was dropped sometime around v6.5+. Linux RPi ARM builds were only during v6.x at v7.x dev cycle (regardless of any false advertising). I like both; old-Macs and Raspberry Pi platforms. I'd love to at least get v9 built for RPi ... I no longer have one but I would get one again, they're cool and fun and can run on next to nothing electricity wise, and two version of LCC will run on it so that's a plus.
FourthWorld
Posts: 280
Joined: Sat Sep 11, 2021 4:37 pm
Contact:

Re: Somebody wasn't listening.

Post by FourthWorld »

richmond62 wrote: Wed May 04, 2022 6:42 am A modded build of 7.1.4 called "OpenXTalk" basic could be included with distro install images
with an update screen to install a "9" version that could then be downloaded and installed.
What I bring here is not theoretical. I've had these discussions with key members of the Ubuntu team.

There's a lot of stuff that goes into a distro. Everyone wants their package preinstalled. A GUI IDE for a language no one's heard of that has at least a dozen broken GUI elements on Linux is a hard sell.

Qt is a much better fit for that audience. And even as deeply as some of the core Qt components are integrated into Ubuntu, they don't include a GUI IDE for it in the distro.

Such things are best put into PPA's in a Snap format where they can be downloaded by those interested (sure, I may prefer other packaging formats, I'm just reporting what gets attention at Canonical). If it becomes a runaway hit the argument for inclusion takes care of itself. And if it doesn't the argument isn't worth making.

You don't just walk up to an OS vendor and say, "Here's my broken app, put it in." Even if it wasn't broken there are a lot of different considerations about what gets included. Every packaging decision Canonical makes affects 50 million users. This is not a small thing to ask. I spent years pursuing it, even covering my costs to travel to Ubuntu Summits, ultimately giving up because I see little hope of any LC-based engine getting the engineering resources to deliver a truly first-class experience on Linux.

The outside world says, "Need desktop? Use Python and Qt. Everything else? JavaScript." With Node.js for servers, React Native for mobile, and straight JS already preinstalled in every device that has a GUI via browsers, it's hard to make a strong case for any also-ran.

And it's even harder when the also-ran has so many Linux issues.
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Re: Somebody wasn't listening.

Post by richmond62 »

has at least a dozen broken GUI elements on Linux
1. Could you possibly list those? As beyond the messageBox I have no idea.

2. Presumably the company we know and love are sorting those out in their commercial version.

(I haven't quite decided if #2 is a sarcastic dig or not.)
https://richmondmathewson.owlstown.net/
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Re: Somebody wasn't listening.

Post by richmond62 »

What was sidestepped in my original post was this:

"LiveCode is an environment that students could graduate to, from Scratch, and go on to do some really amazing things." (my emphasis)

where I would change could for should.

And, I would tend to just let learners 'fool around' with Turtle Graphics a bit, rather than get hypnotised by Scratch; and then more right-along
to xTalk, without 'passing GO'.
https://richmondmathewson.owlstown.net/
FourthWorld
Posts: 280
Joined: Sat Sep 11, 2021 4:37 pm
Contact:

Re: Somebody wasn't listening.

Post by FourthWorld »

richmond62 wrote: Sun May 22, 2022 10:40 am "LiveCode is an environment that students could graduate to, from Scratch, and go on to do some really amazing things."
That article was written while LC was running Kickstarter campaign to go open source. Since OpenSource.com only discusses FOSS, that article couldn't be written today.

LC's open source days are behind it. If there's to be a free and open xTalk that could appeal to educators, it falls on FOSS folk like the OpenXTalk community to do it.
User avatar
richmond62
Posts: 2621
Joined: Sun Sep 12, 2021 11:03 am
Location: Bulgaria
Contact:

Re: Somebody wasn't listening.

Post by richmond62 »

There is nothing intrinsically wrong with using an open source version of LC from the archive.

The ONLY real snag is parents who ask why one is using something that isn't "current".
https://richmondmathewson.owlstown.net/
Post Reply

Who is online

Users browsing this forum: Ahrefs [Bot], OpenXTalkPaul and 5 guests