The end of XTalk / LiveCode for me

[MichaelBluejay]

After nearly 40 years, I'm done with XTalk. In the beginning I loved the ability to quickly make real apps with HyperCard, but today I can make web-based apps with Javascript + SQL, which is easy, cheap, and gives way better performance.

When HyperCard came out in 1987 I was blown away. I actually took a trip to Dallas to hear Bill Atkinson (its creator) speak about it.

Circa 1995 I made a note-taking app with it that I called ScatterBrains. I used an interface that I saw in another obscure app: a list on the left hand side, which pulled in content into the main pane on the right. Prior to this, all apps required the user to open a new document/window to get new data. If you had 20 documents/notes, then you had 20 windows, ugh. Nowadays that kind of list+content pane interface is ubiquitous, but I saw its value early on and adopted it years before it became commonplace.

Some years later I ported ScatterBrains to SuperCard, probably because HyperCard probably wouldn't run on a newer MacOS (or maybe because I wanted color), then later, to Revolution, which became LiveCode.

I just upgraded my computer from a 2013-era iMac to a 2022 Mac Studio, and LiveCode 9 won't run on my new Mac. Upgrading LiveCode costs $440 a year, which is not worth it for me to run some homegrown apps. That pushed me to port ScatterBrains to the web with Javascript + SQL, which I just finished. It's a much better result than I could have achieved with LiveCode, if for no other reason that the LiveCode engine payload is a whopping 8-16 Mb. In 2019, a simple Hello World script required a staggering 28 Mb.

Incidentally, I know there are a gazillion other note-taking apps I could use, but none has the features I want (especially hanging indents), so I rolled my own. It was also fairly easy for me to export my LC data to my new web-based app, while a third-party app would usually require copy-and-pasting a thousand notes.

I also wrote a double-entry accounting app in LiveCode a few years ago after the commercial app I paid for mangled my data and I didn't trust it any more, but I was never happy with my LC-based app. The DataGrid syntax was so insane it really killed the whole idea of XTalk being easy to program, and the performance (e.g., scrolling a table) was just glacial. I'll be porting the accounting app to the web too, using a Javascript + SQL solution, and expect to get way better performance than I got with LC/Data Grid.

LiveCode's advantage is the ability to make apps easily. Is there any other these days? Since I'm content to go web-based for my projects instead of making apps, for myself, I no longer need XTalk/LiveCode.

[keliko]

Really interesting to read about your journey from HyperCard to web-based apps. As for me, I still use LiveCode because, as a solo developer, I appreciate how readable the code is and how quickly I can build something useful.
I also develop with C, Go, PureBasic, and FreeBasic — but I’m not interested in Xojo. When it comes to selling software, I still prefer using LiveCode.
I’ve stopped at LiveCode 10 and now use the Community edition. But if LiveCode Create is officially released (not just a developer preview), I’ll definitely buy it and I’m fine with the 5% revenue share.

So yeah... even though LiveCode drives me crazy sometimes, I still love it.

[richmond62]

I certainly would not consider buying 'Create' until I have had a decent amount of time to play with some sort of trial version.

[richmond62]

LiveCode 9 won't run on my new Mac.

That's odd: I run LC 9.6.3 on Mac OS Sequoia without any obvious problems.

[MichaelBluejay]

LC 9.6.3 on MacOS 15.5 starts to load, then when it says "Loading tools..." it quits unexpectedly.

But at this point, even if I could get it to run, I wouldn't use it any more. I'm nearly finished writing my new accounting app, and it's SO much easier to use HTML tables + Javascript vs. LC's DataGrid.

[stam]

I’ve stopped at LiveCode 10 and now use the Community edition. But if LiveCode Create is officially released (not just a developer preview), I’ll definitely buy it and I’m fine with the 5% revenue share.

While LC Create can charitably be called beta (it’s really still alpha as features are still being developed and added), you can “switch to classic” which basically gives you LC 10 but with all the new widgets that come with LCC, and the licence is noticeably cheaper than v10. I use that now and it works well.

[richmond62]

I wonder how many people are going to pay $100 for alpha software sight-unseen?

[stam]

I wonder how many people are going to pay $100 for alpha software sight-unseen?

Rather than an alpha version, think of it as an improved version of LC10 + the ability for early access to LCC. And as of the latest preview of LCC you can use that to deploy apps. Still, the Classic mode works very well, and you get access to all the new widgets.

Of course this will mainly be attractive to those already familiar with LC… attracting users new to the platform as paying customers is why more changes are needed - either have an actual decent trial version, introduce micropayments so that committing isn’t that scary, or just release v10 for Linux (not LCC) for free and possibly OSS (but limited with the ability to build to Linux only).

To me, the last option seems attractive because you get the full version to play with as a learner/tester and in reality commercial projects will be released on Windows/MacOS/Mobile, so if used commercially they can upgrade to paid version. And it’s not like they don’t already have an LC10 Linux build already…

[MichaelBluejay]

As I'm continuing my accounting sw development in HTML/JavaScript, I peeked into my LC code for my LC-based accounting app, to copy an algorithm to port, and I'm remembering something that really slowed me down in LC development: the code windows scrolls super slow. I'm using an older computer, but I shouldn't need an M-series computer just to scroll a code window. For HTML/JS, I'm using the BBedit text editor which scrolls super fast, and it does syntax highlighting just like LC.

What is your experience with the LC code editor like? Have they made speed improvements since the version I'm used (9.6.3)?

[stam]

What is your experience with the LC code editor like? Have they made speed improvements since the version I'm used (9.6.3)?

Well I'm on an M2 Pro so can’t really say if there is improvement. Performance is fine.
If editing long scripts I will often use VS Code as it has many conveniences apart from a higher frame rate, and there is a formatter/linter/syntax checker for LiveCode. But I don’t feel the need to avoid using the script editor if that helps…

[MichaelBluejay]

As I'm developing with HTML+JS, I found one advantage of LiveCode: It's live. Meaning, I don't have to reload the page to see the results. I've done that some hundred times now with HTML+JS.

[FourthWorld]

As I'm developing with HTML+JS, I found one advantage of LiveCode: It's live. Meaning, I don't have to reload the page to see the results. I've done that some hundred times now with HTML+JS.

That's one of two of the most significant advantages xTalks have (the other being GUI objects as integral parts of the language).

I can't recall any xTalk vendor really playing up that advantage. It's not small.