Open Letter to the managment of LiveCode

[Mag]

The policy of the externals does not work.

After spending quite a bit 'of time developing several projects with LiveCode for my company, I finally did a pretty good idea on the platform. The language is fantastic apps generated are stable everything there works very well and what does not work is fixed in a short time (when possible).

The weak point of LiveCode? Confidence in extenals. Unfortunately, in my experience, this is the worst weakness of the company policy of LiveCode. There are fundamental issues that LiveCode that you can not get without resorting to some third-party software. If you're lucky, everything goes well, if you're unlucky, you'll see, after you worked for months, that the developer of the external does not update the external, that he has no time to make it work or that he is working on something else. And you remain without a key feature of your app.

There are some things that are missing in LiveCode to be considered a mature and complete software for the development yet, I ask the management to not rely on external developers for these features, I ask the company to develop them internally.

This is the only way to ensure a future for the projects of the LiveCode developers.

Thanks,
Ivan Gobbo

[FourthWorld]

Which external were you having a problem with?

[Mag]

Hi Richard, I'm trying to make a general discussion here. From the very beginning when I started developing with LiveCode I realized that it depended for basic functionality (such as sharing an image on a mobile device, just to give an example) from third-party software.

I'm asking that RunRev, implements the basic functionalities directly in LiveCode.

[lpmathiasen]

Hello.
I am also using LC Comm for small project in a manufacturing environment. I agree that externals are a weak point in LC. With people costume to apps, in iOS, in Android, in Google Apps and in Sharepoint webparts, Linux repository andTYPO3 repository one would expect a more seamless integration of externals, or plugins as LC provides. Why not have an online repository we can select and install commercial, opensource, paid and free externals/apps/templates/snip-lets? The framework does call for a seamless integration. I will go as far as to proclaim that LC's success depend on this. LC has an opportunity to be a de-facto framework on every platform out there, but because of lack of seamless use - and following irritations - I think LC is missing a lot of business.

I try to introduce LC as an OS in organisations. I work a lot with BPM and BPMS, in Sharepoint, MS Dynamics AX and NAV, in LEAN manufacturing environments and Six Sigma quality systems. LC could be uses extensively to data collections, to make Value Stream Mapping and Data Warehousing... LC has the capabilities, but lack of best practice, examples, community and most importantly too complex install and external app administration, it seems easier to invest in a expensive iBPMS.
I see LC as the right tool for the job because we can develop on a Company OS like we did develop Linux in the past. LC could be the new universal script framework..

I think it is a matter of lack of maturity in the development strategy of LC.

Kindly
Lars Mathiasen
Senior Technology & Business Developer
www-lpmathiasen-com
www-bj-gear-com

[FourthWorld]

Why not have an online repository we can select and install commercial, opensource, paid and free externals/apps/templates/snip-lets?

Commercial add-ons are here (livecode.com -> Store -> Extensions):
https://livecode.com/store/marketplace/

Freely shared examples, tutorials, templates, and script "snip-lets" are made available through RevOnline - in the IDE click "User Samples" from the toolbar, also available on the Web here (livecode.com -> Community -> LiveCode Share Portal):
http://livecodeshare.runrev.com/

[benjibeaumont]

Hi Ivan,

Thanks for your feedback. I must say that I agree with you to a certain extent, LiveCode's success does rely on the ability of developers to create extensions and share them. This hinges on us providing a simple way to share the libraries, code, externals etc you've written. We have plans to improve this in 2014 so watch this space.

In terms of externals, I also agree that at the moment this isn't as easy as it could be. We have been discussing ways of improving the experience for external developers and your comments have raised the importance of this further in my mind. It is important that we also improve the scope and reach of externals which we are looking at also, for example Android.

Richard has helpfully posted links to our marketplace.

Warm regards,

Ben

[PaulDaMacMan]

One of the things that made HC so great 'back in the day' was the sheer amount of externals freely available on UMICH archives and AOLs HC forums.
I have yet to write a modern LC external (I wrote a few for HC in FutureBasic) but it seems the concept hasn't changed much. You write a code library in a lower level language and then wrap it in 'glue' code that handles the passing of variables from the engine. It still seems like a big time waster to me, particularly when the code library in the lower level language is already written and only needs to be wrapped in order to use it with LC.
Instead of externals / glue, perhaps it would be better to come up with a system of bridges so one could call functions in a C or ObjC lib or framework directly? There are tons of those around. I was hoping that the 'open language' initiative would've included a step or two in this direction (LiveCode scripting of definitions of LC syntax for accessing external libraries).

Just my 0.02

[Mag]

Thanks to all those who have spoken. I'm really happy that RunRev looks carefully at the Externals functionality. I think it's really an important thing for the platform.

The problem I'm talking about, is that LiveCode needs plug-ins to do really basic things, such as share on Facebook with the UI Activity View Controller on iOS. In my opinion, this kind of basic functions can't be delegated to third-party plug-ins.

[FourthWorld]

...LiveCode needs plug-ins to do really basic things, such as share on Facebook with the UI Activity View Controller on iOS. In my opinion, this kind of basic functions can't be delegated to third-party plug-ins.

It would be nice indeed if LiveCode had built-in support for everything we might need, but this one may be more challenging than it appears at first glance.

The iOS dev page describes UI Activity View Controller in very general terms:

The UIActivity class is an abstract class that you subclass in order to implement application-specific services. A service takes data that is passed to it, does something to that data, and returns the results.

https://developer.apple.com/library/ios ... rence.html

The API details are as generalized as we might expect for something with such a broad mandate. While it seem useful for the specific Facebook scenario you describe, should RunRev also provide specific hooks for Twitter, Google Plus, LinkedIn, Pinterest, and others? Would it be possible to abstract that sufficiently for all such services? If not, how would the RunRev team choose which ones to support, and which to exclude?

Making matters more difficult, the dev info Facebook themselves provide doesn't mention Apple's UIActivity class at all, describing their own API instead:
https://developers.facebook.com/docs/ios/share

And they provide a slightly different API for the other 75% using Android:
https://developers.facebook.com/docs/android/share

I'm not familiar enough with the implementation details to be able to suggest whether one should be using Apple's API or Facebook's, or whether both are needed.

But with what little I've come across so far it seems we face a challenge with an engine-based implementation at the moment: It risks being either too specific to be worth doing, or too general to be easily implemented.

What have you learned from the APIs that might help the process of arriving at an actionable specification? Maybe we can get some members of the community to explore the idea of adding this into the engine ourselves, now that the source is freely available.

[Mag]

Hi Richard,

thank you for the post. Monte has an external that adds UI Activity View Controller to LiveCode http://mergext.com/home/mergpop/ he also has a version for Android platform.

I think this feature is very important for a mobile app, so important that could be considered a basic feature. In my opinion of course.