On 2010-12-16, at 12:36, Raju Bitter wrote:

> I'm looking for the 10 best reasons to use LZX instead of JavaScript
> when building rich browser and mobile applications.

I can't give you a list, but thanks for starting this discussion.  Hopefully 
others will chime in with their favorite OL feature.

My personal favorite is the (still evolving) presentation type system, which I 
think will be a very important tool.

> Why are we coding in LZX? What was the initial idea behind creating
> the language, and is the initial idea of LZX (XML + JS + XPath ...)
> still valid in times of a web world which seems to be more and more
> influenced by JavaScript? What does the general of adoption of JSON
> instead of XML in web services mean for LZX?

The initial idea, to use standards where standards are good and powerful and to 
extend when the standards lacked, still applies IMO.  The ascendancy of JSON 
over XML is a fine thing (just as XML over SOAP), and we have a path for that 
(through presentation types, as you have seen in my recent mail).

I don't foresee the use of XML for declarative programming being supplanted by 
Javascript, unless someone can come up with something more clever than simply 
transcribing the XML to JSON.

I believe OL is still alone in supporting constraints, mixins, and now 
presentation types, which are power tools that let the power user develop and 
evolve their applications rapidly and efficiently.  These power tools have the 
potential to give you a jump on your competitors, if used correctly.

> The results of this discussion will be summed up in a blog post on the
> LZX language, which I plan to publish in January 2011, and will be
> part of future presentations on OL as part of the KamiJS project. I'm
> specifically interested in what the OL team thinks the language will
> look like in 12-18 months from now. What are your visions for LZX, and
> how are you planning to better promote the language? And what would
> you have to say regarding
> 
> - official support for coding classes in JS (which means: the next
> minor update of OL will not break my existing class code)

Not sure what you are asking here.  We support class-based development in both 
LZX and LZS.  The latter is an extension of JS based on the proposed ES4 
standard (which never came to fruition).  We continue to track ES standards and 
have every intention of supporting our class-based development going forward.  
We hope that ES6 will include the "traits" proposal, because we believe that 
will allow a much more efficient implementation of our classes in ES-based 
platforms.

> - generally less changes in LZX! There are so many changes to the
> language in the past 3 years, that it makes the LZX the most
> "unstable" language to use in all of my IT projects. The fact that the
> Laszlo has full control over the LZX language makes the language a
> target to syntax changes directly related to the next release of
> Laszlo Webtop, which is not good for the language in general, I
> believe.

This is a choice you have to make.  You can pick a language that will never 
change, you could develop in Objective-C or Java, or you can pick a language 
that is evolving to be more powerful.  You always have the choice also of 
sticking with an earlier revision, just as many people are still programming in 
PHP4...

The advantage that OL has is that it is a small community, not bound by a 
standards committee, so can evolve and adapt to be the best development 
language you have access to.

Nevertheless, we do not make changes to the language willy-nilly, and we do not 
make changes just to support Laszlo.  We always request community input when 
making changes to the language.  When we have made changes, we have maintained 
back-compatibility whenever practical.  When we make breaking changes, they 
have almost always been to repair warts in the language that were the source of 
confusion and bugs.

It _is_ true that at the beginning the language was built to support Laszlo.  
Because the language was designed "as it went along", there _are_ places where 
the design is less than ideal.  I don't feel anyone is well-served by 
preserving design flaws.  I don't want to see the language become stagnant.

> - AS3 runtimes and debugging - getting rid of the dependency on the 
> SWF8 runtime to detect errors (the more Flash 10 /10.x APIs are used,
> the less is the approach to test with SWF8 to get a better
> understanding of what's wrong within your code valid - since the app
> won't compile). That's especially true for any SWC libraries you
> include.

Not sure what you are asking here.  We do plan to phase out swf8 support, 
because it seems no longer necessary.  The compile-time error detection of 
swf10 seems more powerful to me than the run-time detection we shoe-horned into 
swf8.  If there is something missing, I would be interested to hear about it.

> But I'm not only interested in the OL teams opinion, but especially
> interested in what OpenLaszlo adopters with at least 6 months in OL
> coding experience think of the language. What do you like about LZX at
> the moment, and what should be improved?
> 
> Any input and comments are very welcome.
> 
> Raju


Reply via email to