flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Harui <aha...@adobe.com>
Subject Re: [FlexJS] Presenting at the HTML5 Dev Conference
Date Thu, 11 Sep 2014 20:50:41 GMT
Good questions.  My thoughts:

On 9/11/14 1:03 PM, "Harbs" <harbs.lists@gmail.com> wrote:

>Some additional thoughts: (don¹t know if you can fit this in ‹ and I
>don¹t have the answers to all these points)
>FlexJS vs. Dart vs. Haxe. It seems that all three of these have points
>for strong typing.
I definitely try to make sure it doesn't sound like we thought of the idea
of using a higher-level language for generating JS.  I include GWT and TS
in the list.  I try not to set it up as a competition, but rather, a
validation that strong-typing is important in building big things.  We can
all go down to the local lumber yard and hack together a dog house out of
wood and nails, but folks who build skyscrapers use different tools and
techniques.  No high-rise buildings are made of wood and nails.

>Is there an advantage of ActionScript over those? (not sure)
For dog houses, and probably one and two-story houses, there isn't any
advantage.  But eventually, as your application grows, and the team that
is building it grows, the requirement most of these other languages have
of needing all of the code in once place in order to type-check it becomes
a practical limitation.  Yes, these languages let you promise that some
external thing will have a particular contract, but the ActionScript VM
will verify the contract at the point of integration.  The Dart VM might,
I haven't checked, but the JS runtimes certainly won't.

>Is there an advantage of MXML? (probably yes)
I think HTML proved the advantages of declarative markup.

>Maybe some thoughts on the MXML approach vs. HTML/CSS approach?
>advantages and challenges, etc.
I could certainly see someone creating a way to generate HTML from MXML,
so I don't see it as an either/or.  We are just starting with MXML because
we have it, and it is extensible.  I'm still not clear what is happening
with Web Components and HTML extensibility.  But for sure, we have a
better chance of working in IE9.

>What about TypeScript? Do we have advantages over that? What?
IMO, having a markup language gives us advantages over TS.

While it would be awesome if FlexJS became the most popular way to write
HTML/JS/CSS apps, I will be more than happy just to be popular among
enterprise or some other smaller market of folks building really big apps.
 The notion of having markup to basically have a diagram or schematic of
your components, and better connectors in AS and the VM to ensure those
components are connected properly, an IDE that understands all of that, a
debugger that also understands that, and scalability up to really large
applications whose source code cannot be gathered all in one place, and
scaleability down to mobile apps so you can share business logic with the
mobile companion apps should make us attractive to those who are thinking
ahead.  Once the requirement comes out to put a fifth floor on your wooden
house, or add an addition to your house built in another country, you will
find yourself worrying about a lot more things if you aren't using FlexJS.


View raw message