Well, it seems to me that Typed Arrays is the right way to go about 
implementing ByteArray in JS.[1]

IIUC, mimicking ByteArray is the hardest part of AMF.

Of course, this means a minimum of IE10, but that might be OK…

[1]https://developer.mozilla.org/en-US/docs/Web/JavaScript/Typed_arrays

On Dec 1, 2015, at 8:25 PM, Alex Harui <aha...@adobe.com> wrote:

> 
> 
> On 12/1/15, 10:18 AM, "Christofer Dutz" <christofer.d...@c-ware.de> wrote:
> 
>> Well actually it sounded like: We had AMF support, but it didn't perfrom
>> as well so we threw it away ...
>> 
> 
> No, nobody I know has worked on it.  I just have concerns about how well
> it will work.  Of two implementations I found on the web, one had some
> ties to GPL so I stopped looking, and the other was using getCharCode to
> implement ByteArray.  Depending on how folks want to deal with older
> browsers, we could try using Typed Arrays in the JS side.
> 
> Other than cyclic object graphs, I don't understand why JSON and XML can't
> have decoders that convert to/from typed objects.  That's how the SOAP
> code works in regular Flex today.  You might need custom serialization for
> cycles (at least in the early rounds).
> 
> I understand that AMF had lower bandwidth, but I've heard that JSON can be
> compressed over the wire.
> 
> -Alex
> 
>> 
>> 
>> -----Ursprüngliche Nachricht-----
>> Von: Harbs [mailto:harbs.li...@gmail.com]
>> Gesendet: Dienstag, 1. Dezember 2015 18:56
>> An: dev@flex.apache.org
>> Betreff: Re: [FLEXJS] AMF (was Re: AW: lib sprite flexjs,add graphics.as
>> (canvas))
>> 
>> Instead of talking about throwing it away, why not discuss what is needed
>> to implement AMF? It sounds to me like a great goal.
>> 
>> I personally have never used AMF, so I have no idea.
>> 
>> Harbs
>> 
>> On Dec 1, 2015, at 7:47 PM, Christofer Dutz <christofer.d...@c-ware.de>
>> wrote:
>> 
>>> There is no need for a amf to json converter. You would just use a
>>> different serializer instead.
>>> 
>>> What I'm worried about is that amf is way more powerful than json. Let
>>> me name some benefits:
>>> - strongly typed
>>> - able to serialize cyclic object graphs
>>> - uses way less bandwidth
>>> 
>>> Actually one of my talks at the last apachecon dealt with this entirely.
>>> 
>>> Giving up on this makes one of actionscripts benefits sort of useless.
>>> What's the good of being able to use strong types, if these get lost on
>>> the way from the server to the client?
>>> 
>>> If we throw overboard all the good stuff, we'll be just one if the
>>> other frameworks
>>> 
>>> Chris
>>> 
>>> 
>>> 
>>> Von meinem Samsung Galaxy Smartphone gesendet.
>>> 
>>> 
>>> -------- Ursprüngliche Nachricht --------
>>> Von: Alex Harui <aha...@adobe.com>
>>> Datum: 01.12.2015 18:29 (GMT+01:00)
>>> An: dev@flex.apache.org
>>> Betreff: [FLEXJS] AMF (was Re: AW: lib sprite flexjs,add graphics.as
>>> (canvas))
>>> 
>>> Hi,
>>> 
>>> Renaming this fork of the thread...
>>> 
>>> Well, I have no doubts that AMF is quite popular, but I guess I really
>>> should have asked these questions:
>>> 
>>> 1) If FlexJS didn't exist and you couldn't use FlashPlayer, how would
>>> you get data from the server to client (and back again)?  JSON, XML,
>>> some other thing?  The reason I haven't spent any energy on AMF for
>>> FlexJS is because I think folks would have had to stop using AMF
>>> anyway.  Maybe there is an simple AMF-to-JSON module that folks could
>>> implement on their servers to do the job so you don't have to rewrite
>>> your server code.
>>> 2) If FlexJS did have AMF but its performance was worse than using
>>> JSON, would you still choose the slower AMF implementation?  It isn't
>>> clear that implementing AMF in JS is going to perform as well as
>>> browser-native JSON or flash-native NetConnection.
>>> 3) What is the minimum version of IE that needs to support this?
>>> 
>>> Thanks,
>>> -Alex
>>> 
>>> On 12/1/15, 7:24 AM, "carlos.rov...@gmail.com on behalf of Carlos
>>> Rovira"
>>> <carlos.rov...@gmail.com on behalf of carlos.rov...@codeoscopic.com>
>>> wrote:
>>> 
>>>> Hi Alex,
>>>> 
>>>> AMF is "key" for Flex in IT ecosystem. you could make a Poll and, if
>>>> most of people involved in Flex would fill it, you'll be surprised of
>>>> the amount of AMF people is using to comunicate with server-side.
>>>> 
>>>> So, this means, that for me and many others, the AMF is a requisite,
>>>> (more even that Maven, that already is) to start prototyping and
>>>> working with FlexJS in a day by day basis trying to change the
>>>> traditional Flex 4.x layer for a FlexJS layer.
>>>> 
>>>> HTTPService is a must, and is a good base, but it's used in any IT
>>>> app about 5% of the times. People uses RemoteObject (and some times
>>>> Web Services due to some request) as main RPCs
>>>> 
>>>> For me AMF (and I think for many others) is the final wall to start
>>>> investing time in our IT depts with FlexJS.
>>>> 
>>>> Take into account that there are many server side business logic out
>>>> there (Java, PHP, .NET, Ruby...) thats abstract all the things
>>>> happening in the server from the Flex client, and is exposed to Flex
>>>> through AMF - RemoteObjects. So having AMF in FlexJS seems the
>>>> potential keypoint to start trying to change Flex 4.x for FlexJS,
>>>> since you don't have the need to touch server side services.
>>>> 
>>>> So, is a fact that AMF is key for FlexJS.
>>>> 
>>>> Thanks for asking Alex
>>>> 
>>>> Carlos
>>>> 
>>>> 
>>>> 
>>>> 
>>>> 2015-12-01 15:55 GMT+01:00 Vincent <vinc...@after24.net>:
>>>> 
>>>>> +1,
>>>>> All of our projects use AMF
>>>>> 
>>>>> 
>>>>> Le 01/12/2015 15:52, Christofer Dutz a écrit :
>>>>> 
>>>>>> Cause AMF is so much cooler than JSON ;-)
>>>>>> 
>>>>>> I too would like to see AMF in FlexJS ...
>>>>>> Actually if we drop AMF support there's no need for me to keep
>>>>>> maintaining BlazeDS any longer.
>>>>>> 
>>>>>> Chris
>>>>>> 
>>>>>> ________________________________________
>>>>>> Von: Alex Harui <aha...@adobe.com>
>>>>>> Gesendet: Dienstag, 1. Dezember 2015 15:32
>>>>>> An: dev@flex.apache.org
>>>>>> Betreff: Re: lib sprite flexjs,add graphics.as (canvas)
>>>>>> 
>>>>>> On 12/1/15, 5:22 AM, "carlos.rov...@gmail.com on behalf of Carlos
>>>>>> Rovira"
>>>>>> <carlos.rov...@gmail.com on behalf of
>>>>>> carlos.rov...@codeoscopic.com>
>>>>>> wrote:
>>>>>> 
>>>>>>> I start to think the only big problem is now to get AMF comming to
>>>>>>> FlexJS.
>>>>>>> 
>>>>>>> All frameworks (FlexJS, feathers, ...) out there are very cool,
>>>>>>> but all lacks RPC APIs (RemoteObject, ...)
>>>>>>> 
>>>>>>> And without that is impossible to propose a starter project in a
>>>>>>> company or IT dept with FlexJS...
>>>>>>> 
>>>>>> 
>>>>>> Hi Carlos, why AMF?  FlexJS does have HTTPService.
>>>>>> 
>>>>>> 
>>>>>> -Alex
>>>>>> 
>>>>> 
>>>>> 
>>>> 
>>>> 
>>>> --
>>>> 
>>>> Carlos Rovira
>>>> Director General
>>>> M: +34 607 22 60 05
>>>> http://www.codeoscopic.com
>>>> http://www.avant2.es
>>>> 
>>>> 
>>>> Este mensaje se dirige exclusivamente a su destinatario y puede
>>>> contener información privilegiada o confidencial. Si ha recibido este
>>>> mensaje por error, le rogamos que nos lo comunique inmediatamente por
>>>> esta misma vía y proceda a su destrucción.
>>>> 
>>>> De la vigente Ley Orgánica de Protección de Datos (15/1999), le
>>>> comunicamos que sus datos forman parte de un fichero cuyo responsable
>>>> es CODEOSCOPIC S.A. La finalidad de dicho tratamiento es facilitar la
>>>> prestación del servicio o información solicitados, teniendo usted
>>>> derecho de acceso, rectificación, cancelación y oposición de sus
>>>> datos dirigiéndose a nuestras oficinas c/ Paseo de la Habana 9-11,
>>>> 28036, Madrid con la documentación necesaria.
>>> 
>> 
> 

Reply via email to