Smalltalk › Amber Smalltalk

What would be the best way to build amber into an existing project?

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

6 messages Options

Andy Burnett

What would be the best way to build amber into an existing project?

ᐧ

I am currently playing around with Meteor. If you haven't tried it, they seem to have done some very impressive work. What I particularly like is the way they have pulled together various packages to deal with authentication, database integration, and a host of other things. It seems like an interesting development stack. However, what I really want to do is write the code in Smalltalk.

I could see various ways of doing this, such as:

1. generate the app from within Meteor, write the javascript as Smalltalk, and then compile it to js.

2. as above, but somehow 'requiring' amber within Meteor. They have a coffeescript package, so maybe I could hack that to compile Smalltalk.

3. as above, but make Amber live within the Meteor app, and develop from there. I imagine having a basic Meteor framework, and then populate the app with methods developed from within Amber

4. generate the app within Amber, and then somehow 'require' Meteor.

I am sure this will be messy to start with, but I really love the idea of having something a little like Gemstone as a reactive platform that allows us to develop in Smalltalk.

Any have any experience/thoughts on this?

Cheers

Andy

--
You received this message because you are subscribed to the Google Groups "amber-lang" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

sebastianconcept

Re: What would be the best way to build amber into an existing project?

Meteor is fantastic. Is getting hotter and hotter by the minute.

Casually the idea of embracing Meteor and Amber crossed mi mind.

I see 2 paths:

1. inventing our own reactive stack on top of Amber in front end and an Amber-node server app (using MongoDB for persistence and Redis for model cache + pub/sub communication) <- I actually have something like that but on Pharo

2. embracing Meteor and coming on top of it with an Amber app as you suggested

This is the kind of thing we need to discuss in hangouts.

This community has many brilliant people I’m sure they’ll be awesome.

sebastian

On Apr 15, 2014, at 11:33 AM, Andy Burnett <[hidden email]> wrote:

ᐧ

I am currently playing around with Meteor. If you haven't tried it, they seem to have done some very impressive work. What I particularly like is the way they have pulled together various packages to deal with authentication, database integration, and a host of other things. It seems like an interesting development stack. However, what I really want to do is write the code in Smalltalk.

I could see various ways of doing this, such as:
1. generate the app from within Meteor, write the javascript as Smalltalk, and then compile it to js.

2. as above, but somehow 'requiring' amber within Meteor. They have a coffeescript package, so maybe I could hack that to compile Smalltalk.
3. as above, but make Amber live within the Meteor app, and develop from there. I imagine having a basic Meteor framework, and then populate the app with methods developed from within Amber

4. generate the app within Amber, and then somehow 'require' Meteor.

I am sure this will be messy to start with, but I really love the idea of having something a little like Gemstone as a reactive platform that allows us to develop in Smalltalk.

Any have any experience/thoughts on this?

Cheers
Andy

--
You received this message because you are subscribed to the Google Groups "amber-lang" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

Andy Burnett

Re: What would be the best way to build amber into an existing project?

I like the idea of a hangout. And, I think it would be great to be able to demonstrate something, if only to rule out some approaches. I will try to hack something together.

I have vaguely wondered about building our own reactive stack. We sponsored the Sven to create the websockets classes in Pharo, and I have been itching to use them. However, I hate reinventing the wheel, and the Meteor guys have done a great job in making the real-time behaviour work in a whole range of different cases, i.e. when websockets fail etc.

My ideal situation would be to press a button that creates a bare bones Meteor application, but with Amber included, and then to develop the rest from within Amber. Of course this raises all sorts of interesting questions such as:

How does Amber know about all the various packages that were included in the Meteor app?
How do we deal with building/manipulating templates? Or, could we sidestep the templates altogether, and use a Seaside like approach, whilst still retaining the data binding and live updates?

I have looked - briefly - at the Meteor Colleciton class. They have exposed it so that you can write e.g. Objective C apps that integrate with Meteor. It may be possible to link into Meteor at that level. However, that might mean that we can't take advantage of all the packages that people are creating for Meteor.

Personally, I think it would be great to be able to go to use Amber to insert e.g. a D3 chart, tell the chart object that it is linked to a reactive data source, and just hack about with it via a workspace.

Cheers

Andy

ᐧ

On Tue, Apr 15, 2014 at 11:01 AM, sebastian <[hidden email]> wrote:

Meteor is fantastic. Is getting hotter and hotter by the minute.

Casually the idea of embracing Meteor and Amber crossed mi mind.

I see 2 paths:

1. inventing our own reactive stack on top of Amber in front end and an Amber-node server app (using MongoDB for persistence and Redis for model cache + pub/sub communication) <- I actually have something like that but on Pharo

2. embracing Meteor and coming on top of it with an Amber app as you suggested

This is the kind of thing we need to discuss in hangouts.

This community has many brilliant people I’m sure they’ll be awesome.

sebastian

o/

On Apr 15, 2014, at 11:33 AM, Andy Burnett <[hidden email]> wrote:

ᐧ

I am currently playing around with Meteor. If you haven't tried it, they seem to have done some very impressive work. What I particularly like is the way they have pulled together various packages to deal with authentication, database integration, and a host of other things. It seems like an interesting development stack. However, what I really want to do is write the code in Smalltalk.

I could see various ways of doing this, such as:
1. generate the app from within Meteor, write the javascript as Smalltalk, and then compile it to js.

2. as above, but somehow 'requiring' amber within Meteor. They have a coffeescript package, so maybe I could hack that to compile Smalltalk.
3. as above, but make Amber live within the Meteor app, and develop from there. I imagine having a basic Meteor framework, and then populate the app with methods developed from within Amber

4. generate the app within Amber, and then somehow 'require' Meteor.

I am sure this will be messy to start with, but I really love the idea of having something a little like Gemstone as a reactive platform that allows us to develop in Smalltalk.

Any have any experience/thoughts on this?

Cheers
Andy

--
You received this message because you are subscribed to the Google Groups "amber-lang" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "amber-lang" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.

Darius Clarke

Re: What would be the best way to build amber into an existing project?

+1 to working with Meteor in some way

Tim Mackinnon

Re: What would be the best way to build amber into an existing project?

In reply to this post by Andy Burnett

I was talking to a few guys about MV* libraries in JS - trying to understand what the sweetspot might be used to bootstrap some Amber usage.

Obviously Meteor is mentioned here - but then there was also mention of Angular a few days ago - and then when I read a bit more, I see mention of Ember (and Backbone as a more primitive cousin). Are these all questions around the same thing?

I did note that they are re-writing Angular to fit an AMD style mechanism (I think I understood that correctly?). I know many of the guys I was talking to seemed to prefer Ember.

I have noticed there is Trapped as well - but I think I wanted to understand how the JS libraries handle this, to then understand the Trapped approach.

Tim

On 15 Apr 2014, at 17:02, Andy Burnett <[hidden email]> wrote:

I have vaguely wondered about building our own reactive stack. We sponsored the Sven to create the websockets classes in Pharo, and I have been itching to use them. However, I hate reinventing the wheel, and the Meteor guys have done a great job in making the real-time behaviour work in a whole range of different cases, i.e. when websockets fail etc.

Andy Burnett

Re: What would be the best way to build amber into an existing project?

There is definitely no shortage of libraries to choose from ;-)

My interest in Meteor is driven by two main factors:

They have (or at least they are moving towards) a full development stack. I really don't want to waste time trying to glue a framework, to a database to an authentication mechanism to an email system to ... Basically, being very lazy, I just want to have a large class library, and spend my time sending messages to objects. Safe in the knowledge that persistence is being handled through 'magic'.
Real time is baked in rather than an add-on. This is important for the type of applications I have in mind, but it certainly isn't the case for all apps. However, I think that more and more applications will be expected to be real time, as people become used to the experience.

I don't really know enough about all the other libraries/frameworks in the JS world, to know if there is a better choice. And I am definitely not wedded to Meteor, it's just that it seems very good, and they seem to have enough funding to suggest that it will be around for a while.

As a side note, I think live programming of a real time environment would be an amazing sweet spot for Amber, but I am biased in that view ;-)

Cheers

Andy

ᐧ

On Tue, Apr 15, 2014 at 1:09 PM, Tim Mackinnon <[hidden email]> wrote:

I was talking to a few guys about MV* libraries in JS - trying to understand what the sweetspot might be used to bootstrap some Amber usage.

Obviously Meteor is mentioned here - but then there was also mention of Angular a few days ago - and then when I read a bit more, I see mention of Ember (and Backbone as a more primitive cousin). Are these all questions around the same thing?

I did note that they are re-writing Angular to fit an AMD style mechanism (I think I understood that correctly?). I know many of the guys I was talking to seemed to prefer Ember.

I have noticed there is Trapped as well - but I think I wanted to understand how the JS libraries handle this, to then understand the Trapped approach.

Tim

On 15 Apr 2014, at 17:02, Andy Burnett <[hidden email]> wrote:

I have vaguely wondered about building our own reactive stack. We sponsored the Sven to create the websockets classes in Pharo, and I have been itching to use them. However, I hate reinventing the wheel, and the Meteor guys have done a great job in making the real-time behaviour work in a whole range of different cases, i.e. when websockets fail etc.

--
You received this message because you are subscribed to the Google Groups "amber-lang" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.