Re: [e4-dev] Where to seek feedback on Google Summer of Code idea?

[Vineet Sinha <vineet@xxxxxxxxxxxxx>]

Elias, Thanks for the forward.

Jake,
Great to hear of the project. I would love to offer you any help in having this project done. We have been considering getting GMF working on the web as a great next step, and your work will at the very least be a great prototype for the technologies you might consider.

We should take this discussion offline, but off the top of my head here are a few thoughts you might want to consider:
o You need to figure out what code you want to run on the server. You are trying to provide a lot of capabilities, and having them completely run on the browser might make it very heavy weight.

o You might want to use something like GWT's Java to _javascript_ compiler to getting the GMF, GEF, and Draw2D capabilities on the web faster. Using this approach will perhaps also make the code more maintainable.
o You are planning 8-weeks of work for getting GEF running on the web - it will likely make sense to use something like the Logic example in GEF as a test case as you make progress for your work.

There is also the thought that ideally you might want to consider using Canvas for this code to be run faster.

Feel free to drop me an e-mail as you are considering some of these and working on the project.

Regards,
Vineet

2009/4/2 Jacob Beard <jbeard4@xxxxxxxxxxxx>

Hi Paul,

Thanks for the response. Here's what I posted to the newsgroup:

Hi all,

I'm currently in the process of completing my proposal to Google Summer of Code. I had the idea to make it so that the GMF workflow can be applied to the creation of web-based, domain-specific editing environments. The way I envision doing this is that it would slot into the existing GMF workflow such that, at the last step, where you generate an Eclipse plugin that implements a domain-specific diagram editor, you would have the option to generate a web-based, thin client version instead (much like you can for an RCP version), so, in addition to the EMF Java classes implementing the abstract syntax model, it would also generate an HTTP servlet, and all of the necessary _javascript_, HTML, and CSS to implement the editor. You would run it by launching the servlet and navigating to "http://localhost:<port_number>/<project-name>" in your web browser.

I talked to EMF and GMF newsgroups, and received positive feedback, but it was not until I posted on the soc-dev mailing list that someone pointed me to several resources involving the e4 project, including a presentation from EclipseCon on exporting GEF-based applications to Flash so they run in your browser[0]. So, clearly the goal of that project and the goal put forward in my project proposal are the same. But it appears as though they may go about it in different ways. Please correct me if I'm wrong here, but it seems as though the e4 project would take the entire generated diagram editor plugin and "export" it to Actionscript/Flash by running it through a compiler. The way I am proposing to solve the problem is one step higher, before you generate the code the implements the generator. I'm not sure what the relative advantages/disadvantages are between these two approaches. But, because my goal and the e4 project goal seem to be so closely aligned, I'd like to make sure that the work that I do: a) may at some point integrate well with the work done so far by e4 developers, and b) does not reduplicate the work already done be e4 developers. An example of a component I might be able to reuse is, for example, if e4 has exposed some core Eclipse functionality via RESTful interfaces over HTTP. Exposing EMF classes via REST is definitely going to take up some time during my project.

Another difference between my proposed project and the work that was laid out in the presentation is that I feel like it is important to minimize Flash dependencies. I intend to target the Dojo libraries, including dojox.gfx for a vector graphics canvas. So, as the work that has been done on e4 appears to be mostly Flash-based, immediately reusing my work might be difficult.

I'll be submitting my proposal tonight, and I'll still have the opportunity to iterate on it after that, but I'd appreciate it if people could let me know what they think about my project idea and how I might be able to align it with the work being done on e4. I'll also post a link to my proposal in a bit, after I've had some time to finish it up. Also, if anyone would be interested in mentoring me, I'd like to hear from you.

Oh, and one last thing: I can't seem to find the eclipse.e4 newsgroup on the news.eclipse.org server. I'm submitting this message through the web form... Is this newsgroup located on some other server?

I'd greatly appreciate any guidance anyone can offer. Thanks,

Jake


And then:


If anyone is interested in the details of how I intend to implement this, the draft of my project proposal is here: http://docs.google.com/Doc?id=ddrn5wx9_1240gztszsg6

This proposal is a moving target, even after I submit it, but still I think the core is stable.

I'd love to hear what people think. Thanks,

Jake

On Thu, Apr 2, 2009 at 7:43 AM, Paul Webster <pwebster@xxxxxxxxxxxxxxxxxxx> wrote:

Hi Jacob,

Here is probably a good place to ask.  Traditionally eclipse -dev
lists are for the development of the component, but our newsgroup list
is so low traffic that I would probably be the only one to see your
proposal there :-)

PW


--
Paul Webster
Hi floor.  Make me a sammich! - GIR
_______________________________________________
e4-dev mailing list
e4-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/e4-dev

_______________________________________________
e4-dev mailing list
e4-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/e4-dev