Re: [technology-pmc] RE: Sapphire draft proposal

[Konstantin Komissarchik <konstantin.komissarchik@xxxxxxxxxx>]

I am deeply disturbed by comments like this. As someone who is a big supporter of the Eclipse Foundation, I do not want to see it turn into a community that gives monopoly on ideas to projects. If we are wrong, this project will fail and get archived. No harm done. If we are right, perhaps we will better the community with our ideas.

We have considered using EMF and in fact devoted many months to trying to get it to work to solve our usecases. We had several senior engineers working on that, reading EMF bibles, tutorials, etc. In the end, after all that effort, we still couldn't produce something that worked well. Later on, a separate group at Oracle took another stab at the problem. After several more months of investment they got a solution that was a bit more stable, but it was still far too difficult to use.

That was multiple senior engineers working on the problem, but our goal in creating Sapphire was to enable rapid UI development by fairly junior developers. We found EMF to be poorly suited as the basis of this.

I believe we have done enough in scope statement to clearly state that it is not a goal of this project to compete with EMF. I do not believe that this project proposal should need a stamp of approval from EMF to proceed. What do you expect the response to be? I am sure that EMF community believes passionaly in the technology they are working on and would view us simply not grasping the pure truth of it all. Who is right? Only time can tell.

- Konstantin


----- Original Message -----
From: eclipse-bugs@xxxxxxxxxxxx
To: mike.milinkovich@xxxxxxxxxxx, technology-pmc@xxxxxxxxxxx
Cc: konstantin.komissarchik@xxxxxxxxxx, teri.whitaker@xxxxxxxxxx, emo@xxxxxxxxxxx, greg.stachnick@xxxxxxxxxx, pieter.humphrey@xxxxxxxxxx
Sent: Tuesday, June 15, 2010 8:30:22 AM GMT -08:00 US/Canada Pacific
Subject: Re: [technology-pmc] RE: Sapphire draft proposal

On 6/15/10 10:13 AM, Mike Milinkovich wrote:
> I actually don't have any comments on the proposal itself. As far as I am
> concerned, it can go live with the edits proposed by Gunnar on the PMC
> thread.
>
> I do have some worries, but ones which will have to be solved by the project
> once it's underway.
>
> I have concerns about introducing another modeling framework. Personally, I
> am suspicious of the explanation that EMF is too general or too hard to
> learn. If you don't like EMF, wrap it with a DSL via Xtext. That certainly
> seems like a cheaper and more consistent solution than writing and
> *maintaining* a whole modeling framework.
>
> We also seem to be growing a plethora of projects focused on simplifying
> client-side application construction (Riena, Scout). Have you looked at
> those projects as potential consumers of Sapphire?
>    

I've had the same concerns since I read the proposal. In fact, this 
isn't the first proposal that's come in front of the PMC that I've had 
similar feelings about; I just didn't want to be a big nay-sayer or 
discourage developing project teams. But I do have a major concern about 
this kind of duplication.
I'm not necessarily saying that Sapphire shouldn't move forward 
eventually; but I'd be much more supportive if there was evidence that 
the Sapphire team had engaged the EMF and other teams with which it has 
some overlap. In other words, make an effort to work with existing 
projects and technologies, and only after all parties agree that a 
totally new project is warranted, then move in that direction.

My personal opinion is that new project creation should only be 
initiated after engagement with existing projects has shown it to be the 
best direction. Although the Sapphire proposal tries to address the 
topic, there is no evidence that such engagement has taken place.

Eric