| Re: [modeling-pmc] Sven's comment on Comment to Ed's new main text for the Eclipse Modeling Project |
|
Philipp, Comments below. Regards, Ed On 18/08/2012 1:53 PM, Philipp W.
Kutter | Montages AG wrote:
Hi, Sven and PMC members.This again has the sense of "straying for the basic tenets of some orthodox religion." We need to remain pragmatic and avoid being dogmatic. That won't change. As I said, we want to have a page focused on standards. This aspect of EMF will be prominently highlighted along with things like XML Schema and the growing repertoire of supported OMG standardsSorry for not at all agreeing on this split, I will explain below why. Yes, and that's very well supported at Eclipse. Nowhere better I think. It's a key value point that supports the integration we have achieved. E.g., OCL can work nicely with Xtext-based models (even if Sven doesn't care about that) and whether a model is serialized as XML, XMI, or binary, or persisted with Teneo, CDO, EMF Store, or MongoDB, the rest of the frameworks can work reflectively on that data. It's a great achievement and folks who use our technology will be surprised that we've solved more than just the original problem they came to solve. We need to get them past the front door so they can see all there is in the room. As Sven pointed out, actual serialized EMOF interchange, while valuable in principle, is little used in practice. As the OMG community accepts ECore as the pragmatic implementation of Mof, Eclipse Modeling Project, with its "onion layers" like Ed Merks described in his text is thus an a pragmatic implementation of MDA.Exactly. I don't fully agree. Folks working with Xtext, for example, need know nothing about MDA. They might not even like it, perhaps through ignorance, or perhaps it conflicts with the agile processes they follow. In the end, not all model-based development is necessarily MDA. I prefer to provide technology that is process neutral. There is no splitting. There are only different perspectives from which one can view the landscape. The landscape itself is not altered by the point of view, merely the perception is different. This sounds like you're agreeing, but also arguing it should only be viewed from the MDA perspective.CDO is an excellent example of a technology that builds transparently on ECore and thus extends the pragmatic implemenation of MDA, which EMP is with an important runtime component. CDO shows, that even things like persistence , that where considered programmers omain, can be overed with a generic framework, if you leverage modeling technology. I'm sure folks like Vlad Varnica will not tend agree on this point. He'll argue everyone should be using UML2 and that Ecore is only good for implementing UML2. Sven also pointed out perceived bias of MDA.95% of people who know about the real problem in business do not even know what an API is, but they know there business concepts, their properties, dependencies, logic. We all worked for decades to make modeling work for these 95%. As a result, the MDA initiative found, that doing everything with UML2 is the wrong way, and thanks to EMP we where able to explore many more ways to express models. MDA as well found, that the main problem with the 4GL's in the 90is was not that they could not do anything we can do now, but that the metamodels of these DSLs where all defined in proprietary, diverse ways, and thus, all we need is to agree how to define the metamodesls, and the common agreement is: ECore.Yes, this is a fundamental integration point and the reason for this sentence: At its core lies the Eclipse Modeling Framework (EMF) a rich abstraction for describing, composing, and manipulating structured information.But as Sven points out, people would use Xtext even if that were not the case. Fortunately it is the case, so people using Xtext get more than they bargained for with some powerful consequences. Indeed. Ultimately though, the aim is to produce software and well-structured software generally has API for manipulating the specified data. Of course one can use reflection instead of a specific generated API. But that's just a different perspective on the same landscape. It should be clear that this diversity of perspectives on a common landscape is a key value of the modeling architecture we have at Eclipse. Users of our technology will realize this, if we can get them past the front door. A big sign "MDA inside" is less likely to draw a big crowd than would a sign advertising "free drinks inside". I don't see how we gave up anything. The text I proposed highlights the value of a common metamodel, though not using those words, because for a large portion of the people landing on our pages for the first time, that's like speaking in Latin to someone who only knows English.This is exactly the disaster I try to avoid: In the end, the PMC listens to the opinions of the community, and then we act as leaders and make decisions based on what we think is best. It's our primary responsibility. I'm the lead for the modeling project and I, along with the other PMC members are responsible for making decisions. We won't be running a referendum to delegate decisions to the populace. Maybe I am too Swiss, but for changing this weSorry, we are open and transparent, but we are not a democracy. Each group of committers sets the direction for their respective projects. The PMC and I will set it for the modeling project and we will focus on a diversity of perspectives to give the best overview of the full landscape. Sorry, but that's a no go. We aim to attract and inform and will do so. Certainly we intend to have a main page that dispatches to interesting focal points/perspectives. The text I proposed (or some improvement thereof) will ink to those focal points. No, we need a clear message, i.e., a way of leading new arrivals to their appropriate baggage claim area or to their connecting flights.100% disagree. Yes, we'll dispatch to additional pages with a clear perspective and focal point. Of course I understand the value of standards, and the importance of the OMG standards in particular so that will definitely be a page the PMC manages. In addition to that, each project can have its own clear message, and when there is such a message, we will ensure that the PMC-managed pages provide appropriate linkage. There's one darned sentence that says almost nothing. It's a bit incredible to me that an attempt to more informative is viewed much as would an attempt to revise the 10 commandments handed down by God. I don't want another laundry list page. We mention the value of textual notations and that will be a launch point for the DSL perspective (which I must say is something I feel is a key perspective that I believe will drive the future evolution of software development). it as well doe not clear enough tell thatSorry but that information is just Latin, and while it's very precise and has a long history of deep meaning, it's not consumable for the non-Latin arrivals. Yes, I don't see them as separate religions. and "defining your own DSL and generating results from it is much a better way than following the MDA initiative to have a common metamodel of your domain models."Fortunately you get both even if your focus is only on the former. You seem to treat these as mutually exclusive when in fact our technology supports both at the same time. In the end, you don't have to accept it. We will make changes to improve it and we'll discuss those openly and transparently. Of course Microsoft would be happy to see this, because honestly, and this is only my personal opinion, the Eclipse Modeling Project, bringing together the dynamics of open source and open standards is THE killer argument for the Java platform.Microsoft has no coherent solution in this landscape. I doubt they pay any attention to what we're doing, especially not after the demise of M. Otherwise, if you look just for cooler, better, smarter software, you may very probably land with the .net stack.A basic premise of open source is that he who makes the commits makes the decisions. In this case, that's the Modeling PMC. I don't think anything is being destroyed. We're trying to broaden the base of adoption and we firmly believe with reinforce all the value we already have. I doubt the world relies on me all that much. In any case, my responsibility is to keep the ship moving forward in the direction of the most promising horizon not to moor it in a harbor or run it aground on a hidden shoal.
|