[XSL-LIST Mailing List Archive Home]
[By Thread]
[By Date]
RE: DCOM is now open code
Subject: RE: DCOM is now open code From: "Didier PH Martin" <martind@xxxxxxxxxxxxx> Date: Sat, 6 Mar 1999 17:04:10 -0500 |
HI Jeremy, <YourComment> Hmm. The relevant page at Open Group, http://www.opengroup.org/comsource/ contains this definition of the DCOM package: While COMsource does not include the application programming interfaces (APIs) for object linking and embedding (OLE), or ActiveX? Controls, it does provide the underlying technology upon which they (or similar technologies) can be built. I read this as telling us that we get a *drawing* of a box on four wheels, based on which we can design a car... It does absolutely *nothing* toward making the real-world OLE and ActiveX apps freely interoperable across platforms. Somebody thinks we are all mushrooms... <bg> <YourComment> <Reply> More and more with this thread I feel like a guys in the middle age having said something wrong against the wholly institutions and that the next step is the inquisition :-) Active X control are simply a marketing plug period. In fact active X interfaces are DCOM interfaces. However script languages like Ecma script, PERL or Python needs a IDispatch interface which basically is used for late binding (same concept as CORBA late binding interface). The package includes a IDispath marshaller and anyway, because it includes the MIDL you can compile yourself a IDispatch interface with it. But the marshaller is included. So it is technically feasible to implement late binding objects that could work with script language. But what this means is that any visual stuff is not part of the package. Obviously DCOM has to be Windows independent and all helper function provided to implement OLE which is essentially for visual objects or objects that will interact with the user. So to be really windows independent, the package should not contain any windows specific libraries dependencies. With practical experience, I would say that the package is OK for non visual objects or objects that not depend user interactivity. For example the package could be used to implement a DOM or object like that. It could be used also to implement an interpreter or parts of an interpreter. Anyway, you understand the point anything that do depend on windows, button, menus, etc... which are in fact platform specific. So the package cannot include OLE stuff witch is windows dependent because it is about menu negotiation, in place editing etc... To tell you if you are or not a mushroom is not within my competencies :) However, I can tell you that the package could be used for any "back end" or "model" type of objects like for instance interpreters. That any objects not using desktop libraries will work. Any object that need user interactivity with the desktop library elements won't work because it is platform dependant and the consortium choused not to include any dependencies with Windows stuff. And, the writer made a mistake, you can implement a certain class of active X objects as long as they do not have any user interface. The package do includes the late binding marshaller named IDispatch to implement such objects. The MIDL can also generate TypeLibs with could be included in objects for late binding method discovery. So, I cannot say if they provide you with a car without any wheel but I can tell you that this is about the same kind of car that CORBA provides you. A set of tool to implements services not visual objects or interactive objects. But if you need fast services you can create a DCOM interface, map it to a compiled language and compile it on different platforms without having to rewrite your code. If you need cross-platform visual object this is not the right tool. Java is better. Is you want to implement a service and that maximum speed is not an issue, Java is better. </Reply> Regards Didier PH Martin mailto:martind@xxxxxxxxxxxxx http://www.netfolder.com XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: DCOM is now open code, Jeremy H. Griffith | Thread | Beginner's question: The | doesn't , Rebecca Chan |
RE: DCOM is now open code, Didier PH Martin | Date | W3C-transformation language petitio, Oren Ben-Kiki |
Month |