Talk:Common Object Request Broker Architecture

From Wikipedia, the free encyclopedia

Modification note (7/1/2003)

I just thought I'd clear this up to avoid a mod-war. IDL stands for Interface Definition Language, *not* Interface "Description" Language. It was recently changed incorrectly to the latter. I've put it back the way that it was. See the CORBA spec p 3-1.

 (Derek)

I noticed that someone broke my change above when doing some edits. I've re-edited it so that it is now correct once more.

 (Derek)

Contents

[edit] CORBA/Common object request broker architecture

68.167.249.197 23:44, 7 Apr 2004 (UTC): I'd suggest that it is standard practice to have the main article body for CORBA/Common object request broker architecture to be under the fully expanded version of the name (cf. Remote procedure call/RPC, and File transfer protocol/FTP to name two examples).

That's right! I am sexy 23:46, 7 Apr 2004 (UTC)
I would argue that File transfer protocol should be moved to FTP, though Remote procedure call should stay as it is. This is because there is not just one file transfer protocol. Having a page title like "File transfer protocol" or "Common object request broker architecture" implies that there is more than one - for example - "a file transfer protocol is..." or "a common object request broker architecture", whilst "an FTP" or "a CORBA" is less appropriate. Dysprosia 06:40, 8 Apr 2004 (UTC)
68.167.249.2 09:02, 8 Apr 2004 (UTC): Admittedly "file transfer protocol" can be generically used, but I disagree about "common object request broker architecture" which has established itself as a specific standard. That is emphasized outside of wikipedia by capitalizing all the words. Additional examples, like OSI, DNS, SNMP, LAN, FDDI, SMTP, ICMP, RARP all redirect to the full words, though HTTP doesn't.

[edit] an exception proving the rule; disambig COBRA and CORBA

68.167.249.197 00:03, 8 Apr 2004 (UTC): One good standard practice within wikipedia is to not disambiguate misspellings. I agree that this is a good guideline to follow. As I also note at Talk:COBRA, I would contend that COBRA vs. CORBA is a case where the exception proves the rule. This is a case of two five-letter acronyms, which is unusual. For two acronyms of that length (and, to get perhaps overly analytical about it, the fact that B and R are both typed by the left index finger on a QWERTY keyboard), isn't a reasonable to anticipate our audience would appreciate such a disambiguation here? I would never advocate disambiguation for TLAs. but this seems a case where a good rule has a reasonable exception.

68.167.249.2 09:02, 8 Apr 2004 (UTC): Dysprosia's suggestion elsewhere of a disambiguation block prefacing the CORBA page seems like the best of both worlds to resolve this.

[edit] KDE Using CORBA

Wasn't KDE using CORBA and then dropped it? 81.151.193.21 17:10, 9 January 2006 (UTC)

That was GNOME, I believe; KDE fanboys at the time often pointed to the CORBA-usage as a sign of the over-engineering which they believed plagued the GNOME project. --217.128.105.9 14:50, 3 February 2006 (UTC)
The CORBA in GNOME is not dropped and still used. Audriusa 07:54, 25 April 2006 (UTC)

[edit] Images?

Hi. What are these supposed to indicate? Image:100%.png Image:50%.png Why are they not defined? Thanks. --ChrisRuvolo 01:53, 6 Apr 2005 (UTC)

Developmant Status. In researching the links I found project which are not fully developed and/or which have not been updated for many years. With the % rating on what to expect when following the link. If you know the project and find the rating unfair you are free to correct - just be honest - a project with barely make minimum feature set and has seen no update for 3 years should not be marked 100%. --Krischik 06:49, 6 Apr 2005 (UTC)
Ok, thanks for the info. I don't mind their use, but in the article it should be defined what they represent.. and then what the various implementations are lacking (justification of each 50% rating for example). They won't be able to be interpreted usefully otherwise. --ChrisRuvolo 16:23, 6 Apr 2005 (UTC)
I don't understand your context - what images? 68.100.20.83 (talk) 00:35, 17 November 2007 (UTC) Sorry for not logging in, I'm Seaneparker (talk) 00:36, 17 November 2007 (UTC)

[edit] Current status

Information about the current status of CORBA is needed. What alternatives are there? Is it evolving? Are there new developments that use CORBA?

I'd love to provide a status but I'm neither a member of the OMG nor do I work with CORBA at the moment (I did at my last job last month :-) What would you like the format to be? i.e. links to "current projects"? White papers? Conference references? New standards in process? Seaneparker (talk) 00:38, 17 November 2007 (UTC)

[edit] VMCID

Is the VMCID really important enough to deserve an article? For that matter, is it even worthwhile to merge into the CORBA article? Wikipedia isn't a comprehensive reference guide for CORBA. That's what the CORBA specs are for. -- Whpq 17:34, 7 September 2006 (UTC)

I agree with this. An article (or section) on VMCID, in my opinion, doesn't have enough encyclopedic value to justify an article on Wikipedia. That sort of thing belongs in a programmer's handbook. --Tschel 22:10, 14 September 2006 (UTC)

[edit] The criticisms and success sections

The language of these sections seems more like an internet flame war than an encyclopedia entry. —The preceding unsigned comment was added by 194.25.80.243 (talk) 09:56, 7 December 2006 (UTC).

I agree, these two sections are full of contradictory statements. A neutral reader will end it up wondering if CORBA was really successful or a real failure at all. Also found the section for Success lacking for not citing enough references. (SirGalahad 21:09, 4 January 2007 (UTC))
I dont think the article can argue Waldo's A note on distributed computing is biased. His paper is the seminal "limitations of dist-comp" paper and was written while Sun were wholeheartedly backing CORBA -they were one of the co-founders of the OMG, after all. Java RMI came along a lot later; Java ships with a (dated) ORB and even EJB defaults to IIOP as its protocol. Admittedly the paper applies to more than just CORBA; it applies to anything that tries to make remote stuff look local (SOAP, Java RMI, DCOM). But it is valid in Corba, because a corba OID doesnt include any location hints at all; you have to use an Orb to find things. —The preceding unsigned comment was added by SteveLoughran (talk • contribs) 12:02, 27 February 2007 (UTC).
Also, I don't know how "fundamentally false" the notion of location transparency is. I do think its impossible to make a service on a remote system appear to be something local, and not have it fail in unusual ways (e.g NFS timeouts blocking file access operations). What may be possible is for programs to be written to talk to a system that is remote, without caring where the remote system is. Perhaps "fundamentally false" should be downgraded to "impossible to maintain in the face of network failures" SteveLoughran 17:37, 11 April 2007 (UTC)

The tone needs to be made more neutral and more examples of these supposed failures/short comings are needed; otherwise the criticism is really unsubstantiated.


If you look at the source of the criticisms, its quite full of commented out opinions. That discussion should be moved to the talk pages, where it belongs. SteveLoughran 22:48, 20 June 2007 (UTC)

FYI - I altered the language to be a bit more neutral (if that's possible for me to do) however I did not add any more references, etc. Please give feedback on whether improvement or detriment... seaneparker 21:26, 1 September 2007 (UTC)

Some of the language is in the past tense. This makes it sound like CORBA is no longer in use. 194.72.110.12 (talk) 11:39, 7 February 2008 (UTC)jac

Still in use, but it is not as ubiquitous as the vision. You can still find it in the Gnome desktop for Linux/Unix. SteveLoughran (talk) 18:39, 15 March 2008 (UTC)
I had a reread of the criticisms section. 1. We need to mention Microsoft and their lack of support. If all of the vendors at the time embraced CORBA, it would have stood more of a chance. As it was, MS pushed COM/DCOM. The other issue is that the criticisms says that Java RMI is the way forward. But Web Services, built on WS-*, and HTTP-based REST services have changed that. Maybe the big limitation was not the competing technologies, but the emergence of the Internet itself. CORBA was built for LAN-based distributed computing, and you now have enterprise wide and internet-scale problems, for which alternate solutions are needed. It was HTTP that stopped CORBA. SteveLoughran (talk) 18:39, 15 March 2008 (UTC)

[edit] NTCIP

I'm not sure where a link to National Transportation Communications for Intelligent Transportation System Protocol (NTCIP) would fit into this article, if anywhere. NTCIP offers CORBA as one of two possible choices for an application layer protocol. Is there a more appropriate listing for standards or other products that make use of CORBA? Squideshi 18:59, 27 June 2007 (UTC)

[edit] Red Hat delivers

"(Red Hat Linux delivers with the GNOME UI system, which has its IPC built on CORBA.)"

In addition to being grammatically unsound on several levels, this sentence is difficult to understand . . . and misleading at best. Red Hat neither engineered nor is the maintainer of GTK+ ("the GNOME UI system"). I'm looking to move/modify this, but I'd like some input on what the original intent was. Was the author merely confused as to Red Hat's responsibilities and what GTK+ is? —Preceding unsigned comment added by Brian Geppert (talk • contribs) 17:47, 12 March 2008 (UTC)