Talk:Object database
From Wikipedia, the free encyclopedia
 |
This article is within the scope of WikiProject Databases. |
| ??? | This article has not yet received a rating on the assessment scale. |
| High | rated as high-importance on the assessment scale |
"Object databases are generally recommended when there is a business need for high performance processing on complex data."
I can't let that one pass - whatever the advantages of object orientation, "high performance" isn't one of them. Flexible, probably. It may help with the design process. But the corollary of this is the performance hit you have to take.
--Fjleonhardt (talk) 13:50, 16 May 2008 (UTC)
Cleaned up the links -- removed three, one of which was dead, left the one to the open source benchmarking site and to the ODBMS.ORG site, which is sponsored by OMG through their European agent, Prof Robert Zicari but is hosted (physically) by an open source ODBMS vendor, db4o. "Unspammed" the two remaining. My WikiEditing may stink so hope others can clean it up (if need be) and make it right.
--Charwing 17:59, 10 May 2006 (UTC)
Made the mods discussed below and added some text to explain better what the Object DB Technology WG is intending to do (am a member of it). Thought about adding a link to the Request for Info that was the basis of the 'announcement' cited in the article but decided not to. However, let me use this as a forum to encourage response from those who are interested in this technology area. Responses are due 1 June 2006. Here's the link: http://www.omg.org/cgi-bin/doc?mars/2006-2-18.
charwing 15:52, 3 May 2006 (PST)
Should correct the page regarding OMG -- they were granted the right to develop new specifications based on ODMG 3.0; they did not "acquire" it.
charwing 15:03, 3 May 2006 (PST)
Thanks for your updates, Doug: nice to have solid information from the horse's mouth, so to speak. I would criticize your contribution as being just a little bit lacking in dispassionate analysis, but that's the price you pay for getting a contribution from someone who was intimately involved!
Mhkay 21:04, 20 February 2006 (UTC)
Following all the indecisive debate about merging, I took the plunge and wrote a pretty-well new article on object databases, changing the OODBMS article to redirect to it. (Although the two terms have different meanings, this wasn't reflected in the articles, which were both writing about the same thing). Neither article was very good, both were rather opinionated, and I hope the overhaul is a better basis for moving forward. It's not anywhere near perfect yet and I hope others will improve it, but I felt something had to be done!
Mhkay 20:05, 17 October 2005 (UTC)
- I agree, well done. Here is the old talk page:
Contents |
[edit] For History
The list of the notable object-oriented management systems is a little excessive right at the top of the article like that. The reader starts to lose interest before he or she gets into the meat of the article. I'd recommend striking most of that and then including a list of the notable systems in an infobox to the side, linked from the history section.--Bmccaff 15:20, 15 July 2007 (UTC)
[edit] For Technical Features
In Technical Features, it's written that joins aren't necessary because objects can be located directly by pointers. Are these pointers indexed? How are the pointers referenced by a query? --Bmccaff 15:24, 15 July 2007 (UTC)
[edit] For merge
I just read them both and I'd say go for it. – 24.193.221.118
JG: They are either synonymous, or at least most people understand them as synonyms. I agree with merging. – 62.168.3.222
I'm saying go for the merge. I was thinking about what I would write in the object database article but it seems that most of the content would just be repeated from the OODBMS article. They're not synonymous, I realise that, but they are inextricably linked — they're not two separate ideas, and that's why it's so hard to write one article on object databases and another on object-oriented database management systems without having extensive repetition between them. --BSTRhino 00:46, 29 August 2005 (UTC)
They should be merged. Although there is a valid distinction between a database and a DBMS, the current articles do not make that distinction. They overlap greatly (and they are both rather weak, to be honest). Also, it's in the nature of objects that there is less distinction between state and behavior than there is in the relational world. Mhkay 20:13, 4 October 2005 (UTC)
- I completely agree that the articles need a lot of work, but merging two poor articles isn't going to solve that problem, and (as outlined in the No merge section below) it would only create more problems. Also, let's think about the long-term consequences here: What happens when the articles improve and the concepts are properly distinguished in the text? Do we separate the articles again?
- I should also point out that both of the first two (unsigned) pro-merge comments were from anonymous users, and both comments were submitted before there was any discussion about the pros and cons of merging. I attached the submitters' IP addresses to differentiate them from BSTRhino's comments. – Ringbang 15:14, 14 October 2005 (UTC)
[edit] No merge
No merge. Not only are the two terms not synonymous, merging these articles would create an organizational inconsistency. Currently, relational database management system and relational database have separate articles; similarly, database management system and database are separate articles; federated database system and federated database are also separate articles. Meanwhile, for some reason, object-relational database is a discrete article, while ORDBMS and object-relational database management system both redirect to OODBMS.
OODBMS is not a very good article as it stands, and some of the redirects to it are inappropriate since object-oriented databases and object-relational databases are not the same thing.
By the same token, treating data models and database management systems as the same thing is like saying that automobiles are synonymous with internal combustion engines. In the case at hand, it creates a wrong impression of how we regard these entities in database theory. I say undo the merge-in-progress and concentrate on improving the individual articles. Ringbang 00:45, 5 August 2005 (UTC)
- What about merging relational database management system and OODBMS with database management system instead of with object database? --R.Koot 00:51, 5 August 2005 (UTC)
-
- Honestly, I don't think that would be a good idea. For one thing, the resulting organization would still be inconsistent. Also, database management system is a long article, and most of its size is attributable to the subsection on RDBMS's. A distinct RDBMS article already exists; dedicated articles for the other system-specific subsections also exist. The system-specific subsections of the RDBMS article are lengthy and potentially content-rich enough to have their own articles.
-
- As it stands, no doubt there is a considerable amount of repeated information between the database management system and the wikis dedicated to each respective DBMS model (RDBMS, ORDBMS, etc.). If anything, I think the bulk of the text in the system-specific subsections of database management system should be respectively offloaded and merged into the dedicated articles. A monolithic merge into the DBMS article would make it enormous, unwieldy, and more difficult to maintain. Ringbang 14:26, 5 August 2005 (UTC)
- I very much agree with you. Pavel Vozenilek 00:53, 7 August 2005 (UTC)
- No merge. Different topics, big articles . Rather, a common cleanup is required, to avoid duplicaiton. mikka (t) 00:23, 14 October 2005 (UTC)
[edit] SQL isn't relational
I am editing relational references which are actually SQL references.
Not sure what that comment is doing on this page, but it's a bad idea. Objecting to "relational database" for systems that are a practical engineering approximation to the mathematics of the relational model is like objecting to people describing wheels as circular or roads as straight.
Mhkay 20:09, 4 October 2005 (UTC)
--R.Koot 00:16, 31 October 2005 (UTC)
[edit] Claims re: relational database versus object database
"a relational database becomes cumbersome to use with complex data." -- This sentence does not seem to meet NPOV criteria at all. "It is sometimes claimed that..." would be a good way to start this proposition, and it should be accompanied by some sources or quotations.Ryandaum 03:39, 4 February 2006 (UTC)
[edit] NPOV
Further to my above comment, I think this whole article has serious NPOV issues! It reads more like an advocacy article ... for example : "Although some commentators have written off object database technology as a failure, the essential arguments in its favour remain valid" -- this is opinion unless someone can provide references, or at least say "some consider that ...". Ryandaum 01:03, 9 February 2006 (UTC)
[edit] Smalltalk
Is a Smalltalk programming language image (such as in VisualWorks) an Object Database? Wouter Lievens 10:54, 23 March 2006 (UTC)
[edit] Perhaps a minor issue
I come from an undergraduate computer science background and have some experience with object-oriented programming languages like C++ and Java as well as exposure to the object-oriented design model of software engineering and knowledge of the theories behind relational databases, but after reading this article, I still don't have a clue what an object-oriented database is.
The way the article currently describes it, an object database seems no different from any object-oriented framework like the Microsoft Foundation Classes (MFC). It mentioned traversing objects using pointers. Does this mean it's another name for a binary search tree and similar abstract data types?
I just did a Google search and found The Object-Oriented Database System Manifesto. Maybe citing information from this will clarify the subject for people who are trying to learn about something they don't know.--NeantHumain 02:28, 28 July 2006 (UTC)
[edit] OODBMS features in mainstream databases
It would be good to have a summary of OODBMS features in Oracle, MySQL, DB2 etc.
