Talk:Database management system

INGRES AS THE BASE OF MANY COMERCIAL DATABASE PRODUCTS -- Is it true?
The article includes this paragraph, which has no references associated with it.

"Many of the people involved with INGRES became convinced of the future commercial success of such systems, and formed their own companies to commercialize the work but with an SQL interface. Sybase, Informix, NonStop SQL and eventually Ingres itself were all being sold as offshoots to the original INGRES product in the 1980s. Even Microsoft SQL Server is actually a re-built version of Sybase, and thus, INGRES. Only Larry Ellison's Oracle started from a different chain, based on IBM's papers on System R, and beat IBM to market when the first version was released in 1978."

I have first hand knowledge of several of the products mentoined here and I do not believe they were based on INGRES, as is claimed.

I do know that SQL Server did begin as a version of Sybase, that is correct. I do not know that Sybase was a version of INGRES, as is claimed, but it may be.

It is my belief that Informix shares no heritage with INGRES, (IE: no common code) and was an organic outgrowth of Informix's early products, which were development languages. Of course people familiar with INGRES maybe have worked on it. (That is not the claim being made here, though.)

I am equally sceptical about the Non-Stop SQL Claim.

I'm going to do a little checking up, but I may delete or rewrite this paragraph. If someone is invested in keeping these claims here it would be helpful if they would source them.

24.22.76.36 (talk) 20:11, 21 September 2012 (UTC)

Misplaced comments
It is a set of prewritten programs that are used to store, update and retrieve a Database.

How it is possible............. —Preceding unsigned comment added by 192.8.211.11 (talk) 08:16, 16 July 2009 (UTC)

There are articles for relational database and relational database management system but they have very less content compared to the related section on this page. Much of the information on this page can be moved there. Jay 15:54, 2 Jan 2004 (UTC)

First RDBMS
copied contents from Talk:Oracle database

The claim is made that: "Oracle is the world's first RDBMS." Surely, this is not the case. I am unsure which is the first but, amongst others, IBM System R and Logica Rapport were around before Oracle, I believe. Can someone clarify please? Geoff97 18:09, 1 Jan 2004 (UTC)
 * Multics too claims to be the first RDBMS. We need a consensus.
 * "...(RSI) was founded in 1979 and released Oracle V.2 as the world's first relational database.".
 * "Multics Relational Data Store (MRDS)... is believed to be the first relational database management system ...".
 * Jay 07:29, 2 Jan 2004 (UTC)

RSI was started in 1979. There are references to other RDBMSs before that date, which seems to eliminate Oracle as the first. The question is which was the first? See #10 here for a reference to RAPPORT-3 from Logica:  Geoff97 10:33, 2 Jan 2004 (UTC)
 * The entry for relational database in List of firsts can be edited accordingly. Jay 21:05, 2 Jan 2004 (UTC)

The following appears to be misleading:

''Relational databases took over the entire database market,there are some tasks which they are not very good at. In particular the use of keys to link related records together can be particularly slow when dealing with many common databases. For instance, to find the address of the user named Bob, a relational system would look up Bob in the USER table, find his "primary key" (the login name), and then search the ADDRESS table for that key. Although this appears to be a single operation to the user, it still requires a complex and time consuming search through the tables.''

The described method is not the usual way to retrieve related data from multiple tables in a relational database. The usual way is to use a join or a subselect, the behavior of which is not determined by specific underlying implementations (such as searching one table, then searching another). In a real, production database one might, very likely, construct an indexed view encapsulating the needed join, specifically instructing the DBMS to optimize for the case of SELECT on that view. This allows for the benefits of optimization without sacrificing relational integrity or explicitly handling pointers. --FOo 22:14, 10 Feb 2004 (UTC)


 * Agreed. It looks a bit as if a multi-value database enthousiast has been writing this. Any suggestions on how to rewrite this? -- Jan Hidders 18:55, 12 Feb 2004 (UTC)


 * Unfortunately, I don't know enough about "multidimensional" databases to present a correct introduction to them. It might to do drop the introduction and simply say, "Another data model is ...." However, an awful lot of the text of that section is written from a misled "relational bad" POV, now that I look at it. I was hoping that the writer would speak up and clarify what s/he actually meant! --FOo 04:13, 13 Feb 2004 (UTC)

-

I offer these comments as a retired software engineer who worked in the Oracle Server Technologies Division for 15 years beginning in 1994. These notes are based mainly on my recollections from that time, with at least one note from other personal experience between about 1979 and 1983. Also, one of my friends from UCLA in the 1960s, Stu Feigin, was Oracle employee #5.

What I had heard from inside Oracle was that the first RDBMS was Ingres, from UC Berkeley in the 1970s, and that Oracle was the first RDBMS to be delivered as a commercial product. IBM was also early in the RDBMS market, but apparently was slightly later than Oracle as a shippable product. When I attended a session at a convention around 1983 or '84, it appeared that the commercial RDBMS market consisted almost entirely of Oracle and IBM, perhaps even exclusively Oracle and IBM. IBM's large mainframes contributed to Oracle's early growth indirectly because Oracle licensing charges at that time were set at 10% of host hardware cost.

Ingres went from academic project to commercial project at a date that I'm not aware of. UCB chose to use a Quel (is it spelled QUEL with all caps?) as the Ingres query language. Meanwhile, IBM and Oracle used SQL, originally named SEQUEL. It's possible that IBM might have started its RDBMS work first but that Oracle, starting as a garage shop, moved faster in bringing a product to market. Also, Oracle probably was more aggressive in marketing its RDBMS than other companies. —Preceding unsigned comment added by Paul Raveling (talk • contribs) 05:52, 8 September 2010 (UTC)

2 items caught my attention. One is the indication of 'relational' to 'relationship' and related data. The term 'relational' has to do with describing data as a set of normalized tables or RELATIONS (set theory, I believe).

The other, as has been discussed, does sound like Pick propaganda. Is Pick still around? It has no advantages. The statement indicates that normalization is a bad thing and that de-normalization is a good thing. The performance cost of normaliztion usually can't be measured and there are a number of circumstances where normalization improves performance. Where denormalizaiton pays off (and it can) is usually very specific applications that require retrieving very large amounts of data in a linear fashion. A de-normalized table essentially replaces a view. it can also be very difficult to manage since it may have a lot of duplicate data that may change. All that redundant information has to be located and changed - at a high cost in performance. De-normalization may make some retireveals faster, but it doesn't help much at update time. — Preceding unsigned comment added by DanFfx (talk • contribs) 16:50, 21 December 2011 (UTC)

Term clarification needed?
copied contents from Talk:Data model

I have the impression, that the whole area of data model, database model and database management system might have to be made clearer and restructured. The practical use of terms often is unclear and leads to inproper use of terms.

Proposal: --Udo Altmann 15:26, 17 Oct 2004 (UTC)
 * Database model : everything what is related to general design/concepts how data can be structured (relational, hierarchical, network, object-oriented, ...)
 * Database Management Systems : technical aspects of the implementation of database models (e.g. languages like SQL, indexing, ...)
 * Data model : aspects of designing a specific application (Entity-Relationship- Diagramm, UML, ...)
 * Database : the implementation/instance of a data model


 * I certainly sympathisize with your effort to clean up the terminology-mess in this area, but I'm not sure if your solution is the correct one. You seem to suggest to limit the content of articles to a specific meaning but articles should deal with their subject in the wider sense of the word. If the term is often used in a sloppy way and/or has multiple meanings then all these can and should be discussed in the article. So the article on "data model" should discuss all the meanings of the term, including the sloppier ones. Also note that redundancy in articles is not a bad thing and in fact desirable because articles should be more or less self-contained. I'm also a bit wary of the term "database model" because I don't think this is a widely used term. -- Jan Hidders 00:03, 18 Oct 2004 (UTC)


 * I would like to point out that currently (April 2005) this article gives Sage Payroll and Microsoft Money as examples of a "database application", but the actual wiki entry for that term redirects to "Database". This doesn't seem consistent.  -- User:Anonymous13242354 20:37, 23 Apr 2005 (UTC)

Paul Raveling (talk) 05:56, 8 September 2010 (UTC)

Perceived shortcomings of RDBMSs
The article now says:


 * But a more serious cost is that relational database management systems are unable to store many to one relationships without normalisation: As a consequence Oracle, Sybase, MS SQL Server etc are not well suited to the storage of denormalised data. Other database management systems such as Pick are used instead.

I strongly object to this. It's not only very close to the usual Pick propaganda, and therefore should be formulated in a more NPOV manner, but as it is formulated now also simply nonsense. So I would like to ask the author of this to reformulate it such that (1) the terminology is correct ("many-to-one relationship" is probably not precise enough, and "denormalized" is probably also not correctly used here) and (2) it is made clear that this is a position that is not widely held by the experts and in fact denied by most of them. Of course, good arguments that show I'm wrong are also welcome. :-) -- Jan Hidders 14:57, 22 Nov 2004 (UTC)

I contributed towards the para you quote and I agree with your comments. I did change it a little afterwards. It is similar to some of the (often laughable) Pick propaganda. You might remember me arguing in comp.databases.pick in the early 90's with Pick proponents. However, some of the propaganda was true! And there is nothing incorrect in the paragraph you quoted. However, why anyone would set out to store denormalised data when having the power of Oracle available is the question. Paul Beardsell 12:00, 15 Dec 2004 (UTC)

Directory service
Why leaving out such important area of real world DBMS application from the article? Why is it so "unusual" compared to other examples? Novell's entire product line is based on such kind of DBMS. And Novell is not some small and obscure company. In fact the functioning of our entire network here depend on that kind of DBMS, including human resources and such. In the views of way too many people the DBMS is well suited for some sort of accounting applications only. And all examples in the article are suggesting just the same. But directory service is not just the user database with their passwords as some believe. It's an object DBMS of it's own right, though without so much emphasize on transactional integrity and with high access speed and scalability. Neglecting _any_ reference to DBMS for managing collections of network objects doesn't seem at all like NPOV to me.


 * I don't see it as NPOV. I see it as: Somebody has to spend some time writing about directory services as a type of database management systems. Feel free to do that. If you are unsure how to phrase it, then you could propose an addition here in the talk-page before you integrate it into the article.
 * As a start, I've added Directory service to the "See also"-list.
 * TroelsArvin 12:42, 14 Dec 2004 (UTC)


 * Actually my complain was about its complete removal from the article without any replacement. I feel fine about moving it to the "See also" at this time. 13:49, 14 Dec 2004 (UTC)


 * I removed it because it was right at the top of the article giving it undue importance. Seems to me each of us could have our own pet database application area listed (data wharehousing, etc etc) - but not at the top of the article.  I am happy with (any number of) relevant See also links, of course, and I would have acted better to move the link there.  Paul Beardsell 11:53, 15 Dec 2004 (UTC)

Associative Model of Data
A new database architecture called the Associative Model of Data has been patented. This is a non-relational database and a fully functioning Associative DBMS, called Sentence's, has been developed. I have used it and I would like to discuss the Associative Model of Data briefly. Comments? --Certus 08:28, 12 Feb 2005 (UTC)


 * I'm not familiar with this particular data model, so I can't comment on the details. However, I'd like to caution that many times in the past, someone has come up with a brand-new super-powered data model that is supposed to take the world by storm and fix all the problems of previous models. Frequently their claims are based on misunderstandings of previous data models (for instance, the idea that "relational databases are two-dimensional") or on "flexibility" gained by relaxing the constraints which permit data integrity guarantees (for instance, the idea of the MySQL developers in the early days, who claimed that transactions are for lazy programmers). It's crucial that we avoid hyping any new development, particularly if it is so new that it has not really been thoroughly tested in practice and in the market. --FOo 19:45, 12 Feb 2005 (UTC)


 * I'm going to delete this section as it is cut and pasted from . Note the copyright statement at the bottom of that page. --User:Craig Stuntz 6 April 2005

Did Stonebraker found Sleepycat?
At a quick glance I haven't seen any evidence for this, not in the BerkeleyDB or the Michael Stonebraker Wikipedia pages, or elsewhere on the web. -- Greenman 28 Feb 2005

Stonebraker founded Illustra, which was later purchased by Informix. I've never heard of a relationship between Stonebraker and Sleepycat. Joelm 04 Aug 2005


 * The claim that he founded Sleepycat's been in this article quite a while, without much comment. I don't know enough to decide whether to remove it or not. Does nobody have better sources? Greenman 09:12, 5 August 2005 (UTC)


 * I mailed Sleepycat. Stonebraker didn't found Sleepycat (he founded Illustra). Sleepycat was founded by Keith Bostic and Margo Seltzer. I'll correct the article. TroelsArvin 14:20, 9 August 2005 (UTC)

History of MySQL
MySQL is an open source database whose slogan lays claim to be "the worlds most popular open source database" see slogan on MySQL.com  O'Reilly article indicates MySQL was developed from msql which was originally developed as a frontend to Postgres. At the time Postgres did not have true SQL so, msql translated SQL into Postgres' Quel language. msql was developed [in 1994?] by David Hughes at Bond University in Australia as part of his PhD project developing a network management system. MySQL was developed by Michael Widenius while working at a Swedish company TCX using the msql SQL parser attached to a TCX devloped "Unireg B+ ISAM handler."  As for marketshare claim, BusinessWeek article quotes Evans Data survey and International Data Corporation (IDC). "A growing headache for the database giants," Steve Hamm, Business Week, October 21, 2005.  Jim Callahan JimCal80@aol.com

Desktop DBMS
Does anybody remember any desktop databases prior to dBase II on the CP/M operating system? dBase II became dBase III when ported to the IBM PC-DOS. A later version was named dBase IV. The .DBF file format used by dBase III is still in use. FoxPro was a competitive implementation of the xBase language and was purchased by Microsoft in part for its very fast "Rushmore" indexes which Microsoft intended to use in a forthcoming product which became Microsoft Access, which was sold as part of the Microsoft Office suite. Lotus, the company that developed Lotus 1-2-3 worksheet introduced Lotus Approach as its database offering and the company that develeped the WordPerfect word processor introduced DataPerfect. On the MacIntosh the leading database, was and is FileMaker. This all from memory and needs to be fact-checked. Jim Callahan JimCal80@aol.com

DBMS's vs DBMSes vs DBMSs
I am wondering what the consensus about the appropriate plural form of DBMS should be. I have style guides that say never to use an apostrophe-S to make a plural (as it is normally used to indicate possession), and a few that say to use it when it would be more confusing not to do so. I am inclined to use DBMSes, because the pronunciation of the letter S ends with the S sound, which in standard usage is typically pluralized with es (this follows from the same logic as using "an RPG" instead of "a RPG" because the pronunciation of the letter R begins with a vocalic sound). And there's always DBMSs, which could be confusing, though perhaps not nearly as confusing as Ss (in reference to multiple instances of the letter S). I know this is really minor, but it should be consistent throughout the article, which requires some decision about which to use.&mdash;Kbolino 01:55, 26 May 2006 (UTC)


 * I agree with "DBMSes". --FOo 07:06, 26 May 2006 (UTC)


 * I also agree that "DBMSes" would be the best choice. --Cgsguy2 05:47, 3 August 2006 (UTC)

SQL
The article says multiple times that SQL violates the model of a true RDBMS. Can we figure out someway to clarify that. Is there an article to link to? Or can we put in more about where SQL violates the model? Considering that SQL is the primary form of DB out there and most uninformed users probably think that SQL is relational.--Parcelbrat 07:02, 12 July 2006 (UTC)


 * I wrote a bit about that in Relational database. --Craig Stuntz 13:03, 12 July 2006 (UTC)


 * It's not so much that it violates the model of a true RDBMS, it's that SQL gives database programmers the leeway to build other types of DMBSes other than just the relational model. Same for Access, ORACLE, and most others.  - Mugs 16:56, 7 March 2007 (UTC)


 * In a truly-relational DBMS, SQL is adequate for a query language, but is cumbersome. Saying that SQL violates the model of a true RDBMS is not an incorrect statement, SQL allows for nulls, nameless columns, duplicate rows, ... And We're talking about the SQL language here. If you want me to describe how MS SQL Server isn't relational, that's a whole different ball of wax. McKay 14:05, 8 March 2007 (UTC)

List of Database Brands
I think this is unencyclopedic, contributes nothing to the article, and is a waste of space. I think it should go and isn't even worth keeping as a separate "list of" article. We already have Category:Database management systems. Any dissent? --Craig Stuntz 13:46, 23 August 2006 (UTC)


 * Sounds good to me. --FOo 05:07, 24 August 2006 (UTC)


 * I was tempted to work on improving this, since it appears to be in random order and to be missing some major ones (especially the largest, IBM IMS) by subdividing it into Mainframe DB's, early PC Db's, and current PC DB's. But I'm not sure it's worth the effort.  What value would it add to this article?  206.55.187.170 (talk) 20:24, 5 October 2008 (UTC)

What the?!?
Okay, I've been lurking on the page for a while now, intending to give it some serious treatment (I did a major overhaul to Relational database), but I've only been able to troll vandalizers. It wasn't until today that I gave the page a serious read. Just the opening paragraph got me a little frustrated.

The opening definition defines Database System (which is a redirect to this page sadly enough). DBS and DBMS are *not* synonymous. The numbered section under Description, accurately describes a DBMS. Hmm, on second look, the rest of the article looks fine (as a DBMS), but the opening paragraph needs to go into it's own article (under DBS). There wouldn't be a lot of content there (at least initially), but they are totally distinct topics. McKay 04:48, 23 January 2007 (UTC)
 * DBMS is the software used to manage the databases. MySQL, MSSQL, Dataphor, MS Access, FileMaker. These are DBMSs. One can use a DBMS to make a DBSs.
 * DBSs are programs that accesses a database (and hence, use a DBMS). SugarCRM, any corporate accounting program, human resource programs.

Vandalism
On the history page of this article, you can see that many recent edits are reverts and undos. Personally, I also find this page hard to read, since it doesn't seem to have a clear structure. So maybe this page should be locked, and/or be reworked too. It's just a suggestion, but I hope some of the people interested in this topic will consider it, in an attempt to clearly explain to beginners what Database Management Systems are, without scaring them off.--81.240.182.77 14:20, 20 October 2007 (UTC) Julle praat net n klom POES!!!!!!!!!!! —Preceding unsigned comment added by 155.238.4.85 (talk) 05:01, 13 May 2010 (UTC)

Opening
The opening definition doesn't ring true to me, perhaps because it reads funny:

A 'database management system' (DBMS) is computer software designed for the purpose of managing databases based on a variety of data models.

It sounds like every single DBMS manages databases based on more than one data model, but it seems like it should read more like a single DBMS manages databases which are based on one (or possibly more) data models. Maybe a good sourced definition is in order? Unfortunately all I have handy are RDBMS books, and their definitions are slanted toward the relational model. SqlPac (talk) 02:35, 22 March 2008 (UTC)


 * Fixed. --FOo (talk) 20:51, 5 October 2008 (UTC)

Should low-end database application programs be a separate section or article?
Should low-end database application programs such as Filemaker Pro or Microsoft Access be contrasted in a separate section or article? It seems a little buried in here, and most computer users are probably more concerned about how to use them than the larger systems. Oicumayberight (talk) 21:26, 15 November 2008 (UTC)
 * Added to Database section on applications. -- Beland (talk) 07:36, 16 March 2013 (UTC)

DBMS Definition Seems Weak
To me essential to a DBMS is the concept of maintaining data integrity and backup and recovery. Data is changed via atomic units of work called transactions which cannot be split up (for instance a bank transfer transaction would include taking money from a savings account and moving the amount to a checking account - either both the credit and debit to the respective accounts takes place or neither takes place). A DBMS ensures that transactions once completed are never lost by utilizing a system of backup and in the case of a failure recovery. These are essential features of a fully functional DBMS and only backup is lightly touched upon in the article. Jmurtha (talk) 12:32, 10 April 2009 (UTC)

Recent history
The history section needs information on the history of database management systems after the mid 1980s. -- Gordon Ecker (talk) 06:09, 19 June 2009 (UTC)

Duplicated introduction (signed) moved from article
DBMS: AN INTRODUCTION (By: Manik Sharma)

DBMS stands for Database Management System, is a system which manage the overall working of database. The concept of DBMS was given by Dr. E.F Codd who is also known as guru of Database technology. As from its name it is clear that DBMS is collection of three words:

Database:In simple words database is collection of related data i.e. database is defined as collection of interrelated files/tables. It should be noted that collection should be interrelated.

Management: It is the set of techniques, which are used to manage the overall activities of entire organizations

System:	It is set of interrelated components having a specific goal to achieve. For example: An education system, in this system various interrelated components are Teachers, Students, Non Teaching staff, and the objective of system is to develop the carrier of students

Moved by --Boson (talk) 11:10, 11 October 2009 (UTC)

Question
How does this DBMS Works! —Preceding unsigned comment added by 41.188.139.98 (talk) 12:49, 9 November 2009 (UTC)

Lead section needs copyediting
CAN WE PLEASE HAVE PEOPLE WHO UNDERSTAND SENTENCE STRUCTURE DO THE WRITING IN WIKIPEDIA? 70.64.0.253 (talk)


 * It's a fair comment (belongs here though, not in the main article). Andy Dingley (talk) 15:08, 3 December 2009 (UTC)

what is dbms?
DBMS is a software that enables designers to define ,create,maintain and control access to the database (KUSHAL)

Current trends, where xQuery!?
This is a relevant part of the "current trends", and a concrete one! —Preceding unsigned comment added by 189.111.77.219 (talk) 12:06, 21 January 2010 (UTC)

Single File Database: Develop a student record management program, which uses single file as its database. It is not supposed to create additional files at any part of the project and should never rewrite total content. It must be capable of storing 100000's of records. It should support variable length records to be added, deleted, modified, listed. —Preceding unsigned comment added by 202.53.10.54 (talk) 15:36, 2 September 2010 (UTC)

BDSM
"Not to be confused with BDSM" at the beginning of the article seems a joke made by a funny kid. Could it be removed? —Preceding unsigned comment added by 188.25.185.170 (talk) 22:10, 20 March 2011 (UTC)

The BDSM article still has "Not to be confused with DBMS" on it. The error is equally possible either way. Expurgating text because it's amusing despite its relevance is far more wasteful and inefficient than adding text just because it's amusing, despite its irrelevance. This is why serious people are generally disliked. -Ryan —Preceding unsigned comment added by 68.175.3.89 (talk) 23:02, 23 May 2011 (UTC)

== Elements of Database ==

1) Field: It stores one type of data element. 2) Record: It is collection of related fields. 3) file: Collection of records. 4) Database: Collection of files is called Database. Sadu jaya chandra prasad — Preceding unsigned comment added by 122.184.53.23 (talk) 12:01, 7 July 2012 (UTC)

Elements of Database
1) Field: It stores one type of data element. 2) Record: It is collection of related fields. 3) file: Collection of records. 4) Database: Collection of files is called Database. — Preceding unsigned comment added by 122.184.53.23 (talk) 12:05, 7 July 2012 (UTC)


 * Bare references to person removed, a more complete reference to a public work would be needed. I would note these definitions are not compatible with an Oracle Database for example. Djm-leighpark (talk) 05:26, 22 August 2018 (UTC)

Length
Well, I managed to condense about 11K out of this article, but it's still overly verbose. I'm sure the intro and first two sections could be rewritten to be more concise. Certainly the "Database storage" section goes into a lot of detail which can probably be deferred to linked articles. In the long run, this article could probably be combined with database to make a better article that is shorter than the sum of two, but both need to get cleaned up separately first, I think. -- Beland (talk) 19:11, 5 March 2013 (UTC)
 * Gave a big push of consolidation and merging; this is now redirecting to Database. -- Beland (talk) 07:33, 16 March 2013 (UTC)