From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Split Database management systems again[edit]

I am reasonably minded Database management systems (DBMS) should be split from this Database article to separate and clarify the concepts and to avoid undue weight. This would likely be a contraversial split. The predicate restricting this article to DBMS controlled databases is likely inappropriate, (although it is broken at one point) especially given its a level-4 vital article. Id' go even farther to say article with this name should have a broader and summarising view of the subject broadly covering the scope of the Outline of databases page. We should surely be asking is the article compatible with Microsoft Access, Apache derby, Embedded database or is it over-focusing on internals or databases like Oracle or DB2 to the omission of a user view of a database like say Wikispecies ? I currently calculate I don't have the bandwidth or energy to pursue this.Djm-leighpark (talk) 12:42, 21 August 2018 (UTC)

Article Improvement August 2018[edit]

I've spent a few days looking at this article and believe it really needs a rework. One key issue is this is the root article for databases and therefore Wikipedia:Make technical articles understandable should be an important objective for the wider audience. Some issues I have are:

  • The current constraint of the article to DBMS controlled databases is overly restrictive for the root article
  • Technical discussion on DBMS dominates the article and the expense of general database discussion. I am personally inclined regret to that the merger from Relational database management system took place but the state of the articles prior to merger meant something really have to be tried.
  • The summaries where the main article elsewhere are often too long
  • The article at points is biased towards the corporate enterprise databases an the expense of others.
  • There are various statements throughout which are biased, outdated or just plain wrong.

I currently have on loan the Connolly/Begg book which is suitable for sourcing a lot of items (ISBN: 978-1292061184).

Ideal world I'd set up a page and try things asking collaboration. In practice I'll try a series of sprints/pause, mainly because I may have to stop at any point.

Plan A:

  • Sprint 1: Refactor article gathering DBMS together under single level one heading (this will include history). Some Heading changes. (Examples to Classification; to Languages under Database Access, both can be developed later). No content changes
  • Sprint 2: Move DBMS definition from lede to its own section. Rework the lede as need to cover that. Point inbound redirect links to Database Management System to the DBMS Section.
  • Sprint 3: Introduce structured content early, probably in form of structured work expansion/rework of the Applications/Classification section, examples of corporate, personal and embedded databases.

Thankyou. Djm-leighpark 20:53, 23 August 2018 (UTC)

  • I have now completed Sprint 1 and refactored the article (no content changes) and will now pause for a while. I will add the need to add a little prose start of the Data access section. I also am aware eventually I am likely to use Connolly/Begg's definition of DBMS to add or clear things from that section. But for the moment a pause. Comments welcome.Djm-leighpark (talk) 22:41, 23 August 2018 (UTC)
@Djm-leighpark: Hmm, I'm not sure making the distinction between database and DBMS makes a lot of sense. For example, what is there in the core history of DBMSes that's not a core part of the history of databases generally? A pile of data isn't very exciting without some way to access it, and the way the data is organized is always closely related to the DBMS, or at the very least the type of DBMS. -- Beland (talk) 01:21, 24 August 2018 (UTC)
@Beland: Thankyou for your comments. I see the data as import in itself, and I believe there is content that can usefully be written than concentrates on the data, application, automated feeds and dataset extraction while treating the DBMS/Database system as a black box, and that such content can be of use to less technical users. I'd tend to agree that if you asked by to rattle of a history timeline of databases I'd actually rattle off a timeline of DBMSs. In the timeline of Database some events such as the theoretical work are not directly focused as the DBMS. On review there view history section best needs to move back out of the DBMS section to top level so I've just done that. I'll proceed with Sprint 2, but not before late Saturday or Sunday, which will help scope the DBMS section ... (I'm only able to access intermittently before then for short replies/fixes etc). ThankyouDjm-leighpark (talk) 08:51, 24 August 2018 (UTC)
@Djm-leighpark: How would you propose splitting up "Performance, security, and availability"? Most of the material there seems like good info to have in the main database article, and especially things like storage arrangements and transactions which to me seem like core topics. -- Beland (talk) 12:43, 24 August 2018 (UTC)
@Beland: The basic idea first to use Begg/Connollys 8(from Codd)/(plus 2 of their own) and perform a two gap analysis(both ways). Handle and resolve any gaps. The more techie stuff stays where it is. Important topics to go to an overview section summarised non technical language, easily said not always so easily done. A key objective is to try any achieve goals of 'Make technical articles understandable'. Djm-leighpark (talk) 16:31, 24 August 2018 (UTC)
@Djm-leighpark: Hmm, that doesn't really answer the question of which subtopics belong with "database" and which belong with "DBMS". Without having the book in front of me, I'd say if "database" covers "data that is stored and how that data is arranged" and "DBMS" covers "how that data is accessed and manipulated", it seems to me that all the subtopics have to do with both sides, which is why I didn't split them up. Note that (to pick a random competitor) the article on databases also just discusses both at the same time.
You said above you didn't think this article should be limited to "DBMS controlled databases". I assume you're talking about something like SQLite? Certainly this article should cover all databases, and the fact that some exist without DBMSes can certainly be noted and examples given. If we're doing that then I think we definitely wouldn't want to put "Performance, security, and availability" under DBMS, since those are considerations for serverless databases as well. A bunch of sentences may want to be reworded to reflect a broader array of possibilities, though databases running as services with DBMSes are a big part of what there is to say about databases.
Making the article more accessible I think is an orthogonal problem. What I would do for that is to put the "Applications" section before "History", and add a new section before History that gives an elementary-school example of a SQL database and for diversity some type of non-SQL database. -- Beland (talk) 18:07, 24 August 2018 (UTC)
@Djm-leighpark: thanks, generally expect to take on board most of your points last tomorrow. Djm-leighpark (talk) 21:44, 24 August 2018 (UTC)
@Beland: Over the course course of 24 hours I have come totally round to moving "Performance, security, and availability" backup up to the top level for the reasons you suggested and perhaps more besides and have done so. I have a desire to call the section something other than "Performance, security, and availability" and would prefer to use the term "Database Features" and to add specific "Availability" and "Performance" sections to approach from the reasoning for these features. But I need to refocus on the 'Sprint 2' and "Database Management System" which I intend to do tomorrow morning. Thankyou.Djm-leighpark (talk) 17:31, 25 August 2018 (UTC)
@Djm-leighpark: OK, cool. "Database features" sounds like a good and more concise title. It might also work to simply promote all the subsections of that section to be top-level sections, since they're mostly major topics, but whatever you think is best. Should "Database management system" go under "Database access"? -- Beland (talk) 02:50, 26 August 2018 (UTC)
@Beland: Overslept. Good feedback. Using 'Database features' makes we want an 'availability' and 'performance' sections within it ... these concepts being explainable at a minimal technical level with an overview of technique used to overcome (e.g database copies).( hmm Database concepts might be an interesting alternative heading?? need to think and can be changed anytime). Adding 'Application (program)' and some on API's under 'Database access' (actually database interactions). The 'Materialized views' and 'Relication' may be able to be covered under Performance/Availability and may not need a top level heaving. I'm inclined to defer move decisions until after some content addition at the moment.Djm-leighpark (talk) 11:13, 26 August 2018 (UTC) I've done a bit today .. not much. I need to keep/working on that DBMS section. Will need a section on Non-DBMS databases … something cardindexes, books calling themselves databases, and e.g. use of spreadsheets as a database. My have spurt tomorrow morning or evening.Djm-leighpark (talk) 21:28, 26 August 2018 (UTC) I am doing sprint 2 at slow walk rather than a sprint. apologies. Hope to continue tomorrow.Djm-leighpark (talk) 09:03, 28 August 2018 (UTC)
I've sort of finished sprint 2 and possibly have limited time for a short while. I'm open to suggestions: section movements, renames, discussions, etc etc. Of note there are subsections on Availability and performance that are needed (I think I have good ideas and approaches for these and they are self contained) and for a non-DBMS database section (card indexes, books called database and use of e.g. spreadsheets as databases). I have emphasised use of 'application' as the application interacting with a database, and tried to avoid using that word in other contexts. I've just reworked the lede to focus majorly only on defining "database" as per the article to a redirect to "database managment system" section is possible. I no longer have a desire to more DBMS to its own article. And another library book is in for me do I dont have to keep citing Connolly/Begg only (Hopefully). Thankyou.Djm-leighpark (talk) 06:48, 31 August 2018 (UTC)

The lede - October 2018[edit]

I just noticed a minor edit on the lede (lead section) ... then said to myself ... this is too complex in this position and the best option was to remove it. I feel we certainly need to be trying to get to a lede that is self contained so that it makes suitable reading for an 8 year old who looks up database and doesn't want to go any farther. And having removed that sentence I can now see the paragraph on 'Database Management System' contains too much complexity! I hope to look at that in the next while, and perhaps carefully tweak some things out of that into the body. I admit to being a C-Class editor but I'd like to see the lede on this important article brought up to 'A' class ... and I am very sure that is not simple.Djm-leighpark (talk) 20:03, 30 October 2018 (UTC) After walking away for an hour I was able to think of a second sentence that was able to mention design and modelling in a more simple way, and yet in a way that allows say people with MS Access etc to be able o skip the formal design if dealing with a handful of tables. That also gives the lead paragraph more that one sentence which is a better balance for the lede. Thankyou.Djm-leighpark (talk) 21:35, 30 October 2018 (UTC)

missing classification[edit]

I miss the distinction between DEDICATED (in any manner server-based) and NON-DEDICATED DBMS (local only DB, flat-file-based (DBase) or relational (ACCESS, SQLite)) (talk) 13:07, 29 January 2019 (UTC)


I think there should be a reference to the symbol, which is commonly used to represent databases in charts. I also found the explanation to what is exactly the origin and meaning of it here and here. It can even just appear in the leading paragraph as the description of the images. Galzigler (talk) 13:12, 13 August 2019 (UTC)

I am not against it, I just don't think it is an important symbol. I am not sure that a answer is a good enough source for this. It might be possible to find something more academic/journalistic. BernardoSulzbach (talk) 19:13, 13 August 2019 (UTC)
Obviously. That is the reason I didn't make this edit yet. I was hoping someone may have a more reliable source to use for this, because I couldn't find one. I think about adding the images to the article without the description until this issue is resolved, since if for a while they won't be in use they will be deleted. Galzigler (talk) 01:09, 14 August 2019 (UTC)
I am far from convinced there is the symbol. I know of perhaps fours different forms (at least), some being more relevant to particular diagrams rather then others. I suspect different standards may use a different standard symbol. Stackoverflow responses can be dodgy ... some can be great ... others are simply looking stuff up in wikipedia or voicing an indivuals opinion. Djm-leighpark (talk) 03:23, 14 August 2019 (UTC)
  • @Galzigler .. I've chosen at this point to revert your addition. A combination of: two side by side icons in one image in the lead is too messy especially with neither removing nor relocating the previous image; the caption may be trying to claim too much which is essentially your opinion, and thirdly there are other candidates for that icon. There are various candiidates for the database icon ... for example in [1] ... ( is a possible indicator that is an accepted standard ). Additionally [2] is currently indicating license problems needing. Djm-leighpark (talk) 07:46, 14 August 2019 (UTC)
I agree that there is no definitive database symbol, but even this should be mentioned if the article goes over how databases are represented in diagrams. I'd say the main issue here is that there is no reliable source for any claims about what symbols are used and what is their origin. BernardoSulzbach (talk) 16:03, 14 August 2019 (UTC)