Page 1 of 2 12 LastLast
Results 1 to 15 of 21
  1. #1
    Join Date
    Mar 2008
    Posts
    11

    The Conceptual Schema

    Hi there,

    I am currently studying and focusing on databases. I have heard that most conceptual design stuff is done using entity relationship modelling.

    However, having studied both ER and ORM (object role modelling) I have formed the opinion that ORM is vastly superior to any flavour of ER or UML for conceptual db design.

    So, I was wondering if anyone who works in this area could shed some light on why ORM has not been embraced to the same extent as its lesser counterparts.

    Is it just that the ER tools are cheaper and/or more readily available?
    That has been my only gripe with ORM as a student. There is no software that you can use for it (without paying big bucks for the enterprise edition of visual studio). So all of our diagrams are done using MS Word.

    Otherwise ORM cooks!!!

  2. #2
    Join Date
    Apr 2002
    Location
    Toronto, Canada
    Posts
    20,002
    Quote Originally Posted by onefootswill
    So all of our diagrams are done using MS Word.
    there's your main problem right there!

    rudy.ca | @rudydotca
    Buy my SitePoint book: Simply SQL

  3. #3
    Join Date
    Feb 2004
    Location
    In front of the computer
    Posts
    15,579
    If the only tool you've got is a hammer, then every problem looks like a nail... Don't try to repair watches though! The problem with ORM (Object Relational Modeling) is that it can only cope with O, and treats RM as an afterthought.

    There are many ways to create code, and object orientation is only one of those methods. Many of us deal with databases that were designed in "flatland" when data was kept in files that were managed by applications, and only the applications that had access to the file definitions could use the data. Some of use have to deal with NPL (Non-Procedural Languages). ORM fails miserably in trying to cope with either extreme.

    If you are sure that you'll only have to work with Object Oriented Programming (OOPs), then ORM is "just the ticket" for your modeling needs. ORM is a good methodology, and it is quite adequate for many purposes as long as you don't exceed its limits.

    -PatP

  4. #4
    Join Date
    Mar 2008
    Posts
    11
    I think you are mixing up Object Relational Modeling with Object Role Modeling.

    Object Role Modeling has nothing to do with OOP.

    But I agree with r937, it needs a good tool. Has anyone used a good Object Role Modeling tool? I have heard of some tools that spit out a bunch of tables based on your relational schema diagram, but I have never experienced one yet.

    I'd love to hear what they are like.

  5. #5
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    I think it is you, and Dr. Terry Halpin, Distinguished Professor and Vice President of Conceptual Modeling at Neumont University, who are mixing things up.
    "ORM" is most commonly known as Object Relational Mapping.
    Mr. Halpin should come up with a new acronym....
    If it's not practically useful, then it's practically useless.

    blindman
    www.chess.com: "sqlblindman"
    www.LobsterShot.blogspot.com

  6. #6
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    ...and my favorite modeling tool? - Microsoft PowerPoint. Fast, cheap, simple, readily available, and lets me get right on to the important stuff which is implementing the design.
    If it's not practically useful, then it's practically useless.

    blindman
    www.chess.com: "sqlblindman"
    www.LobsterShot.blogspot.com

  7. #7
    Join Date
    Feb 2004
    Location
    In front of the computer
    Posts
    15,579
    Quote Originally Posted by onefootswill
    Object Role Modeling has nothing to do with OOP.
    Thanks for that, I needed a good laugh this morning!

    It is somewhere between a BFD and an HK that TH used ORM when he meant ORM instead. If that doesn't make you laugh, you are WAY too serious for me today. If people are going to snag an acronym, why can't they at least pick a unique one or add some sort of qualifier to make them unique.

    My first observation still stands... The ORM you are referring to is hopelessly mired in the O and only uses the RM to support the O.

    As a demonstration, try to model a schema that would be useful in a COBOL or an Eiffel/SmallTalk environment. ORM falls apart, just like ORM does.

    -PatP

  8. #8
    Join Date
    Mar 2008
    Posts
    11
    OK yes. Halpin and I have stolen an acronym. I'm more than happy to give it back.

    If you can model a schema for COBOL or an Eiffel/SmallTalk environment using ER, then you could certainly do it with Object Role Modelling.

    You could almost say that ER is a subset of Object Role Modelling. It can do everything ER can and more. Also remember, the Object Role Modelling space is the conceptual schema. Implementation issues are irrelevant at that point. It basically just captures business rules in the same way predicate logic does. Only, because it uses natural language, it does it in a simpler manner. You then map your schema to tables (the logical schema). At that point, you start thinking about implementation issues.

  9. #9
    Join Date
    Jun 2003
    Location
    Ohio
    Posts
    12,592
    I've yet to meet the client who will pay me simply to produce an ER diagram, so I don't really give a f.
    If it's not practically useful, then it's practically useless.

    blindman
    www.chess.com: "sqlblindman"
    www.LobsterShot.blogspot.com

  10. #10
    Join Date
    Apr 2002
    Location
    Toronto, Canada
    Posts
    20,002
    i've met several clients who wanted to pay me to produce an ER diagram

    unfortunately, they were all students needing to hand in an assignment, and none of them could afford me!!

    rudy.ca | @rudydotca
    Buy my SitePoint book: Simply SQL

  11. #11
    Join Date
    Feb 2004
    Location
    In front of the computer
    Posts
    15,579
    Take a quick shot at modeling the SYS hierarchy from IMS. This is easy to do with ERM because you simply model the levels indendantly, then show the decendant and sibling levels using standard notation. An object model can model each hierarchy level, and it can handle descendant levels as though they were multiple inheritance, but at least as far as I know there is no clean way to model the sibling levels within a hierarchy.

    I'm sure that ORM is a wonderful tool for doing what it does well. IDE1FX is what I learned, and I thought it could model anything and it is still the best all-around modeling notation that I've found even though it does have some shortcommings. Experience is a wonderful thing, if nothing else it keeps reminding us that we're human.

    -PatP

  12. #12
    Join Date
    Jan 2007
    Location
    UK
    Posts
    11,434
    Quote Originally Posted by blindman
    ...and my favorite modeling tool? - Microsoft PowerPoint. Fast, cheap, simple, readily available, and lets me get right on to the important stuff which is implementing the design.
    ...my favourite modelling tool is a classic: pen and paper.
    George
    Home | Blog

  13. #13
    Join Date
    Mar 2008
    Posts
    11
    To be honest, I do not know what the SYS hierarchy from IMS is. But it sounds like an implementation. So IDE1FX would be the appropriate thing to document that. It doesn't sound like something which is being modelled conceptually; rather an implementation that is being documented. IDE1FX does not really occupy the conceptual realm. It certainly has its strengths and place. But the conceptual realm is not it.

    It seems the theme here is "go with what you know". And that's fair enough. It does give some insight as to why object role modeling is still a minor player.

    As I said, I've studied IDE1FX, Barker notation and Chen’s original notation of ER. None of them have the power of object role modeling. And that is why I was perplexed as to why it hasn't been taken up as widely. It takes a bit longer to learn as well. That could be another factor.

    And don't be fooled by the word object in the name of object role modeling. Apart from the subtyping aspects of it, there really are no parallels with object-oriented technologies. (In fact, the things being modeled are actually called Entity Types.)

  14. #14
    Join Date
    Feb 2004
    Location
    One Flump in One Place
    Posts
    14,912
    A shame certus is not posting on here much nowadays. I imagine he might have some comparative experience. I think the gist amongst those who have responded so far (me included) is "OR what?"

    BTW - I propose ETRM as a new FLA.

  15. #15
    Join Date
    Nov 2004
    Location
    out on a limb
    Posts
    13,692
    academics are always coming up with new tools and techniques
    the knack is to spot the tools and techniques that actually have a usefull role in life, and those that can stand the test of time.

    there are some tools that are better than others, but die because of lack of support, NIH syndrome or others sticklers.
    there are others that are not so goof but continue becuase they are easy to use, well understood and straightfoward to use.

    just beacuse an Academis say blah, doesn't mean that blah is the technique to sue right here right now, sometimes it takes a while for the academics ideas to get translated into a working toolset.

    so I' dsuggest if ORM is the dogs dangly bits that you seem to beleive, and there are no modelling tools to support it.. then theres a market opprotunity to develp such a tool. mind you I wouldn't mind betting that something frotm he Eclipse stable could well be an answer.

    Academics are jst as prone to companies (and religions for that matter.. there is one true tool/god/religion, all others are blasphemers). A colleague of mine once claimed religion is a bit like detergent, each manufacturer proclaims their produt as the only tru product that will wash your souls whiter, and each manufacturer churns out promotional material to support their brand awarness (whether that be the bible, the new testament, the Koran, the whatever.

    as to whether ERM is a subset of ORM.. seems unlikely it sounds as if ORM's inventors have grabbed what best bits or ERM and other techniques and repacked them.

    who cares, as long as the tools are upto the job it really doens't matter. somebody who knows ERM, suing an ERM tool is going to be a far better prospect than someone using ORM who doesn't know what they are doing. what matters is the quality of the finished system. there is nothing worse in my books that someone introducing new tools and techniques wihtut adequate information or experiance.

    you assert ORM is far superior.. is it? Not to me, because I've yet to come across it, and I need toolsets right here right now that can deliver workng systems. I have neither the time or resources to go away on some speculative technique that as yet isnt in common usage
    I'd rather be riding on the Tiger 800 or the Norton

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •