Thanks, Rami, I don;t think we are in disagreement, although I can come across as a bit harsh myself! I went through all the object v relational database stuff in the early 1990's, and always tended to be on the side of relational vendors who want to extend, rather than the revolutionary object vendors. I remember UniSQL and something from HP whose name I forget, they have both fallen by the wayside, but Oracle carries on (never been a fan of Oracle, always was an Ingres supporter, but that shows the power of marketing v technology).
But I still love the way Stonebraker writes - 'they just need educating' - how many times have we all wished we had the balls to say that? Stonebraker has held fast to the relational principles and the need for query languages, so there is probably not too much space between him and Codd/Date. Completely off topic, but I was a participant in some workshops about 'what is a good data model' which is rarely talked about. This got into ISO standards the EXPRESS language, and Shell published some useful work, it's amazing how there are lots of books about the semantics of SQL (the nuts and bolts), but almost nothing about how to evaluate data models for quality and extensibility and maintainability and support for different business purposes. There have emerged quite sophisticated templates for modelling various aspects - my main interest was plant and machinery and maintenance and procurement. Not a simple topic. I can give references if interested. OK, you forced me, here it is, I participated in many of the workshops he chaired in the early 1990's. http://www.amazon.com/Developing-High-Quality-Data-Models/dp/0123751063 How did we get here from 'IN' ? On May 30, 1:04 pm, Rami Ojares <[email protected]> wrote: > I do apologize if I came across as arrogant or impudent. > I do not think that this is a competition. > And as a fairly liberal person I do enjoy when people can do things in a > way that suits them. > > All I am really saiyng is that Date's book have changed my way of viewing > the relational databases. The clearest and most practical book I have > ever read about sql is Date's "A Guide to SQL Standard". > Looks small, looks boring but the content is pure knowledge. > > So while I do agree that a programmer needs to test his works in practice > and work out the final solution often by trial and error, I also have > experienced > those fantastic moments when an idea has been so solid and logical > that the code flows out beautifully from clearly defined logical principles. > > "So you want to store an image as a byte/bit in each row" > > Atomicity is always a matter of perspective. > Atomic to whom? > Image can be seen as atomic from the dbms perspective, when it has the same > operators defined for it just like an integer. Like equality, order > (which it does not have), > and so on and so forth. So using array or any other type is completely > fine with me > as long as the type specific ideas do not start their journey as part of > the relational machinery. > > To be fair that was not your intention. You just wanted that the dbms > would do this type > specific optimization for the array type. And frankly there is nothing > incorrect about that! :-) > I simply reacted against array because often people start pushing all > kinds of datastructures > into relational model. If you want to use datastructures freely then I > would recommend using > an object database. Or maybe JackRabbit where Thomas is also involved. > > The point of relational database is to use the relation as the > datastructure with which > you can model anything. > > And finally rather than speak of correctness I would talk about > designing systems > whose complexity does not grow exponentially when applied to more and > more complex tasks. > > - Rami > > On 30.5.2012 7:34, essence wrote: > > > > > > > > > It is only the naive who want correctness, I want pragmatism and > > performance. > > > All it takes is a benchmark, use an array data type or put all the > > values in rows, > > > Time a query, time a deletion, look at memory. > > > So you want to store an image as a byte/bit in each row? > > > You can keep your 'correctness'. > > > On May 29, 2:06 pm, Rami Ojares<[email protected]> wrote: > >> On 29.5.2012 9:06, essence wrote: > > >>> Date's manifesto is a lot less entertaining, with fewer bon mots, than > >>> Stonebraker's, so I am staying with Stonebraker. > >> Boring? Maybe. > >> Correct? Definately. > >> For wit and wordplay I read Shakespeare ;-) > > >> - Rami -- You received this message because you are subscribed to the Google Groups "H2 Database" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/h2-database?hl=en.
