It's always a matter of how the data you NEED happens to be organized. For data that needs random access, a method of access designed to optimize random access would be best. For data that needs to be accessed in a certain order, a method of access designed to optimize that particular order is best.
If you don't know how the data will be accessed, then the random-access design is best. When y ou have information about the data to guide you, some other design might be far, FAR better, especially when dealing with yottabytes of data. Example: I want to chose everyone in my city who has a zip code that ends with 17. Using a structure that keeps track of zip coes in order with the ability to index off of the 4th digit of the zip coe will certainly be faster than using a random access technique. To suggest that an enlightened programmer wouldn't provide a zip code index to a database of mailing addresses because that isn't the "enlightened way" of looking at the problem is silly. Database design depends on the USE to which the database is put. Going to great lengths to create the best relational database design for a bunch of addresses in a mailing list is hardly an enlightened use of your time. --- In [email protected], TurquoiseB <[EMAIL PROTECTED]> wrote: > > --- In [email protected], "larry.potter" > <larry.potter@> wrote: > > > > I'm not sure that we can conclude from the database realm to > > the spiritual. that being said it is true that relational d.b. > > is being used far more than hierarchical. The main reason is > > that it was difficult to model a many-to-many relationship, > > There are other reasons, speed of access being one > of them, but the many-to-many thing is interesting > in itself. To me it seems as if the hierarchical > model is something that a seeker (and especially a > deist seeker) would think up. They're always trying > to "climb the tree" to reach something "higher" than > themselves. > > But would a realized individual, who even on the > level of perception relates to the many *as* him- > or herself, be tempted to think hierarchically? > When your entire world view is Unity, are the > devatas (assuming they exist) any "higher" than > you? They *are* you. :-) > > > However, we now that in a web world that demands quick > > access to data, XML is being used to facilitate this, > > and XML has, of course, a hierarchical structure and > > for good reasons: speed, better handling of data elements, > > easier administration, and handling of unexpected elements. > > I honestly suspect that XML's success so far is > based on the relatively small datasets it's called > upon *to* access. When dealing with small amounts > of data, hierarchical structures appear fast on > today's computers. But if you were trying to access > and process yottabytes** of data, do you think you'd > still be using XML or forcing that data to traverse > a tree structure? > > ** Love that term, 'yottabytes' > > kilobyte (kB) 2^10 > megabyte (MB) 2^20 > gigabyte (GB) 2^30 > terabyte (TB) 2^40 > petabyte (PB) 2^50 > exabyte (EB) 2^60 > zettabyte (ZB) 2^70 > yottabyte (YB) 2^80 >
