>>>>> "Russell" == Russell E Glaue <russ@stripped> writes:
Russell> I don't mean to say what way is wrong or right, but if a feature is added
Russell> to MySQL, should not that feature comply with the design and strategy in
Russell> which MySQL is built upon?
Russell> If we are going to make the syntax similar to DB2, should we not do this
Russell> for the entire database and not just one feature? Or lets edit the MySQL
Russell> design model to explicitly support the integration of IBM DB2 grammer.
Russell> To have the entire database compliant with a design model in which
Russell> everyone follows, and then have one (or just a few) feature(s) sticking
Russell> out differently seems a bit chaotic. -- It's not a part of the model, yet
Russell> once in the code it will have to be supported for who knows how long.
Russell> It's just my opinion that to have standards and a design model is to make
Russell> sure the road in which technology is built on is straight and accurate and
Russell> less breakable. It may not seem like a big deal to waiver from this model
Russell> this one time for this feature, but if we are not strict and do not stick
Russell> to this model, then what about the next time, and the time after that?
Russell> Am I wrong?
Russell> What do other people think about this issue?
With MySQL we have the following design ideas:
- We are not afraid to add new useful things to MySQL, even if they
are not 'strictly' SQL.
- When we want to add syntax to MySQL, we first check if it's in the
SQL standard, then if some other SQL server supports this and if
neither of the above is true we add it in a way that should be most
- If some feature exists with a different syntax in some other widely known
database, many people is using this feature and we can add this
syntax with a minimum of changes, I am for doing this.
- In the case you can do things in two way's we will have one of the
syntaxes as the 'recommended syntax'.
This is quite close to your ideas, isn't it?