Re: Replacing WebSQL with a Relational Data Model.

>So, if you *did* specify an API, and it *was* possible to implement that by
"just" using SQLite, you would have
>provided a spec that defines what SQLite does. That would be a good thing,
but I think it'll be tricky to do.

Okay with the relational algebra I am proposing, the specification would be
relational algebra, nothing to do with the SQL database.

The SQL database is simply a backend to implement the relational algebra,
just like I am proposing a reference implementation on top of IndexedDB.

So I have no need to explain anything about SQLite, MySQL etc, because the
query and the response is _completely_ defined by the relational algebra.

So the specification only references the well known mathematical definition
of relational agebra. How that is mapped to the database is an
implementation issue.

We can say with certainty that _all_ compliant implementations should return
exactly the same data if they start from the same state and execure the same
relational statements. Some backends may take longer to execute the
instructions though, just like some JavaScript interpreters are faster than
others.


Cheers,
Keean.




On 26 October 2010 15:53, Julian Reschke <julian.reschke@gmx.de> wrote:

> On 26.10.2010 16:37, Keean Schupke wrote:
>
>> ...
>>
>> Then the superior performance of the SQLite version might persuade more
>> people to implement the standard that way?
>> ...
>>
>
> I think you really need to go through the mailing list archives to
> understand the problems with SQLite. It's not about the *quality* of that
> engine, but the fact that it's not sufficient as a specification.
>
> So, if you *did* specify an API, and it *was* possible to implement that by
> "just" using SQLite, you would have provided a spec that defines what SQLite
> does. That would be a good thing, but I think it'll be tricky to do.
>
> Best regards, Julian (knowing little about databases)
>