[V4RB2] Time trials vs. MySQL

[Olivier]

Le 8 avr. 05, à 08:51, Robert Brenstein a écrit :

> Considering that SQL code in V2 is no longer custom developed by 
> Ruslan, how realistic is getting performance anywhere close to what V1 
> had?  I expect improvement but I doubt V2 will ever be so fast. Having 
> to give up some performance may be a tradeoff of having a more 
> standard-compliant parser, though.

OK but in that case, there is a problem. All the marketing of Valentina 
is at present built on its extreme speed.
Its reputation and the advertisements (in RB developer for example) 
speak a great deal of its very important speed.
Many developers chose and choose Valentina because they needed a really 
very fast database.
it risks to have a lot of disappointment on the performances there are 
sharply less good than with Valentina 1.
I hope that the team of Ruslan will manage to optimize Valentina 2 to 
approach performances of Valentina 1.

In my case, I chose to use API of Valentina rather than SQL so that my 
application is the fastest possible.
But it is sometimes necessary to use SQL. So for all the network 
applications and all the applications already developed with Valentina 
1.
I think that for all the developers which are not Japanese, the use of 
the UTF8 will already improve the performances (many? I do not know).

In the magazine RB developer number 3.3 (in January February, 2005), 
page 32, Paul Lefebvre made benchmarks on SQLitePro (and the database 
interns of RB 5.5), it would be interesting to see already (before 
optimizations) which are the performances of Valentina 2.
Anybody knows Paul Lefebvre ?
We could ask him to make the same tests with Valentina 2 or to ask him 
for its database of test.

olivier