Some comments

[Erik Mueller-Harder]

On 8/28/06, Ruslan Zasukhin <sunshine at public.kherson.ua> wrote:

> Important point.
>
> * !!! While you develop SERVER product -- AVOID API way. You have read
> article about ways in WIKI ?
>
> You must use SQL way: SQ: commands, Joins, Cursors. Because this way support
> record locks.

Yikes! Really? Here's all that I've found in the Wiki:

> The main disadvantage of the Pure API style compared with the SQL
> Style of coding arises when working with a remote server.  Using the
> SQL Style programs send small string-based commands to server and
> the remote DBMS can do a lot of work to provide the requested
> result. With the API Style developer must obtain and operate on
> potentially large intermediate results in the client computer which
> can result in the network becoming a significant bottleneck.
>
> On the other hand, given a fast high-bandwidth network connection
> the API Style still can be used quite effectively; especially since
> fast modern networks can often be faster than accessing data from a
> slower local hard disk drive.

I'd assumed from this that I could use the pure API style as long as I
took care to minimize the volume of data transferred. It's true that
the server won't cope with record locking if I use the API style?

-- Erik