On 2005-07-04 18:31:03 -0700, Darren Duncan wrote:
> 5. All details used to construct a connection handle should be 
> completely decomposed rather than shoved into an ungainly "data 
> source".  Examples of what should be distinct (not all being applicable 
> at once) are:
[list deleted]

The main problem I see with this is that *each application* needs to
handle the decomposed form. With the current opaque "data source" it
doesn't have to know what it means: It just reads it from a config file
(or asks the user in a dialog box or whatever) and passes it on to the
connect method. If this is a complex data type like a hash, it has to
construct the hash. If the application programmer only knows about
databases which know about hostname, port, usename and password, he will
write an application which gets these four parameters and stuffs them in
the hash. He will never anticipate that Informix (as Jonathan explained at
length) needs a database and a server name instead of the hostname and
port, so his application won't be able to connect to Informix.

        hp

-- 
   _  | Peter J. Holzer \Beta means "we're down to fixing misspelled comments in
|_|_) | Sysadmin WSR     \the source, and you might run into a memory leak if 
| |   | [EMAIL PROTECTED]     \you enable embedded haskell as a loadable module 
and
__/   | http://www.hjp.at/ \write your plugins upside-down in lisp". [EMAIL 
PROTECTED]

Attachment: pgpCAN29sc755.pgp
Description: PGP signature

Reply via email to