RE: Which DB/Software suite combo is best for Web<->DB implementation

From: Thomas Tatum <thomas_at_koz.com>
Date: 1999/04/13
Message-ID: <7f0c6i$4n9$1_at_news.xmission.com>#1/1


>
>
> I said:
> > > Don't even think about Informix's web datablade.
 Thomas Tatum wrote:
> > I disagree.
 

>
> Validating a date in Informix Web-Datablade tags.
> This resembles no language I ever heard of, except maybe RPG. But not
> really.

I agree the Web Datablade language is very cryptic.

However, you could have written the date validation in perl(via NSAPI), C(via User Defined Routines) or SQL(in a stored procedure).

>
> <?MIBLOCK COND=$(EQ,$MMI_VALID,TRUE)>
 [...]
> <?MIVAR
> COND=$(EQ,$MMI_VALID,FALSE)>$(SETVAR,$MMI_VALDATE,FALSE)<?/MIVAR>
>
[....]

my first web datablade pages took forever to write and debug too. I feel your pain in this learning curve (learning sickle).

>
> Another gripe I have is that each page is an atomic transaction.
> To select for update, or perform certain other kinds of operations,
> requires
> control over transaction boundaries. This is impossible with IUS/WEB-DB
> unless
> you queue a request to an external process with it's own session
> connection.....
>

I wouldnt recommend selecting for update via _any_ embedded web page language, because of the whole web client/server model. If you select for update in a web page on a page pull and the client times out becuase their modem is on the blink you can lock a table out for other users trying to do the same thing.

I think the page-level transaction handling is great, cause if there is any errors/traps during any stage of the web page execution(even nesteed UDR calls or SPL calls), it will throw an exception and roll back all of the transactions for you. What more can you ask for?

If you are catching your exceptions correctly, you can configure an "ERROR" page(s) that will be shown to the users

> And the exception-handling is really cruddy.

does cruddy mean non-intuitive to set up for the first timer? If so I might see what you mean there. most things are.

>
> John
>
Received on Tue Apr 13 1999 - 00:00:00 CEST

Original text of this message