[prev in list] [next in list] [prev in thread] [next in thread] 

List:       nextgen-online
Subject:    [Nextgen-online] Module Registration
From:       sean () datamage ! net (Sean Legassick)
Date:       2004-09-29 11:53:31
Message-ID: 415A86AC.5030007 () datamage ! net
[Download RAW message or body]

James Scoble wrote:
>>>> sean@datamage.net 09/29/04 11:22 AM >>>
>> I think you meant 'like "1@init"' unless there's a reason I've
>> missed for adding some leading zeros?
> None at all, and it doesn't. But its simple way of showing that the
> numbers start at 1 and count up.

I see what you mean, but now we are using strings there is no 
commonality between '0001@init' and '1@init' - they are completely 
different PKs - so perhaps best to stick to what's actually used to 
prevent confusion.

> If someone really does need to add that specific text in default
> data, we can make a workaround. But except for that one possible
> problem, the PKVALUE method seems  simple and functional to me.

I think we should be shooting for higher quality design in KEWL.NextGen, 
as I think it's very achievable. Not that we need to develop all the 
higher quality designs right now, but if anything has a "possible 
problem" then it's on my list to be improved, regardless of how likely 
you or I might think the problem is to occur.

We are intending that many thousands of people (if not more) will be 
using KEWL.NextGen, so we cannot accurately predict whether what seems 
like an insignificant problem to us will also be insignificant for others.

>> but we can wait until we convert the register.conf to an XML-based
>> format to rectify this.
> 
> 
> We will still need to have some method of putting primary keys in.

No, we will just need a method of specifying that primary keys should be 
put in, something like:

<defaultdata>
   <record generatePrimaryKey="true">
     <field name="some_field">
         some data
     </field>
     <field name="some_other_field">
         some other data
     </field>
   </record>
</defaultdata>

(we'd probably make generatePrimaryKey="true" the default so in practice 
you'd miss that attribute out).

Sean

-- 
Sean Legassick
sean@datamage.net

[prev in list] [next in list] [prev in thread] [next in thread]