| Vincent,
>>I wonder how you get the attributes of the URL objects.
>>The performance depends very much on this.
To search if the URL exists already, we do an APOQuery from CPL with apodll32.dll.
Do you have a better idea?
>>I would implement the unique URL in a different way by making (a part of) the
>>object reference in the URL.
Let me try to explain the customer context.
He wants to use LNX to get Object framework functionalities: Workflow, access right control, storage,
programming, and so on.
He doesn't want to be linked to Digital, Microsoft and so on.
So, he wants to use the internet technology.
Most of documents will be HTML documents.
Using HTML, to link a document to another, the URL is used.
So, it is not possible to put something like
<ahref="http://www.digital.com/www2lnx?FF0100000000000000000000000010BD00C3></a>
in the documents.
If you stop to use LNX, what could you do with your document and what users can understand?
So, we will use readable URL. Something like:
<ahref="/DECCabinet/Software/LinkWorks"></a>
That's better, isn't it.
Thanks for help
Patrice
|
| Vincent,
>>In this case the query would look like
>>select x where x.urlattribute='yyyy'
>>Right ?
Yes, that's the query, but my question was:
performances expected when more than 500 000 objects will be created?
>>To make this perform you have to set an index on urlattribute for all
>>classes attribute. You can specify this during attribute assignment.
The URL attribute (String 132 caracters) is assigned to the
UserObject class. Then all classes inherit this attribute. Of course
this attribute is indexed. So, I don't think I should specify it for each
class, isn't it?
Thanks for help
Patrice
|
| If this attribute is indexed, you don't have to worry. It will give a good
performance even if you grow well behind the 500.000 objects.
Attribute assignments are derived. So if you assign an attribute to UserObject, all
classes from that point on will have the attribute assigned automatically.
Regards,
Jos
|