n tier development strategy question

I don't know where to post this question-or maybe its a comment and I am
just looking for reactions.
We are developing a multi-tiered web app where all database access goes
through a Business Object layer which resides on a machine that is not the
db server.
I am trying to develop the Business Objects in such a way that the web tier
which uses them does not access datasets directly themselves. Instead, the
Business Objects layers provides objects and collection classes representing
the data in the database.
It seems that I am always presented with a trade off of choosing between
encapsulation and performance.
For example, in order to instantiate an object A, the object needs to talk
to the database. So the code for this is encaptulated in A.New. But then, if
object B needs three local instances of object A, each instance will result
in A.New making 3 round trips to the database! To code for performance, I
would have have B return ALL of the data needed to instantiate 3 A's and
pass that to a Friend A.New constructor, but this is at the expenses of
encapsulation and reusuability of objects!
Does this make sense!
HHHHHHHHHHHHHHEEEEEEEEEEEEEEEEEEEEEEEEEE
EEELLLLLLLLLLLLLLLLLLLP
It's driving me nuts not knowing what to do!
Is this a common design problem faced by apps that have a Biz Object layer?The business logic itself may be unique to your applicaiton, but there are
proven and widely accepted design patterns for passing data between the
tiers. For example:
Data Access Application Block
http://msdn.microsoft.com/library/d...aab.
asp
Designing Data Tier Components and Passing Data Through Tiers
http://msdn.microsoft.com/library/d...Gag
.asp
"Chad" <chad.dokmanovich@.unisys.com> wrote in message
news:djj90u$qo0$1@.trsvr.tr.unisys.com...
>I don't know where to post this question-or maybe its a comment and I am
>just looking for reactions.
> We are developing a multi-tiered web app where all database access goes
> through a Business Object layer which resides on a machine that is not the
> db server.
> I am trying to develop the Business Objects in such a way that the web
> tier which uses them does not access datasets directly themselves.
> Instead, the Business Objects layers provides objects and collection
> classes representing the data in the database.
> It seems that I am always presented with a trade off of choosing between
> encapsulation and performance.
> For example, in order to instantiate an object A, the object needs to talk
> to the database. So the code for this is encaptulated in A.New. But then,
> if object B needs three local instances of object A, each instance will
> result in A.New making 3 round trips to the database! To code for
> performance, I would have have B return ALL of the data needed to
> instantiate 3 A's and pass that to a Friend A.New constructor, but this is
> at the expenses of encapsulation and reusuability of objects!
>
> Does this make sense!
>
> HHHHHHHHHHHHHHEEEEEEEEEEEEEEEEEEEEEEEEEE
EEELLLLLLLLLLLLLLLLLLLP
> It's driving me nuts not knowing what to do!
> Is this a common design problem faced by apps that have a Biz Object
> layer?
>
>|||On Mon, 24 Oct 2005 14:27:22 -0400, "Chad"
<chad.dokmanovich@.unisys.com> wrote:
>Is this a common design problem faced by apps that have a Biz Object layer?
Yes.
Google is your friend, it's been endlessly discussed.
My position is that a better appreciation of relational databases as
hold a data model, rather than loose tables, is the first step towards
wisdom.
J.

mysterious restore

I am a support rep for an app running on SQL Server 7 sp3.
The customer had a table mysteriously restored along with the dependencies [sp.s, indexes, etc] for that table.
Is there any logging in sql server that will show when tables or objects have been dropped, recreated or restored?
thanksLook up the msdb..restorehistory table.
--
HTH,
Vyas, MVP (SQL Server)
http://vyaskn.tripod.com/
What hardware is your SQL Server running on?
http://vyaskn.tripod.com/poll.htm
"tlr" <anonymous@.discussions.microsoft.com> wrote in message
news:86C067AA-72E8-49CD-A34A-1E79E4E03315@.microsoft.com...
I am a support rep for an app running on SQL Server 7 sp3.
The customer had a table mysteriously restored along with the dependencies
[sp.s, indexes, etc] for that table.
Is there any logging in sql server that will show when tables or objects
have been dropped, recreated or restored?
thanks

My questions

I'm new for ssas and I'm studying NS samples,i found all the subscribers and subscriptions are created manully,my app have millions of subscribers and many different kinds of subscriptions,so how to create those objects through reading data from tables?

Much appreciate for your reply

In SSNS 2000, subscribers, devices, and subscriptions have to be created using the SSNS API using a programming language like C#, VB.NET, or VBScript. In SSNS 2005, we know have some views that can be used to create those rows in the database.

Here are a couple of links that may help.

http://sqlns.blogspot.com/2005/10/creating-subscribers-in-2005.html

http://sqlns.blogspot.com/2005/09/creating-subscribers-in-v20.html

http://sqlns.blogspot.com/2006/08/creating-test-subscriptions-in-sqlns.html

HTH...

Joe

|||Thanks in Advance

my issues with FT Search

Hi, I am trying to implement a global full text search on our SQL
Server. Our app has several entities that are stored in the DB. I would
like to be able to search for 'John Doe' and get results in all types
of entities. Problem is:
- FT Search does not crawl views. Unfortunately, each entity type in
our system is not stored in full in one table. This is because we are
using pick lists and look ups. For instance: Industry type is stored as
a code that represents an entry in an industries table. However, all
these values are joined in a view to create all the required fields for
the entity a full text search query and index are created per table.
What would be an effective way to work around this problem? (I don't
see any solution to this problem in Yukon either.)
- Ranking: both 2000 and Yukon do not allow for merging of rankings
from several tables/queries. However, it is important to us to be able
to display results from several tables and sorted in a logical way. Any
suggestions/work arounds?
- Performance and scalability: How many rows and/or how many GB can I
have in my DB and still get good FT search query performance (less than
5 seconds)? (I need data for both 2000 and Yukon)
- Same with regards to indexing: ideally we would like to keep the
index up to date in real time. We would like to use the track changes
feature. However, our app allows many users to be logged in and
edit/delete/add entries in the DB. What is the maximum amount of DB
transactions per minute (second?) that will still allow us to keep the
index updated in real time? (I need data for both 2000 and Yukon)
-
I would greatly appreciate any suggestions/tips/info/workaround.
Thanks!On 25 Jul 2005 18:35:20 -0700, barak.benezer@.gmail.com wrote:

>Hi, I am trying to implement a global full text search on our SQL
>Server. Our app has several entities that are stored in the DB. I would
>like to be able to search for 'John Doe' and get results in all types
>of entities. Problem is:
(snip)
>I would greatly appreciate any suggestions/tips/info/workaround.
Hi barak,
I think that your message will attract more answers if posted to the
group microsoft.public.sqlserver.fulltext, where the FTS experts prefer
to hang out.
Best, Hugo
--
(Remove _NO_ and _SPAM_ to get my e-mail address)|||Hugo,
Thanks for your reply. I posted it there and got an answer.