True but the whole idea of XML is so it is easy to communicate with other
systems . Just send them the XML and let them use it as they would and other
piece of data. Nobody would write a Java version of a Dataset nor do you
need to  - just send the XML back.

This works well with clients and partners. If you are designing a diverse
system a custom collection is a must - but anyone is mad to design there won
system.

Ben

-----Original Message-----
From: Moderated discussion of advanced .NET topics.
[mailto:[EMAIL PROTECTED]]On Behalf Of Robert Chartier
Sent: Tuesday, 8 October 2002 4:43 PM
To: [EMAIL PROTECTED]
Subject: Re: [ADVANCED-DOTNET] Strongly-Typed DataSets vs. Strongly-Typed
Collections


Another thing to consider..which im sure im not the first to mention it,
is how is this very complex xml which the dataset generates hindering
integration with non-.NET systems?  Something as simple as classic VB
(STK 2/3) attempting to consume the standard .NET "DataSet service"
would be tiresome.  Especially if that service expects the data set back
to provide a list of changes (diffgram...?).

I would hate to the the non-.net programmer that had to implement the
DataSet functionality, etc..  And wouldn't the integration with a more
slimmer-easier/understand to use "Collection service" facilitate the
integration?

.just some thoughts.






-----Original Message-----
From: Moderated discussion of advanced .NET topics.
[mailto:[EMAIL PROTECTED]] On Behalf Of Philip Nelson
Sent: Monday, October 07, 2002 9:38 PM
To: [EMAIL PROTECTED]
Subject: Re: [ADVANCED-DOTNET] Strongly-Typed DataSets vs.
Strongly-Typed Collections


> Since now it's easy to make/use a collection for databinding, I guess
> most of the reasons for using a dataset are eliminated.
>

It sure would be nice to see if a DataSet is significantly worse than a
collection.  When you think about it, and given an axiom I really
believe: "premature optimization is the root of all evil", I would like
to think choosing a DataSet first is the right choice.  If in some
application you discover problems because the number of simultaneous
occurances causes a memory bloat, or massive cpu ticks are needed to
create and tear down your datasets (if it's proved that datasets are
worse), the worst that can happen is that you refactor and use a custom
collection.  In the meantime, one less block of code to check into
source control, one less library to maintain, one less skill to teach
junior programmers, etc, etc, etc...

You would want to do as similar comparison as possible, one simple table
with one column, strongly typed, rows added at runtime,
enforceconstraints off and the like.

__________________________________________________
Do you Yahoo!?
Faith Hill - Exclusive Performances, Videos & More
http://faith.yahoo.com

You can read messages from the Advanced DOTNET archive, unsubscribe from
Advanced DOTNET, or subscribe to other DevelopMentor lists at
http://discuss.develop.com.

You can read messages from the Advanced DOTNET archive, unsubscribe from
Advanced DOTNET, or
subscribe to other DevelopMentor lists at http://discuss.develop.com.

You can read messages from the Advanced DOTNET archive, unsubscribe from Advanced 
DOTNET, or
subscribe to other DevelopMentor lists at http://discuss.develop.com.

Reply via email to