automated code vs. manual

convert21 · October 9th, 2009, 12:22 PM

I'm working through this book, and I keep wondering just how much of the automated code/wizards a pro uses. I have about 10 years web development experience with non oop languages.

Let's take a pet inventory data entry form, for instance. I'm used to writing the sql to retrieve the appropriate records from the db, iterating through the records to set up a table or div display of the data, and adding edit inputs. On data submission side, I would validate the data and write the insert/update query. This method actually seems much easier to me than using all the wizards and auto-configuration tools that then generate the code for you. Plus, I you can't see the code that is generated, so if/when I have to troubleshoot, it makes it much tougher.

So my question is what is the middle ground? How would a seasoned .net developer approach the above scenario? Would he/she use the girdview wizard to bind the data, etc., or use that same tag code that gets generated - but code it him/herself. Or would he/she retrieve the data into a dataset, iterate through it, manipulating it as it may need and present it with manual html, css markup? Why NOT do it that way if one has the skills?

Imar · October 10th, 2009, 04:10 AM

Hi there,

The answer, as always in this field, is: it depends.

Controls like the GridView and technologies like LINQ to SQL or ADO.NET Entities are real, and for real. I've built or designed many large, professional web sites that use these technologies. They work well and can result in great sites.

However, they don't always cut it. A DetailsView connected to a LinqDataSource is not always enough. Often, I create separate details pages with separate controls (no DetailsView, FormView and so on) and get data from a business layer that in turn talks to the database. How you get the data from the database is up to you: Entities, custom in-line SQL code, stored procedures and so on.

In short: I try to rely on code generation (from Visual Studio or from external tools) where possible. Once you have a good feel of these tools work, you can really rapidly create well functioning applications. For more advanced scenarios, or where more control is needed, I typically use more custom code.

You mway want to take a look here to see a design of a real world application: http://imar.spaanjaars.com/QuickDocId.aspx?quickdoc=476

Cheers,

Imar

convert21 · October 10th, 2009, 09:54 PM

Hi Imar,
Thanks again! I understand what you are saying. It makes good sense.
Any chance you could expand on this statement? "Often, I create separate details pages with separate controls (no DetailsView, FormView and so on)"

What kind of "control" would that be then? A custom control? What would an example of this be?

Chapter 15 almost down... 3 more to go! Then it's on to Beginning C# 3.0: An Introduction to Object Oriented Programming by Jack Purdum - already ordered!

I appreciate your responses so far.

convert21 · October 11th, 2009, 02:41 AM

duh - you meant html controls, i.e., radio button, text box, etc., right?! I was thinking too hard...

Imar · October 11th, 2009, 04:25 AM

Yes, exactly. One way or the other, the browser usually needs HTML to build up the UI. Controls like the DetailsView serve as composite controls enabling you to configure one control while the control itself manages many different controls internally.

While manually creating forms is a bit more work, I typically don't mind the increased complexitiy it brings because of the increased flexibility...

Imar