Followed your post on this question over at stackoverflow and thought I’d pass on an observation I made very late last night that may help you get your head around what’s going behind this comment of yours:

<blockquote cite="Duncan">My main problem was that I was trying to access data BEFORE it was bound – so trying to access the data on Page_Load was too early. Although the odd thing was that it worked first time round, but in subsequent requests that data was not available."</blockquote>

I was/am having a nasty problem reading the state of a lot of controls produced by a repeater, I’ve still got my problem in that it fails to get the settings for the first OnClick() event after the initial page loads and is presented to the user, but every subsequent one works – not the problem I’m describing here.

I eventually put in handlers for most common events (Page Load, Item Created, Item DataBound, On Click) and single stepped it in debug mode what I discovered helps explain your observation:

This is ASP.NET 3.5 WebForms

On first page load by the user, ie. initial presentation of the web page the following is the order of the events:
1) Page_Load
2) foreach entry in the repeater:
2a) OnItemCreated()
2b) OnItemDataBound()

(stage 2 gives you Header, then every item (and Alternating Item), then the Footer

On subsequent events like my OnClick event from a button, this is the event order:
1) foreach entry in the repeater:
1a) OnItemCreated()
2) Page_Load

(stage 1 gives you Header, then every item and Alternating item, then footer.
There is no DataBound (which matches the documentation)

So you can see that your page_load happens in a totally different place and different events have and haven’t fired off.

I hope this helps you get some insight into what might have happened.

JOhn

Not really.

As with most things, the answer as to whether it is better is "it depends".

Think of SOAP like RPC over the web – that’s all your doing. It’s procedure-oriented.

Think of REST like object (or rather RESOURCE) -oriented. You always have the same Operations (HTTP verbs) – GET/POST/PUT/DELETE.

In short – it gives a much nicer interface, thereby encouraging people to consume your data. I’m betting that Twitter wouldn’t have been half as popular if it didn’t have a REST interface.

not sure if this what you’re asking for, but I always use the indexer on the attributes collection instead of adding them; the end result is that I never get the dreaded duplicate issue.

Check the bottom section http://tortoisesvn.net/docs/release/TortoiseSVN_en/tsvn-howto-minlogmsgsize.html