mocella

I'd also missed the Scale method. I'll have to keep that one in mind the next time this topic comes up. Thanks

I think I was paying about $8,000US/year for a state (read: semi-taxpayer supported) university in the late 90's. I'd imagine that same school costs more like $10,000US/year now. And the best part is how they manage to make it virtually impossible to pass in 8 terms - almost everyone in my program went 9 terms (4.5 years) to get everything done. My degree is purely in CS, so I feel like I learned a lot of the concepts I needed, but then I learned whatever languages I need now on the job - that's pretty much any kind of path from education to real-world though. Now, was it worthless? I'd say definitely not for me. It's pretty hard to quantify, but I do know the "better" your school is considered, the more money you're probably going to make, at least from what I've seen from developers with pretty similar experience, just the university differing on the resume. Crazy stuff.

The easy way out is to try and set up control anchoring (like to top/bottom/left/right), but I'm not sure that'd allow the user to resize the control themselves, this just makes the control stretch/shrink in relation to the form size. The more labor-intensive way: First you need to determine what the standard relative size is - taken as a width_percentage = current_control_width / current_form_width (and the same for the height) Next, on the event of the user resizing the control, do the same thing after they end their resize and use those percentages if the form is later resized - at which time you'll have to use the override percentages to size the control again.

Sounds like the poster-child for a smart-client (href-exe, url-execute, whatever-the-heck-they-call-it-these-days) opportunity. The basic idea here is you develop your rich client as a win-forms application, but when you're done developing, you can just plot the .exe/.dll files onto a web-server and point to that file (http:\\localhost\someapp\runapp.exe) through the browser, and your win-form app will launch. There's some requirements though - like having the framework installed on your server (1.1 is more geared towards this as 1.0 was much more locked down and restrictive in some of the settings needed for this type of app). Another issue you'll probably face is code-access security restrictions, since your app will run by default in the Internet security zone, which is probably less than ideal. You can get around this by creating a caspol script to trust your app to whatever level you require. See this link for some info on that (although he has the "-pp" statements backwards in his example): http://weblogs.asp.net/mreynolds/archive/2003/07/11/9975.aspx Read up on smart-client here: http://msdn.microsoft.com/visualc/using/deploying/smartclient/default.aspx

Good idea - I had actually done something like that recently but had also done it the other way I suggested a while back. Probably easier to use the callback.

If you're doing a synchronous thread, then you can pass you label/progress-bar (whatever) variable into the thread's constructor and store it in some global variable in your thread class. You can update that now local variable to whatever you want and the resulting changes are seen in your form. Let me know if you want more detail.

Depending on how you're using your services (ie which platforms you want to support - java, .net, etc), you could try to use a dataset for your return and just create a datatable of your 2d array info. Another approach, to keep with standard data-types, would be to use a string and use two delimiters to determine what was a column and what was a row - say the column delimiter is "," and row delimiter is ";": value1, value2, value3; value1, value2, value3; Then your caller would have to use a Join function to reassemble your strings into an array, but at least this is a supported type for web-services to return.

Based on how the VS.Net IDE generates its code, we've gone with this approach - each line item is a region, and we generally create sub-regions within each to keep like-code functions together. This approach has helped our developers new to .net get up and running and understanding things a lot faster than just tossing code where-ever the developer feels appropriate. Also, we have code-reviews to enforce our standards, which has helped keep things more uniform in our code. Win-Forms: UI Field Declarations Global Declarations (constants would be a sub-region in here) Public Methods Private Methods Protected Methods Windows Form Generated Static Methods Event Handlers Any other type of modules - business-logic, data-access, etc will follow a similar approach, minus the forms-specific regions in the above. All in all, this approach is working out well for us.

It would help to know where the classes are instantiated from. Does Main create the instances of the other two classes, or does creator instantiate taker? Does Main care what's going on between creator and taker? If not, I'd probably just instantiate the stream object in creator and have creator create an instance of taker and pass the reference to the stream to taker's constructor.

Have you tried looking at the samples on the crystal website? http://www.businessobjects.com/services/support/default.asp

Check the Code Library or use the Search feature on the site to dig up what you're looking for - there's tons of good info on this site. If you can't find it here then google it (check google groups if the web search doesn't yield what you're looking for).

Run your app, or even in designer - hit alt+PrintScreen (or just Printscreen) key (that copies the image of your current window to your clipboard). I usually use alt+PrintScreen since it just copies the currently active window and that's usually what I'm looking for. Just PrintScreen copies your whole desktop image. Then open up Paint/Photoshop, whatever and hit Paste (ctrl+v) in a new document. From there you can crop, edit whatever you want.

Turn off Autopostback on those controls that shouldn't be causing a postback.

It's really similar to how you'd to it with a web.config file. Just add an application configuration file to your project, and start filling in the blanks. The easiest way is to add the following: <config> <appSettings> <add key="DataSource" value="server=localhost;user id=someUser;password=somePassword;database=someDB"/> </appSettings> </config> Then in one of your classes in the project - I'll do this in VB.net: Imports System.Configuration public sub DoSomeWork() Dim sqlConn As New SqlConnection(ConfigurationSettings.AppSettings.Get("DataSource")) end sub

Note - making the assumption you're on a Win2K machine with this explanation: You can set the IIS timeout to longer than the default (20 minutes) through inetmgr. When you right-click your web-service web and hit properties, you should see the "Directory" tab displayed. Hit the Configuration button and then go to the "App Options" tab. There you can see the Session timeout setting (probably set to 20 minutes, unless your server has been modified from the default settings). Up that to say 40 minutes (to be safe) and give your test another shot. The other methods you mentioned may also fix your problem (web/machine config file, etc), but I've had luck with this sort of thing by changing this setting.

GUI Bloopers by Jeff Johnson is pretty good. It's more web-centric, but the general discussion really applies to any user-interface design. A good web-site used to be http://www.webpagesthatsuck.com, but I don't know if that site is still around or not.

There's been discussion on this (and other Test Driven Development - TDD) over at http://www.testdriven.com/modules/newbb/viewforum.php?forum=6 Basically, from what I've read, this falls into the Mock Object realm, one that I'm not all that well versed in personally and wish to change. I looked into this, but for the size project I was working on, the mocking of 30 interface screens seemed a bit too time prohibitive (since we're just getting into TDD here). If you're looking to just automate some tests (pre-record then be able to re-run) on a GUI, something like WinRunner may do that trick for you, but this is a different approach than TDD.

I'd try something like this: SELECT userAct.username, userAct.password, userAct.role, userAct.email, userAct.studentcode, userAct.SignDate, tblUrl.URLValue FROM tblUserAccounts as userAct INNER JOIN tblUrl as URLs ON URLs.username = userAct.username WHERE username = @Username

Add this before your if statement starts - see if that fixes it: cboItemId.DataBindings.Clear()

Try specifying your return types: <WebMethod()> Function GetEmployeeList() as System.Data.DataSet Me.daEmployeeList.Fill(Me.dsEmployeeList) Return Me.dsEmployeeList End Function

Try creating a new Access file and re-create your table (with defaults to start - then tweak your columns)/stored procs in that file. That's the way I got it to work.

Now this is weird - I created a new Access database with one table like your, and then added your queries. I didn't apply any security, etc to the file, and everything ran exactly as I expected. Now, when I started to model my table after yours, as soon as I set the DomainName.Required = Yes (from default of No value), I got the error you got. So, I deleted the table and re-added the same layout, just setting the column names and data-types (left all else as default).

I got a quick chance to walk through the code and it seems like your datatable is the cause of truncation. I put a breakpoint on: DOMAINS.Update(ds, "DOMAINS"); and in command window: ? ds.Tables["DOMAINS"].Rows[1].ItemArray[1] "ODO1S" which should be "ODO1SIMULATOR" I'll see later if I can narrow this down some more - perhaps if you create a schema for your data-table it would work?

That's probably a better approach, especially as I re-read that there's updates a lot. Good call.

Okay then, that answers my first questions. Now, did you interrogate the data-table before you issue the Update on your data-adapter to see that the values there are as expected?