Replies

I had a similar problem about 2 years ago. Although in my case it was more then annoying. When you have a domain, where you have to log onto your computer, once it has booted, that ‘ghost’ image sits behind the logon prompt too. It’s a pain. In my case, I accidentally right clicked a file, and instead of hitting ‘save as’, I hit ‘save as wallpaper’. Let’s just say that image wasn’t for children!

What I did to get rid of it, is I just did a simple search for that file and deleted it. No more problem. I probably could have gotten away with ‘renaming’ it, or moving it, or I could have taken more time and searched the registry for it. But it was just quicker to delete the file.

What I believe is happening, is I remember back in the days of when I was using 98 (yeck phewy ), I had found the registry setting that let you set a picture as your ‘splash’ screen for either the bootup or the shutdown. I think what is happening, is that when you save an image as wallpaper, it just automatically uses that file to fill all three positions, and when you pick another image, it doesn’t replace it for the splash screen options.

If I wasn’t heading out the door, I’d goto regedit.com, and look for the registry keys that cause this…..

How are you ‘deleting’ an item from the listbox. Are you building the listbox values with a Value list, and then rebuilding that value list through code? What I would recommend, is after the delete process, just set the listbox value to an index of 0, to have it jump to the first value. (You will have a problem with that if you can end up with NO values in your listbox.).

Does that sort of answer your question?

Whoever told you that images had to be in the root directory should be burned at the stake…..or at least forced to repeat ‘I know nothing about websites’ 1,000 times.

A webserver is a fileshare PLUS. When you open a word document, you can link pictures into it, right? Do you need to have the images in the same folder? Of course not. Files are found using what is called a URL, or Uniform Resource Locator. Using a URL allows you to specify EXACTLY where a resource can be found.

Webservers are even slicker. Let’s say you have index.htm in the root. You then have a ‘virtual folder’ called images, where are actually on another machine, in another folder. If you put an image into index.htm, and set it’s source to “imagesMypicture.jpg”, the webserver sees ‘images’ translates that into the path of the virtual folder, and retrieves the appropriate file.

Another item to toss into the mix. Why are you FTPing files to the webserver? If it uses Front Page extensions, you should be able to just drag and drop the files into the file view of FrontPage. In fact, that is how you SHOULD transfer files. Why? Because FrontPage keeps track of where files are. You can build a page, and set images into that page, put links to other pages etc. At some time in the future, you may move an image to another folder, or move a page which is linked to a new folder, and Frontpage will change the links on all of your pages. If you go behind FP’s back, and just upload files, you aren’t giving Frontpage the chance to keep up to date. FrontPage extensions are really just ‘files’ on the server, which Frontpage uses to keep track of your site. There is no ‘thinking’ on the server, it is done within your local session of FrontPage. Change the files without FrontPage, and you are heading for a site that is going to be, shall we say, ‘unfriendly’ to FP.

Quick question. Are you globbing all of your files into the root folder, or do you have subfolders for various files. Having your entire site within the root may cause problems, even though theoretically, it should work fine. It is similar to any folder/file structure. I have seen quite a few Windows 2000 profiles go corrupt, and in every case, the user’s profile is loaded with documents (In their ‘My Documents’ folder). We typically set people up to have a ‘data’ drive or parition, and we recommend for people to put their data files there. You can even ‘point’ your MyDocuments to that drive/partition/folder. I bring this up because Windows 2000 is a rock solid OS, but it get’s quirky when it’s over loaded.

A Web Server is just a fancy file share. Thus, it could very easily be susceptible to the same ‘quirkiness’. Depending upon the server’s OS, you may easily run into the 8.3 file format issue (8 character filenames, with 3 character extensions).

So, you should immediately split your site up into subfolders (not necessarily subwebs).

If you have already done this, and you are still experiencing these problems, a new subweb may help, but you are going to have to talk with your Hosting company to determine how to setup a subweb. A subweb is a ‘virtual directory’, that can also have FrontPage extensions placed on it. In this manner, a subweb maintains it’s own FrontPage settings (Theme, navigation, etc.).

You may also want to talk to your hosting company about killing the FP extensions on the end completely, and then reinstalling them. It would require work on your end to ‘rebuild’ your settings, but you may not be ridding yourself of corrupt files by just ‘resetting’.

If you are running your own IIS server, I can give you more detailed advice, but from your post it sounds like you are using a hosting company.

Hope my babbling helps…..

Explain what a fragmented data structure is? Okay, here’s Drew’s tyrade on Relation Databases. ….just kidding.

A relational database allows you to structure your data so that you do not repeat data. For example, if you had an Address list, and you had a family of 5 with the same address, instead of making 5 records with the same data, simply make 5 records with Names only, then put the addresses in a seperate table, linking the addresses with an ID field from the names table. You can then query your data so that all 5 names show up with the same address, but the address data is pulled from one record. There are many advantages to this method of data storage. Size for one, the less ‘repeat’ records you have, the smaller your database. Also consistency is nice. If the family of 5 moves, you only need to change one record instead of 5. As soon as that one record is changed, the address for all 5 people is automatically ‘correct’.

That example explains a semi-fragmented state. You have split the data across multiple tables, however, a simple query will pull up the data into one recordset. You can fragment your data even further. For example, in my original post’s attachment, I describe a ‘product’ database. Instead of a ‘prebuilt’ product table, the structure is designed to allow for virtually any type of product. Products are grouped by a generic product type. That grouping puts products with the same ‘specifications’ together. In other words, you may sell cars, which would have specifications such as Engine size, Paint Color, Interior Color, Transmission type, etc. You may also sell Air Freshners, which would have specifications of shape, size, scent, color. These are two seperate ‘product types’, but you can list multiple cars under the ‘car’ type, and multiple Air freshners under their type. The table structure I explain in my attachment shows how to record both products in the same tables.

To do this, there is a specification table, which lists the specifications for all product types. The product type is a field, so you can retrieve the specifications for each product type on their own. For the ‘values’ of the specifications, there are 5 tables (Currency, Date, Long, Double, and Text), which hold the data. The specification type determines which table holds the data for that specification.

Is it possible to use this type of data structure without classes, yes, certainly. Is it a pain to do so? Absolutely. Using classes to store and retrieve the relevant data makes front end development a lot easier.

Beware, reading my example will take you that much closer to the dark side of the force…….unbound forms!

It’s a resource issue. If you have ever programmed in VB, you may have used a TreeView control. In fact, within VB you can quite easily build a mock Windows Explorer. There are a few issues though. First of all, Windows explorer ‘refreshes’ itself from time to time. Open two instances of Windows Explorer, go to the same folder or drive, and add a folder. It will show up in the other instance, though maybe not immediately. Also notice that the ‘order’ may be changed. New files may show up at the end, even if they should be sorted differently.

Refreshing Windows Explorer will fix sorting issues and missing/file folders from the right pane, but that is because it is just a ‘listbox’ of files/folders in a file/folder, so it simply has to rebuild it’s list.

The left pane, however, is a tree view, where each drive/folder is a node (or branch) on the tree. It’s not as simple as refreshing a listbox, more complex logic must be applied to rebuild the tree. That is why it does not refresh right away. The reason that it shows up on the left pane, when you open that folder from the right pane, is because whenever you arein a folder on the right pane, it is shown as the ‘open’ folder on the left pane. When it does this, it forces the logic routines to determine if the left pane has built a ‘node’ for that folder, and if it hasn’t, then it rebuilds that treeview.

Just make a macro (I know, I hate macros too, but it’s a quick solution). Start the macro with SetWarnings and turn warning off. Then put OpenQuery statements for each query. The last argument should be another SetWarnings to turn them back on.

What is it doing wrong? If you are getting a ‘big’ delay, that is because I thought the MetaTag refesh worked for values under 1 second. It doesn’t, it’s either a full second, or an instant. So just set the MaxHundredths to something under 100, and it will just ‘type’ as fast as your system refreshes.

Another option would be to use client side scripting, and I am VERY leary of using that on internet web pages. (I use it all the time on Intranet pages, because we have a controlled environment, where everyone is using a relatively current version of IE).

Ooops, the MetaTag refresh apparently has a low setting of 1 second. Anything smaller is an automatic refresh. The effect still works. You could code within the asp to make a quicker random refresh.

Here ya go. This .asp code will do what you are asking. The page is refreshing, just have a function that writes the actual remainder of the page and replace a call to that function for this line: ‘Response.write “I will now show you the page:”‘ In the following code:

(Note: Replace the value for MaxHundredths with the maximum number of hundredths of a second that you wish to wait between characters)

<%
dim intCurrent
dim strTextToDisplay
dim intRandom
Const MaxTenths = 10
strTextToDisplay="Accessing Databank… Please wait…"
intCurrent=request.querystring("XYZ")
If len(intCurrent)Len(strTextToDisplay) Then
response.write “” & vbcrlf
response.write “” & vbcrlf
response.write “” & vbcrlf
Response.write “I will now show you the page:”
else
Randomize
intRandom=Int((MaxTenths – 1 + 1) * Rnd + 1)
intRandom=cint(intRandom)*.1
response.write “” & vbcrlf
response.write “” & vbcrlf
response.write “” & vbcrlf
response.write “” & vbcrlf
response.write “

” & left(strTextToDisplay,intCurrent) & “

”
end if
%>

An update query is what you are looking for.

However, I would suggest a different approach for your co-worker. Instead of having the actual text from the dropdown list go into your table, have it record a key.

For example, here’s the ‘list’ table:

1-Dog
2-Cat
3-Mouse

Then, here’s your data table:

Rover 1
Cuddles 2
Squeaky 3

If later, you change the text for Dog, to Canine, you don’t have to change the data in other table, because the reference, through the key of ‘1’ remains the same.

Drew

Okay, I am attaching the whole kit and kaboodal. The zip file has the .dll, and it’s source files. It also has sitemap.asp which is an example ASP page that uses the .dll. It’s a direct copy of what I have on my companies site, with the Site path and URL changed.

Let me know if ya’ll have any questions. One item to note, if you have more then one root node, this code won’t work (off the runway). It can be modified to handle that. I figured not many people use multi ‘root’ node navigation.

Also, here is a little code you can use in Front Page itself, to dump the navigation out to an excel file, with a little modification, this can be used to write an HTML based site navigation. (The asp version would be real time, requiring no modification after the initial setup. But not everyone can install an ActiveX .dll on their hosting machine.)

Dim strTemp As String
Dim exl As Excel.Application
Dim exlSheet
Dim intRow As Long
Sub ExportNavigation()
Dim Web As Web
Dim rootNav As NavigationNode
Set exl = New Excel.Application
exl.Visible = True
exl.Workbooks.Add
Set exlSheet = exl.ActiveWorkbook.ActiveSheet
Set Web = Webs(“http://yourwebsite.com&#8221;)
Set rootNav = Web.HomeNavigationNode
intRow = 1
LoopChildNodes rootNav, 1
End Sub
Function LoopChildNodes(nav As NavigationNode, intDepth As Long)
Dim nd As NavigationNode
exlSheet.Range(Chr(intDepth + 64) & intRow).Value = nav.Label
intRow = intRow + 1
For Each nd In nav.Children
‘strTemp = strTemp & nd.Label & vbCrLf
LoopChildNodes nd, intDepth + 1
Next
End Function

Drew

It is dynamic. It needs to see the _fti_pvt folder, which has the structure.cnf file, which is the file that front page stores the navigation in. A lot of hosting companies that host IIS servers, do let you install ActiveX controls. Either way, the code explains how to read the file, so you could do it straight in ASP if you wanted.

I’ll post it later today.

Drew

Sorry about that. In 97, when Access ‘disappears’ you should have a form on your desktop. In 2000, it’s a version issue. Access 2000 handles windows a bit differently. In either case, Access is still running. I have a module that hides/shows the Access window, and I put it in the MiniCalendar to show how to use it. There is a AutoKeys macro, where if you hit Ctrl-A, Access will show (and repeating that will hide it again.).

One tip, a lot of sample databases have initial startup code. The MiniCalendar hides Access, however some are live dbs, that someone posts a portion for sample usage. A lot of times in those cases, there is startup code that fails, since either part has been disabled, or it is looking for a network resource that doesn’t exist. You can bypass any access databases startup routines by holding the shift key down. You can disable the ShiftBypassKey through code, for your own applications (to prevent users from bypassing your startup stuff), but I have yet to see a sample database that had the ShiftBypassKey disabled.

Actually, zipped, its pretty small, so I’ll post it. I removed my smtp servers address, you’ll have to put that in. Also, ‘review’ the code to see how it works, it’s a pretty bare bones utility I wrote.

Drew