Messages posted by: craig

Hi Matt,
I did come across that snippet that assigns the settable property. Now I see that I should be using input registers instead of holding registers for the points which are read only. I was thinking of input register as "input to the PLC" which would make them writable and holding registers as "output from the PLC".

A checkbox/radio might still desirable in the datasource edit screen for setting the default permissions on a point but using "input register" instead of "holding register" as the register range will sort me out.

Thanks for your assistance,
Craig

Modbus IP.

Perhaps I will start by creating a new dwr method that can return both the alarm data and any new messages. That would allow me to get the obvious issues sorted out without refactoring too much.

Creating a system log type of user message is a good idea, thanks.

Our current web site uses the long-poll method. It only has a handful of users so it is not a problem to have a thread sleeping until new data has arrived to send to the browser. For heavy users of mango maybe having 1 thread per logged on user would turn into a problem? I don't really know how heavy threads are.

Our current website uses the 'register' method you describe. The list of points requested by a JSP page at a specific URL and session ID is stored. When any of the points in the list are updated, the servlet wakes up and sends it to the browser. A bit of javascript updates the appropriate DIV.

For mango I'll dig into views.js so the same callback can handle lists of PointViewStates and WatchListStates (and probably more types I haven't encountered yet)

Cheers,
Craig

Hi,
I've got a variable read from a 2 byte signed register over modbus IP.
It can only take on values of 1,2,3,4,etc. In mango on the data source screen the type of this point is numeric and I can't change it to multistate.

I can use the "Range" text renderer and use ranges of 1-1,2-2,etc so the appropriate text is displayed instead of 1 or 2.

In the graphical view multistate image renderer works as well but some flexibility is lost when the type is numeric instead of multistate. For instance I can't assign the same image to multiple states.

Should I convert this numeric to a multistate with a meta point or is there someway to assign the data type of this point to multistate?

Thanks for the help
Craig

Hi,
Some points, like process variables, are read only. Their value can't be set even if you are super-user.
I see there is the "settable override" in the graphical view and I can assign a user's permission on a point from SET to READ.

With more than a couple of users and potentially hundreds of points assigning the READ/SET property for each user becomes a lot of work!

Could this property be assigned to the point in the data source edit screen and be effective for all users?

If you think this is a good idea I'll get to work on implementing it
Craig

In the situation where more than one person is logged on to mango and potentially operating the same piece of equipment (say they are both responding to the same event/alarm) the two people need a means to communicate with each other.

It is preferable to use mango to pass the messages back and forth so that they can be logged and displayed to other users that may logon later. A bit like IRC. A record of the response to a given event can be created.

If the conversation takes place in an external IM program then the record of the conversation is lost. Voice conversations may also take place but specific operations, analysis, or conditions that need to be recorded could be entered via the "chat" log or maintenance log.

This is a bit different from User Comments:

User comments are attached to a point whereas the maintenance/chat/operator log I am envisioning would be associated with the entire system or mango instance

the messages would need to be immediately displayed to all other users logged on

when a user logs on they should be able to review the log of messages passed back and forth between the other users without having to navigate to the point details.

The messages could be stored in the user comments table.

Should I create a 'dummy' point to associate my maintenance/chat log messages with and then modify the html pages to suit or does anyone see a better way?

Thanks for your thoughts
C

I'm thinking of adding a logging type so that pointvalues can be logged every X (minutes|Days|Months|Years).

Most of the data i've come across in the electricity business is logged every 5,15, or 30 minutes.

It would be useful to have this logging type so watchlists and views could update in near real time without generating excessive amounts of data.

The logging on change by tolerance feature is nice (especially for multistate and binary points) but for analogs it requires a good idea of the amount of noise in every point to set the tolerance properly. Can offer great resolution when required though.

I might tackle this myself. Any input before I start?

It would be nice if the title of the pages in mango would change to reflect what is being viewed. Perhaps it could include the name of the view, watchlist, datapoint or datasource being viewed/edited.

It would make it easier to use the forward/backward history drop downs on the browser and easier to keep track of multiple tabs or windows.

Hi,
When you are using eclipse do you run tomcat within eclipse or is tomcat running standalone and you deploy to the tomcat directory with the Ant build script and then attach the debugger in eclipse?

Thanks,
Craig

- change order in drop down lists like the one in the views page.

I agree it could be easier to find the point you are looking for in the drop down list. The filter by data source option in the edit graphical view page would be handy everywhere that drop down list of points occurs (like in the data point edit GOTO drop down at the top)

Another option might be to display the node's parent folder names before the dataPoint name in the select drop down. This way I wouldn't have to include the location and datasource name in every point's name to be able to tell them apart.

For example if I am monitoring a temperature on several motors I have to call the temperature data point "locationID - dataSourceID - temperature" so that I don't have 10 data points called 'temperature' in the drop down.

I might have got everything wrong but I would like to make the "Graphic views" as superuser and then publish them to other users who can't edit the views. just watch values or change set points, acknowledge alarms and so on.

This would be convenient for me too.

What are your thoughts, inclinations a nd priorities on a non-browser based client these days?

Would you consider the eclipse Rich Client platform as a base?

I'd like to use the http retriever data source to retrieve an image from a webcam or any old URL. Then I'd like to place a scaled copy of the most recent image in a graphical view which links to the full size image or shows it in a pop up layer.

Is this something that you might undertake? If not I'll take a crack at it.

I think I'd start by adding a 5th data type for image files to com.serotonin.mango.DataTypes and taking it from there.
I plan to store the images in a dated directory structure.

Additional libraries or dependencies might be required for image resizing. Do you have any preferences?

Hi,
When I hold CTRL and select multiple points in the point hierarchy page I am able to move the selected points to a folder in one click + drag. The problem is while holding CTRL and clicking the various points to select them I would expect the background colour on each point selected to change, indicating that it has been selected. This does not happen.

Trying to select multiple with SHIFT in IE7 resulted in some strange screen drawing as well, with the inverse of the selection getting a blue background.

Perhaps this is more of a dojo issue than mango?

I have now received an escalation notice. It came out of the blue yesterday when i wasn't working on mango but it happened to still be running.

Thanks for the list of things to check and breakpoint suggestions.

I have all the log level settings in log4j.xml set to DEBUG as I understand this is the most verbose setting but I didn't see anything about email being sent although it does generate a lot of output and I could have missed it.

It might be a couple days before I get back to this but I'll keep you posted
Craig

Hi Matt,
Thanks for testing that out for me.

Yes I have used the test email function and checked spam filters for the message. The inactive notification was delivered to both accounts so I presumed there were no SMTP related problems. I'll keep digging today.