I have (again) some questions about the ActiveX Demo.
Is there any kind of limitation in the Demo / the registered product ? (other than the popup)
I try to use this ActiveX. I can get a IMetaTreeX interface, load and view a tree without any problem.
Nice product indeed.
I cannot access the iMtItem neither the iMtItems : it seems - according to the OLE/COM object viewer from Microsoft - that these interfaces doesn`t exist at all - or at least there are not registered. Is this a n ActiveX demo limitation - or my way of dealing with interfaces ?
Is the document in the Demo, the up-to-date documentation for the ActiveX component ? For the registered product, do we get some extra documentation or activeX source ?
Is the .mt file format documented anywhere ?
Thank you for your answer
The .MT file format is very simple - it is just a text file where nodes are separated with # symbol.
Thank you for your answer.
Some ideas / remarks about your product ActiveX Component MetatreeX
Ghost Label (tooltips ?) Refresh :
launch the demo version and set the window to a medium size. Center the window on screen. Use a dark desktop background. When the mouse hovers over the labels, then artifacts appear.dissapear on the left side of the screen.
It looks like a draw/redraw is beeing made as if the label refresh would occurs on a left copy of the metatree - out of the demo window. Or is this the InfoTip refresh ?
The same behaviour occurs on 3 different computers.
Let use Visual C++ 6.0 or Visual .Net and their ActiveX Container test. Load the component and invoke the method loadFromFile to feel the meta tree with the sample example.mt : This time again, one can observe the same behaviour of a shadow-refresh for the label. It is possible to make thid refresh problem more obvious, using the `slow drawing` option of the ActiveX Container Test Wizard : ghost copies of the refreshed labels stay on screen at the bad places before beeing cleared.
Uneasy to center a node just dragging it to the center.
The Visual Studio .Net, and especially the VBasic system doesn`t load the project included in the demo. It exhibits some conversion errors.
In the demo, a double click opens or close a node, but also centers a node. The twice action is disturbing : I think one of the action should be linked to the right button instead or a long press of the left button.
I know there is a distance between node parameter, but what do you think of a zoom parameter that won`t change the distance between node but would provide a way to shrink the whole graph ? This could also be associated to the wheel. In this view, a set_height, set_width method for the tree would be handy, as well as a set_x, set_y for the whole tree.
What do you think of a lock/unlock feature that would gives the user the capability to move a node where he wants -
or to switch to automatic node placement ? Sometimes, even if the automatic placement is ok, the interface reains cluttered.
Anyway, without a good documentation support, the user has to dive into the tlib infos, that is really not acceptable as it. There is a real lack of documentation for this component. A c example with a simple tree building using MTItem would be very very handy at least. How do you get an instance of MTITem and how do you specify it should be the root or not ? A simple exploration of the ocx/tlib and register database shows that the MTItem is not really registered. The class description in the TLIB seems to be a dummy class, making it more confusing for the programmer.
The simpliest way of providing the interface of construction for MTItem and MTITems would be to provide extra metatreex methods that builds objects of these classes - instead of using separate classes : this would allow a simple straightforward use of the single component metatreex and this ActiveX could - this time - be used in any language understanding ActiveX : only one component CLSID would be required. This would really simplify the use of MetatreeX (my opinion).
With a better documentation and a more simple (and reliable) ActiveX implementation, your component could really be a real winner, standing between startr** and tr**bolic.