You have asked so many questions in a row in 3 different forums on 3 different applications, but it looks like they all have the same root problem. I am answering only here, but this answer is for all your questions.
1. The reason why TestComplete doesn't recognize objects inside your application is most probably because your application was not prepared as it required. This is what you need to do:
- first, you need to know the platform/environment which was used for creating your tested application (e.g. Delphi, c++, flash, etc.)
- after that select the type of the application here and read carefully Preparing section (pay attention, for some application types the version is also important, the preparing instructions may differ)
- most probably this will solve all or most your questions :)
2. If your application type isn't listed in the page I mentioned above, ask on the forum here whether or not TestComplete supports it. You'll need to operate with screen coordinates or text recognition only if your application isn't supported by TestComplete.
3. As I understand, you have just started using TestComplete, tried Record & Play for desktop, web and mobile apps, and came here for help. I suggest that you start learning TestComplete by testing just one app type (desktop or web), do not try to master everything in one day. TestComplete has great help system (both internal and online), there are also many videos available, try them first to get a general understanding of how to use TestComplete.
P.S. TestComplete has abilities to recognize text on the screen (link in your case) or search for an image (your question about desktop icon). But these approaches should be used only if everything else doesn't work, they are more difficult and slow. So please, try to follow my suggestions first.
P.P.S. for those, who still remember me - I'm back :)
After you follow karkadil's excellent advice.... :)
It apopear that what you are trying to do for this particular problem is drill into the objects in some sort of terminal emulator... it looks like that "whole object" is some sort of virtual desktop or something. If the contents of that object are being displayed as hosted on another workstation (like an RDP window or a Citrix streamed application), you're not going to be able to get the text of the object you want using the specifics of TestComplete's object identification. TestComplete interacts with the memory of the workstation on which it is running to determine objects in applications... if that "link1" button is actually on another workstation, then TestComplete has no visibility to it. Your best recourse is text recognition or image comparison as karkadil mentioned but, also as mentioned, be aware that this has some pitfalls as either case is dependant upon a very static initial environment.