Since testcases use logical names, if the object's actual name changes on the current GUI, on another GUI, or localized version of the applications, do we have to change just the tag in the window declarations or do we need to do some more changes? (in Silk Test)
are there any disadvantages of TAGs in GUI/Web based testing?
[This message has been edited by vdevara (edited 10-18-2000).]
I think the answer to your question is YES.
I am a bit confused on your lingo .. so perhaps a simple example for all.
The logical name is the identifier you refer to in your script. Once you decide what it should be (or take the system default) you should attempt to keep that logical name for "all times", in other words for goodness sake don't change it ever!!!
Now if some crafty developer decides to change the name of the control from "activate" to say "energize" you don't change the logical identifier .. you change the physical description or tag itself .. like so (i.e lets use an html push button):
Then you save those changes in your Frame file and viola .. all your OLD scripts work just fine! Because the OLD identifier will map to the correct control via the NEW tag .. no matter how many referances you have.
This is the true value to tagging. You can map controls to immutable (non changing) identifiers, and when the caption or some tag description changes .. you only change it in the one place. Then all your old scripts which may have hundreds of referances to that control can still run!!Preety neat huh??
As for disadvantages .. or gotchas, like everything else they are there. I would like to see more tag forms supported (especially a window id for us reverse engineer types).
Also there are two schools of thought to
"multitags". I like them but I remember a funny subject title in this forum "multi tags are evil".