on 32:00, I think your statement is wrong about the current automation tools failing to recognise the change in text. It completely depends on the automation tester what attributes they have chosen to identify the object with. if they identify with text, it is a poor choice the tester has made. the tool should not be blamed for that.
Hp QTP had this feature called smart identification almost 15 years back and people prefer not to use it since it could lead to false passes. Now just because you want to use the buzz word AI, that wouldn't make it safe to use. Moreover, it is just a basic algorithm not sure why they call it AI.
Good try. But accuracy is confusing me with difficult dom. If there is change in code, should be change in test automation code. Thats what code deals with ryt? Code to code?
This is not a real application of AI what these companies are showing. AI is just a buzz word for these people to market. Let me explain how this works - They have an extension that captures all the html tags and attributes based on these attributes they have a logic of building xpaths on fly. For example of Xpath 1 doesnot work it tries to build xpath 2. This is very simple, there is no application of AI here(if someone have knowledge on AI) its just a logic to build xpaths. The main problem with this is it takes a lot of time if there is a no such element exception since it creates xpath based on trial and error.
This is not a part of AI. This is just a logic to generate Xpath using different strategies. Don't buzz the word AI in testing. It's not yet implemented yet in Testing.
Start at 25:00, if you have background.
on 32:00, I think your statement is wrong about the current automation tools failing to recognise the change in text. It completely depends on the automation tester what attributes they have chosen to identify the object with. if they identify with text, it is a poor choice the tester has made. the tool should not be blamed for that.
37:40
What are optimum weight statements in this context? I understand what are optimum weights in an optimization cost functional.
Hp QTP had this feature called smart identification almost 15 years back and people prefer not to use it since it could lead to false passes. Now just because you want to use the buzz word AI, that wouldn't make it safe to use. Moreover, it is just a basic algorithm not sure why they call it AI.
I recall spending hours troubleshooting an issue which went away eventually by disabling the "smart" identification;
Does this concept of dynamic locater strategy works for complex xpaths?
Are there any open source data sets available for same please.?
Amazing presentation skills.
Thanks for this video! It is useful.
Good explanation
Good try. But accuracy is confusing me with difficult dom. If there is change in code, should be change in test automation code. Thats what code deals with ryt? Code to code?
Can i know if desktop apps also can be automated by AI?
Great T-shirt BTW. Where can I get one?
QTP did this 15 yrs back..... called Smart Identification.... AI maybe did it a bit better but you don't really need AI for this.
This is not a real application of AI what these companies are showing. AI is just a buzz word for these people to market. Let me explain how this works -
They have an extension that captures all the html tags and attributes based on these attributes they have a logic of building xpaths on fly. For example of Xpath 1 doesnot work it tries to build xpath 2.
This is very simple, there is no application of AI here(if someone have knowledge on AI) its just a logic to build xpaths.
The main problem with this is it takes a lot of time if there is a no such element exception since it creates xpath based on trial and error.
This is not a part of AI. This is just a logic to generate Xpath using different strategies. Don't buzz the word AI in testing. It's not yet implemented yet in Testing.
Of course, AI is present in Automated Testing. Here, the NN is responsible for identifiying changes but not to crash a test
i pity the cameraman :P
Learned nothing helpful about ai for testing here this is way too simple
Blah blah blah! Meaningles discussion
What a wierd accent ..
really ?
He’s trying to fake it
@@mikemihayAccent is not important. Key concepts are important in which they explained. Awesome explanation.
Somebody help him to speak Login not lagin. Compromise not Campramise. CHU!!!
I think for u english is knowledge.!!! In ur life have u gave any demo