The online community for software testing & quality assurance professionals
 
 
Calendar   Today's Topics
Sponsors:




Lost Password?

Home
BetaSoft
Blogs
Jobs
Training
News
Links
Downloads



Testing Tools >> HP / Mercury WinRunner

Pages: 1 | 2 | >> (show all)
p_g_in
Member


Reged: 05/25/04
Posts: 19
Keyword driven testing
      #93467 - 08/10/05 09:30 AM

Hi,
Actually before posting my query , I did some search and found lot of articles on Keyword driven testing. But frankly couldnt clear my concept.
Could anyone of you please let me know what exactly does Keyword driven testing means andcan it be used in WR?


Post Extras: Print Post   Remind Me!   Notify Moderator  
bpolitzerModerator
Moderator


Reged: 06/25/02
Posts: 2053
Loc: Cupertino, CA, USA
Re: Keyword driven testing
      #93468 - 08/10/05 10:00 AM

Keyword driven testing involves creating an API for non-WinRunner users (manual testers). Basically you write functions for all business processes in your application. You then create a framework where each Business process function (including paramaters) is identified by a "keyword" inside an Excell spreadsheet. The top layer of the WinRunner test loops through the spreadsheet and looks for these "keywords", then calls the corresponding function. You then instruct the manual testers how to fill out the spreadsheet to create different test cases. The idea is to separate the WinRunner expert from the business process expert.

As I stated in my previous response, I'm not in favor of this approach. My experience has been that the manual testers will not go through the effort necessary to learn the API. The WinRunner code quadruples in size and becomes unmaintainable. I'm sure there are many who will disagree with me, but that's my opinion.


Post Extras: Print Post   Remind Me!   Notify Moderator  
p_g_in
Member


Reged: 05/25/04
Posts: 19
Re: Keyword driven testing
      #93469 - 08/10/05 10:21 AM

But the functions we create are in TSL?And what do you mean by top layer of WR?

Post Extras: Print Post   Remind Me!   Notify Moderator  
bpolitzerModerator
Moderator


Reged: 06/25/02
Posts: 2053
Loc: Cupertino, CA, USA
Re: Keyword driven testing
      #93470 - 08/10/05 10:36 AM

Yes the functions are in TSL.
When I say top layer, I am referring to a call stack. It might look something like this:

code:
if(ddt_val(table,"BusinessProcess") == "Login")
{
login(ddt_val(table,"Username"),ddt_val(table,"Password"));
}
else if(ddt_val(table,"BusinessProcess") == "Create User")
{
create_user(ddt_val(table,"Username"),ddt_val(table,"Password"),ddt_val(table,"Functional Group"))
}
else if(ddt_val(table,"BusinessProcess") == ...)
{
...
}



Post Extras: Print Post   Remind Me!   Notify Moderator  
p_g_in
Member


Reged: 05/25/04
Posts: 19
Re: Keyword driven testing
      #93471 - 08/10/05 10:43 AM

Oh ok.So manula testers are not aware of the functions(TSL) .They will be just given the above script to run. right?and how do we assign function to excel sheet?
In the above script..Businessprocess is just a word given to a function right?


Post Extras: Print Post   Remind Me!   Notify Moderator  
bpolitzerModerator
Moderator


Reged: 06/25/02
Posts: 2053
Loc: Cupertino, CA, USA
Re: Keyword driven testing
      #93472 - 08/10/05 11:02 AM

quote:
Oh ok.So manula testers are not aware of the functions(TSL) .They will be just given the above script to run. right?
Yes, that is correct.
quote:
and how do we assign function to excel sheet?
You simply invent a "keyword" for each business process function and write a document (for the manual testers) describing what each does.
quote:
In the above script..Businessprocess is just a word given to a function right?
Yes this is correct, although the keywords and the function names do not necessarily need to be the same.

Take my advice, and don't implement this - you'll be sorry.


Post Extras: Print Post   Remind Me!   Notify Moderator  
p_g_in
Member


Reged: 05/25/04
Posts: 19
Re: Keyword driven testing
      #93473 - 08/10/05 11:43 AM

thanks a lot.It really helped me.But what according to you are the drawbacks ?

Post Extras: Print Post   Remind Me!   Notify Moderator  
Tony Davis
Veteran


Reged: 03/23/00
Posts: 3082
Loc: Orange County, CA
Re: Keyword driven testing
      #93474 - 08/11/05 06:02 AM

I've gone down a mid route with keyword driven testing - you still need to go through the steps of actions wanted for the process (like for log in you still need to launch the app, feed it the user and pw, and click the submit) but given functins that are at a level anyone can prety much convert their manual tests to automated without any knowledge of WinRunner specifically. The functions we use now (dropped 2500+ to under 30 after getting rid of the person who "had to have it that way" [Wink] are in the format of:

code:
f_enter(ps_activity,ps_taskorwindow,table.xls,log);

and f_bclick(ps_activity,ps_taskorwindow,ps_button);

with a few special ones here and there. The activity and task or window variables are related to how the application is navigated through functionally, but 99% of all data entry into any window is done with the same function and a lokup table behind the scenes that defines the window that needs focus, the gui file associated with it, and any supporting look uptables for the mapping of the user datasheets to the objects available on the window.

Pretty much a couple of buffer layers that allow quick creation of new window functions and if a GUI changes, or the way a control needs to be interfaced with, I only have to change it in one place and it filters out to all scripts.

They can write their "super functions" at the beginning of the script to use the functions I provide if they want to break it down into a business keyword flow.

For my case, maintainability has been greatly lessened and the ease of use increased for those that want automated tests but don't or won't know WinRunner/


Post Extras: Print Post   Remind Me!   Notify Moderator  
bpolitzerModerator
Moderator


Reged: 06/25/02
Posts: 2053
Loc: Cupertino, CA, USA
Re: Keyword driven testing
      #93475 - 08/11/05 08:20 AM

quote:
Originally posted by Tony Davis:
For my case, maintainability has been greatly lessened

This is exactly why I don't like the keyword driven approach. I think the best practice is to train the WinRunner expert to also be a subject matter expert on the AUT. Scripts and functions should be simple enough that a new scripter can easily pick up and maintain where the old left off.

Post Extras: Print Post   Remind Me!   Notify Moderator  
Tony Davis
Veteran


Reged: 03/23/00
Posts: 3082
Loc: Orange County, CA
Re: Keyword driven testing
      #93476 - 08/11/05 11:36 AM

Can't disagree at all.

I think I have this place down to almost "best of both worlds". For me, the WinRunner "expert" in house, I get to work on the coding and solving problems with how WinRunner reacts with the application(the fun stuff) and I work with the inhouse Bus Analysts and Complaince group to design and document test data that will exercise the system in the best way possible.

I have another new person in the group that is here to do nothing but create automated scripts from the data that is documented from above. They are able to follow the steps and automatically convert the descriptions into automated steps. Makes them happy.

And then there are the manual testers that have actions they have to repeditively do for each section of testing they perform. They can create the automated test steps about 90% of the time to create their own auto test scripts when needed. All the data is delivered to WinRunner through formatted spreadsheets designed within the framework we have set up.


Post Extras: Print Post   Remind Me!   Notify Moderator  
akulkarni
Member


Reged: 02/17/02
Posts: 25
Loc: Pune,Maharashtra,India
Re: Keyword driven testing
      #93477 - 08/15/05 10:44 PM

Hi,

I have taken the route described by Tony, I have not yet completed entirely the work that I am doing.

Few questions that come to my mind at this stage:

1. How to use Exceptions (Popup etc).
2. How much performance can you get out of the Suite. i.e. in 12hrs how many integration testcases get executed ?

Your comments appriciated, I am working on these two questions; First one is pretty simple (I have defined some exceptions and have a Sheet in TestCase where you can set them to "Yes" or "No").
I need some fresh Ideas on this as This involves set_excel_sheet leading to increase in FileIO.

For the second one I am not having enough TestCases done yet so cant get figures to compare / budget.


Post Extras: Print Post   Remind Me!   Notify Moderator  
BTTester
Member


Reged: 03/18/02
Posts: 171
Loc: Twin Cities, MN, USA
Re: Keyword driven testing
      #93478 - 08/16/05 08:40 AM

quote:

...
1. How to use Exceptions (Popup etc).
I would code it and make it a re-usable script.

2. How much performance can you get out of the Suite. i.e. in 12hrs how many integration testcases get executed?
Your question is not clear, it's like asking "In 12 hours, how many miles can I run?". I would recommend you to structure your test/test cases around the requirements and utilizing the creation of use cases.



Post Extras: Print Post   Remind Me!   Notify Moderator  
cbueche
Member


Reged: 02/28/04
Posts: 120
Loc: Germany
Re: Keyword driven testing
      #93479 - 09/13/05 03:21 AM

>Could anyone of you please let me know what >exactly does Keyword driven testing means andcan >it be used in WR?

Hi

A keyword driven Test is for winrunner available by using a Framework that support this kind of environment for free:

http://www.cbueche.de/welcome_to_emos_group.htm

Take search and you will achieve a lot of examples how to implement it. As before descibed it helps to bring non programmicle person in a state to doublicate standart buisnescases for different purposes. By just copie the meanng blocks with teststruktur in excel and fill it out. The goal is the maintabilty that is better when only one function support the steps during execution for all different tests with the same or different navigation on the API. so if there are changes, only one palce have to be changed. And not for any kind of action a different script.
All functions that will be need have to be laoded as compiled modul. And will report to the result all necessary info by wrappers over the ususal funktions from winrunner. I.E edit_set(); just edit. FRM_edit_set(); will tell what is set or what is inside the api depends on info from speadsheet how to handle. And also generate testdata with this one function. Whitch object is used from what table and and whitch function will also be shown.

So this function can be part of a block that has more wrappers to navigate and manipulate. This block will be called from winrunner depending on what info is in the spreadsheet that execute this block in order.

Kind regards

Carsten

http://www.cbueche.de\index4.htm


Post Extras: Print Post   Remind Me!   Notify Moderator  
Peter RuscoeModerator
Veteran


Reged: 03/18/02
Posts: 7686
Loc: Tampa Bay
Re: Keyword driven testing
      #93480 - 09/13/05 06:23 AM

Your "alert" was a little slow picking up this time, Carsten!

Post Extras: Print Post   Remind Me!   Notify Moderator  
cbueche
Member


Reged: 02/28/04
Posts: 120
Loc: Germany
Re: Keyword driven testing
      #93481 - 09/13/05 07:17 AM

I was on Hollidays! Sorry about ;-)

Post Extras: Print Post   Remind Me!   Notify Moderator  
cbueche
Member


Reged: 02/28/04
Posts: 120
Loc: Germany
Re: Keyword driven testing
      #93482 - 09/19/05 04:36 AM

More info from Wilson Mar's page

http://www.wilsonmar.com/datadrvt.htm


Test Automation Frameworks
This page compares and describes "data driven" and "keyword" driven automated tests using open source libraries and WinRunner from Mercury Interactive the hardest won (but easily forgotten) insights not obvious from available documentation


Topics this page:

1 Why Frameworks

2 Driver Frameworks

3 Infrastructures

4 Driver Test Specs

5 Summary

6 Flow Diagram

7 SAFS/STAF Install

8 WRSAFS Install

9 WRSAFS Customization

10 Keywords & Functions

11 Fitnesse/Abbot Fitnesse/Abbot

[ 09-19-2005, 04:46 AM: Message edited by: cbueche ]


Post Extras: Print Post   Remind Me!   Notify Moderator  
winrunner_new
Member


Reged: 09/20/05
Posts: 52
Loc: India
Re: Keyword driven testing
      #93483 - 09/21/05 01:52 AM

can any one explain wat is keyword driven testing

Post Extras: Print Post   Remind Me!   Notify Moderator  
bru
Super Member


Reged: 05/08/03
Posts: 1383
Loc: Austria
Re: Keyword driven testing
      #93484 - 09/21/05 04:51 AM

There are several members of QAForums that can explain keyword driven testing. And they already did that in some other threads.

Therefore please use the search functionality at the top right corner of this page!

Regards,
Juergen


Post Extras: Print Post   Remind Me!   Notify Moderator  
bpolitzerModerator
Moderator


Reged: 06/25/02
Posts: 2053
Loc: Cupertino, CA, USA
Re: Keyword driven testing
      #93485 - 09/21/05 11:28 AM

quote:
Originally posted by cbueche:
More info from Wilson Mar's page

This looks frighteningly complicated and unmaintainable:


 -

No Thank You.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Peter RuscoeModerator
Veteran


Reged: 03/18/02
Posts: 7686
Loc: Tampa Bay
Re: Keyword driven testing
      #93486 - 09/22/05 12:45 AM

Now that's MY kind of diagram. Not like that wimpy SDLC thing. THIS one will separate the sheep from the goats.

Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1 | 2 | >> (show all)



Extra information
0 registered and 4 anonymous users are browsing this forum.

Moderator:  Peter Ruscoe, bpolitzer 

Print Topic

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled

Rating:
Topic views: 8613

Rate this topic

Jump to

Contact Us | Privacy statement SQAForums

Powered by UBB.threads™ 6.5.5