HTML(Context Sensitive) 

is usually preferred for several reasons.  One of the 
reasons is because it is "context sensitive" , i.e.each 
mouse click

recorded in the script is made in the context of the 
previous step/page.  In order to maintain context it looks 
through the 

previous page to make sure the current operation is 
available.   This adds extra overhead because it is 
performed by a 

background utility called the runtime parser.   Although 
the parser uses extra overhead, it does saves us work in 

correlation and page checks. This because if a link/request 
is not available, the runtime parser will see that and 
throw 

an error telling us that the request is not there. (The 
old "Requested Form Not Found" error).   URL mode, on the 
other

hand does not check the context of the previous request.  
It will simply make each recorded request individually 
regardless 

of whether the request is still available on the page or 
logically follows in context.  That's why URL mode requires 
more in

the way of page checks and correlation.

This is the incorrect answer.

Its always a best choice to go for HTML mode.

 Url based scripting generates lengthy script and too many 
dynamic values when compared to HTML mode

When u use internet explorer during recording it use HTML 
format, but when u select URL(uniform resourse locater)from 
load runner window it generates script 4 URL.

HTML based recording:
1.It will capture only HTML based resources.
2.URLs are developed based on each action performed by the 
user.
3.Resource=0 for all the URLs generated under HTML based 
recording
4.Functions developed under this are
  i.web_url
  ii.web_link
  iii.web_submit_data
  iv.web_submit_form

URL based recording:
1.It will capture all the Non-HTML bases resources in the 
application i.e, Jpeg,bitmap etc
2.URLs are developed based on the steps performed by the 
user.
3.Resource=1 for all the URLs in the developed script.
Note:Resource=0 is also some times being displayed in the 
script
4.Functions developed under this are
  i.web_url
    ii.web_submit_data
  iii.web_concurrent_start
  iv.web_concurrent_end

HTML:-Using this recording mode LR can generate script by 
performing certain user actions on only HTML context. This 
mode recogneze the user inputs either in the form of data 
or in the form of forms.For each action that we perform on 
the application, LR generates a function called  Web_url. 
Along with this 4 more functions will generated (i.e)
depending on actions we perform.These are

web_submit_data
web_submit_form
web_link
web_image

Generally URl based recording mode will be used on Non-HTML 
resources like Bitmap,Jpeg,gif..., In this mode LR 
generates the function web_url for each and every step that 
we perform on the application. this  mode recognize the 
user input data only in the form of data submitted 
function. some more functions will generates such as

web_submit_data
web_concurrent_start
web_concurrent_end

please correct if i am wrong.

HTML based recording records only the user actions 
where as URL based recording records the requests and 
resources for the server.

HTML
Context Sensitive based script recording

 HTML-Based / describing user actions is the easiest to maintain

For example, assume there is a link called "Continue" that
sends you to http://localhost/ index.html. Your script
records "Continue", and NOT the destination URL. So, if
sometime later your application developer decides to change
the URL associated with "Continue" to something else - like
http://localhost/ home.html, your script will still work
correctly - and go to "home".

LoadRunner parses the HTML it received from prior server
responses to determine what the full URL is behind the
Continue link(i.e. Context Senistive).

LoadRunner does as much work for you as it can, leaving you
with the fewest script statements to maintain.



URL:
No context would be considered during script recording,
every statement get recorded independently

Every thing would get recorded as URL only. i.e. web_url

In this case if URL based recording is used then script will
fail during replay.

URL mode records every URL that LoadRunner is supposed to
ask for –but without the actual parsing.

So, if a URL is there, LoadRunner simply calls it but does
not examine it for embedded URLs. 

This allows(i.e. Not parsing embedded URLs) would make LR to
run very FAST.

But this type of script would become obsolete and you may
not even know it if some thing changes in the application
hence this is very much useful for a stable

In html recording mode each action has one url 
eg:i have take one iogin form in that login form  we can 
enter user name and pwand click on ok this is one tranction 
having a one url
in url recording mode each step has one url
eg login form un having one url pw is one url and click on 
one url