Re: This is a total php newbie question [message #175686 is a reply to message #175685] |
Wed, 19 October 2011 19:12 |
Doug Miller
Messages: 171 Registered: August 2011
Karma:
|
Senior Member |
|
|
On 10/19/2011 12:22 PM, Bill B wrote:
> On 10/19/2011 11:27 AM, Jerry Stuckle wrote:
>> On 10/19/2011 7:21 AM, Bill B wrote:
>>> On 10/18/2011 6:42 PM, sheldonlg wrote:
>>>> Here is a scenario:
>>>> 1 - User does this and gets a page with a link to a download which does
>>>> the download.
>>>> 2 - He doesn't click the that link on that page until 25 hours later.
>>>> Note, he has not closed his browser or that page.
>>>>
>>>> How are you going to know that more than 24 hours have passed? You need
>>>> to do something that will be similar to a timed logout.
>>>>
>>>> Something to think about.
>>>
>>> I do not know the answer to that specific question, but it raises, for
>>> me, the question of how much extreme testing ought to be done to test
>>> the limits of code and browser behavior.
>>>
>>> Is there a generally accepted standard, does it vary by circumstance,
>>> does anyone have personal practices that have served them well?
>>>
>>> Bill B
>>
>> You test enough to ensure your code works properly. One of the reasons a
>> database will help you tremendously.
>>
>
> Jerry I was asking in the context of practices that you, for example,
> might use that would, for you, embody what "test enough" means. I design
> serious games that use computer simulation as the underlying engine. The
> decision a player is responding to might expect an answer between 5 and
> 10. For testing I would answer 500,000 just to see how the engine reacts
> to the number.
Test boundary conditions and invalid input. If the answer is supposed to
be between 5 and 10, I might test using these values in addition to the
obvious 5 thru 10: "A", " ", "5A", "5 ", " 5", 4, 11, 105, 0, -1, -5,
-10, 2^16, 2^16 + 5, 2^32, and 2^32 + 5.
|
|
|