Reputation
857
Next privilege 1,000 Rep.
Edit questions and answers
Badges
4 8
Newest
 Yearling
Impact
~52k people reached

Jul
25
comment Should we reject QA candidates based on basic reasoning and math tests?
I could argue that the average spec is in many ways a mix of word puzzles and math problems. being able to parse that stuff, figure out what is testable, how it should actually behave in a given situation is a useful skill. As is creative solutions and 'out of the box' thinking, troubleshooting, etc. puzzle solving and troubleshooting have a lot in common.
Jul
21
comment What goes in your definition of done (DoD) ?
If you are using proper engineering practices (pretty essential to be successful when building stuff via iterations) then the code as a whole tends to be easier to maintain. (the work done in the 'next iteration' is in many ways exactly like 'maintenance' of the code produced in the prior iteration. If support will need tools etc, the customer should be specifying that as some of the stories they want built during the project, along with things like configurable/customization features etc. In any case, trust me, you'd much rather maintain code produced by an Agile team than a BDUF team.
Jul
1
answered What are the best tools available to test performance of a MVC web site
Jun
27
answered Is drag-and-drop possible in watir-webdriver?
Jun
27
comment Is drag-and-drop possible in watir-webdriver?
The best venue for technical questions regarding Watir or associated gems like watir-webdriver is the main StackOverflow site
Jun
23
comment Testing email generation
I'd known that caps made no difference and periods pretty much didn't exist in gmail addresses (johnsmith is the same as John.Smith) but of course that has a limited number of variations. The numeric suffix is a good trick, I wonder if systems like exchange can be setup to allow that
Jun
23
comment If I didn't use TDD but want to transition to agile, should I go back and create those unit tests?
+1 on the book referral, it cones highly recommened by a few devs I know. I would add in doing a fast search in your code for anything else that might utilize the object/class/method etc that you are modifying, and add unit tests that reflect the behavior expected by those things FIRST (and make sure they pass) before you then write unit tests for the new functionality or bug you are fixing.. that makes a slightly better guard against breaking existing functionality with your changes.
Jun
23
answered What goes in your definition of done (DoD) ?
Jun
23
comment Should unit tests be executed as part of a build?
I would add to the advantages: fewer functionality regressions if someone reverts code by accident, or changes for a new feature are made to existing code and end up breaking existing functionality.
Jun
23
comment Should unit tests be executed as part of a build?
There are means by which developers can tag unit tests as 'pending' if the unit test itself is not complete, or the code that would make it pass is not complete, but the developer needs to checkin their code. The key thing is keep those to a minimum, and make sure that once a dev thinks their feature is complete, there are no tests still in that condition.
Jun
23
comment Should unit tests be executed as part of a build?
it might be better or clearer to say that the 'code can successfully be compiled and built' meaning that process itself does not generate an error, however unit tests (which verify the code is performing the correct behaviors) fail..
Jun
23
comment What is the difference between sanity and smoke testing?
I have also seen the 'sanity' test done as a quick test (in environments without good automated regression suites) ala your description, a random sampling of functional tests across the whole product, to ensure that a recent patch or 'hot fix' didn't break anything. That's a very different thing than 'is it good enough to test' it's closer to 'we don't have time to retest it all, does it look like nothing suddenly started behaving in a crazy way.'
Jun
23
comment What is the difference between sanity and smoke testing?
I've worked places they were the same, and places they were different.. different was rarely the same definition of different.
Jun
23
comment What is the difference between sanity and smoke testing?
Heh you guys are making me remember an old article on new solid state devices that appeared in an april issue of popular electronics long ago (I'm ancient). Three were applicable to this discussion.. SSCG Solid State Confetti Generator (25uF electrolytic 5v cap, placed across 120vac source.. good for one use only. NED the Noise Emitting Diode (Place a normal LED across 120vac source, makes a loud noise, once.) and DED -Dark Emitting Diode, what you have after a NED has been used once.
Jun
23
comment Can exploratory testing replace executing manual functional tests?
Did the people doing the manual testing actually find ALL of the same bugs that were found during exploratory testing plus 5% more? If not then I'd suggest you try another experiment next time, swap the order, do the manual functional tests first, then do the exploratory and see how many bugs you find during exploratory that were NOT found during the manual functional tests. I say this because the way the question is phrased makes it sound like the manual process resulted in finding all the same bugs that exploratory did and then some, which is not what I would have expected.
Jun
23
comment Can exploratory testing replace executing manual functional tests?
I like Ethel's answer a lot here.. The only thing I'd add to it is have you tried to figure out why your exploratory testing missed the 5% that it did, and for that matter did it find issues your scripted tests would NOT have found? I'd be looking at the stuff that was missed, and asking what you can do to improve your exploratory process such that you would have found those missing bugs the next time around.
Jun
23
comment Can exploratory testing replace executing manual functional tests?
For functional testing, creating well done (non brittle, low maintenance automation is generally more expensive than executing the same test manually, often my several multiples. However, once it is created, it's very cheap to run, so run it often if you have it. IOW not cheap to create, but very cheap to use. (the exception being combinatorial esting, where even creating disposable quick and dirty automation beats doing the same thing with small variations thousands of times)
Jun
23
comment Can exploratory testing replace executing manual functional tests?
@glowcoder I'd tend to classify stuff like that as a 'usability' issue (inconsistent UI elements that violates the paradigms that you've used elsewhere make the product harder to use, and not appear uniform and consistent to the user. )
Jun
23
answered Testing email generation
Jun
23
comment Creating test cases for site registration using Selenium IDE
This is a good approach to things like loadtests where you will be creating a lot of accounts over the course of the test and don't want to be deleting stuff or doing any cleanup mid test (and perhaps don't care about stuff like having the email going to a real address). otherwise I'd prefer my tests cleanup after themselves and leave the system in the state they found it to start with.