|
and/or
- cancel dialogs
- close app (with click on "X") while db is in edit mode
- kill app with task manager while someone is editing something
- use special - non ascii - chars where it is not expected
- simulate high load with endless read/write operations
- change screen size/resolution to unusual values
modified 14-May-24 15:45pm.
|
|
|
|
|
It comes down to my biggest beef with tests and that's being only as good, generally, as whoever wrote what they're testing.
For one, some jabroni didn't really write tests for your system. They wrote some documentation.
If you can't run them in an automated fashion - without manually hunting records down and manipulating data? Then they simply aren't tests and if the person's task was to write tests, they 100% failed that task.
You write the test so that it stages the data you know will work and then delete that data post-test.
[OneTimeSetup] / [OneTimeTearDown] are the decorators for NUnit to do stuff like that.
When we started writing a bunch of tests, we settled structuring things linearly like the following with regions for:
Class Members
Assemble/Setup
Act/Invoke
Test Asserts
Cleanup
|
|
|
|
|
Get some experienced testers to bash it hard - much better than test scripts IMHO - writing(and testing / maintaining test scripts) is often harder than coding the application
In a closed society where everybody's guilty, the only crime is getting caught. In a world of thieves, the only final sin is stupidity. - Hunter S Thompson - RIP
|
|
|
|
|
For at least 30 years, we have had/practiced the "5-year-old robustness test":
Put the kid in front of a keyboard and a screen, and tell him: Do anything you want. If you make it lock up, and can show daddy/mommy what you did, you'll have an ice cream cone!"
I know of a few ice cream cones awarded that way 
Religious freedom is the freedom to say that two plus two make five.
|
|
|
|
|
I like it.
"A little time, a little trouble, your better day"
Badfinger
|
|
|
|
|
Brilliant! I need to offer more ice cream cones to my QA team.
Religious freedom is the freedom to say that one plus one plus one does not always make three.
Scientific freedom is the freedom to say that parallel lines in a non-euclidean reference system may shake hands.
|
|
|
|
|
Maximilien wrote: In reality, I have to do about 20 different steps to get there and go through thousands of records (SQL) to find one record that will work.
Why not edit a single record so it will work?
I'm currently doing a project where we're rewriting an old VB6 application that used to use a dBase database.
It's SQL Server now, but everything is not what you'd expect (everything is char(x), even bools, which have "Y" or "" (and sometimes "N"), except some ints, which are int or decimal(x,y)).
Lots of magical numbers and strings too.
Some fields even are like "X X " in which the first X means "option 1 should be used" (whatever option 1 is), the empty second and third position means those options are off and the fourth X means option 4 should also be used (the software only uses four options, even though it's a char(10), also it's not always X, sometimes and option can be, for example, "S" for small or "L" for large, etc.).
Anyway, finding a row you can use and writing a WHERE clause is a pain, so I just edit the top row so it'll fit my test/use case.
|
|
|
|
|
1) Mind internationalization. Is your app translated? In how many languages? Serious work to do if this number is greater then 10.
2) How does the GUI looks like on a native Chinese (Traditional or Simplified)/Japanese/Korean PC?
3) Input long texts in various languages, with many specific characters. Special care when comparing and processing texts - in which culture?
4) Did you specify a culture when installing the app DB? If not, the SQL server default one will be used. Try different cultures/collations etc.
5) Input/use numbers in various cultures, alternating decimal/thousand separators etc.
6) Input/use datetimes in various cultures.
7) Set the date of the PC to 29th february in a leap year and test app functionality.
8) If you read text from disk/stream, create test files containing binary data (e.g. 0x1A/EOF - can your app read text beyond that byte?). Or create test files that have \r\n or only \n as end of line etc.
|
|
|
|
|
When I was working on an ActiveX control embedding a video stream into a web page (a long time ago when others were still working to prevent millenial apocalypse), we had a customer who used the "coffee cup" test - put a coffee cup down on F5 so it would continually refresh. Extremely harsh!
|
|
|
|
|
Maximilien wrote: Seriously, do you know of any good white paper on how to write good test cases ?
No, but I can list what I do:
1. Write use cases. What are users supposed to do with the system? From use cases, build test plans which not only test the system but teach the testers how to use it.
1A. Ideally most test cases start with new records and build upon them. For cases where that is not feasible, construct test records. Create scripts that will insert or revert the test records. If you have to hunt for test records, do it just once.
2. If you have role-based authentication and authorization, conduct each test in #1 for each role.
3. Most applications have pre-defined choices (radio buttons, drop down lists, list boxes). Test every possible combination.
4. This one is the hardest -- make as many totally illogical choices as possible. Try to do things out of order. Paste 10,000 characters into a textbox. Enter a string of special characters only. Save while missing mandatory inputs. Basically the "five year old" part.
5. Ask the end user testers to do as many weird things as they can think of. Trust that they will think of things you will never consider.
I'm no longer surprised at how much basic error handling is NOT built into applications, e.g., the DB field is 20 characters but the text entry field is unlimited. #4 will identify things like that.
|
|
|
|
|
Wordle 1,060 4/6
π¨β¬π©β¬β¬
β¬β¬π©π©β¬
β¬β¬π©π©β¬
π©π©π©π©π©
|
|
|
|
|
Wordle 1,060 4/6
π¨β¬π©β¬β¬
β¬β¬π©π©β¬
π¨β¬π©π©π¨
π©π©π©π©π©
Bad third guess.
|
|
|
|
|
Wordle 1,060 3/6*
π©β¬β¬π©β¬
π©β¬β¬π©π¨
π©π©π©π©π©
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
β¬β¬π©β¬β¬
β¬β¬β¬β¬π©
β¬β¬π©β¬π©
π©π©π©π©π©
In a closed society where everybody's guilty, the only crime is getting caught. In a world of thieves, the only final sin is stupidity. - Hunter S Thompson - RIP
|
|
|
|
|
Wordle 1,060 4/6
π©β¬β¬β¬β¬
π©β¬β¬β¬β¬
π©π¨π¨π¨β¬
π©π©π©π©π©
An L of greens π.
|
|
|
|
|
Wordle 1,060 3/6
π¨β¬β¬β¬β¬
π©β¬β¬π©β¬
π©π©π©π©π©
|
|
|
|
|
Wordle 1,060 5/6
β¬β¬β¬β¬π©
β¬β¬β¬π¨β¬
β¬β¬π©β¬β¬
π©π©π©β¬β¬
π©π©π©π©π©
|
|
|
|
|
Wordle 1,060 4/6*
π©β¬β¬β¬β¬
π©β¬β¬β¬β¬
π©β¬β¬β¬β¬
π©π©π©π©π©
Happiness will never come to those who fail to appreciate what they already have. -Anon
And those who were seen dancing were thought to be insane by those who could not hear the music. -Frederick Nietzsche
|
|
|
|
|
Wordle 1,060 4/6
🟩🟩β¬β¬β¬
🟩🟩β¬β¬β¬
🟩🟩β¬🟩🟩
🟩🟩🟩🟩🟩
Ok, I have had my coffee, so you can all come out now!
|
|
|
|
|
Wordle 1,060 3/6*
🟩β¬β¬β¬β¬
🟩🟩β¬β¬β¬
🟩🟩🟩🟩🟩
|
|
|
|
|
In my experience we web devs don't actually use or at least watch a video of a screen reader. This is bad. Very bad. Because we talk about accessibility, but it's amazing how many people still don't use aria tags, leverage semantic elements, etc.
Every web dev should watch this video. It's only 4 mins long.
Screen Reader Demo for Digital Accessibility
Not trying to sound preachy (but I am  ). The man is spot on. The world revolves around the Internet now. We can't forget about our blind brothers and sisters.
Side note, one of the cool features (as it pertains to the skip to content thread below) is the screen reader shown already has the ability to jump to a header. Also notice he skipped over the skip to main content link.
But even outside of that, peeps need to see a screen reader in action at least once. In particular how it allows him to scroll with a focus box over content.
Jeremy Falcon
modified 13-May-24 21:37pm.
|
|
|
|
|
That's a really interesting video and provides great perspective on the issue.
I know that Bootstrap is really good about providing the aria tags as part of their controls, etc. and using Bootstrap helped me to gain some understanding of the importance.
This video was really informative and helpful and made me re-think about how important it is.
Thanks for sharing. 
|
|
|
|
|
raddevus wrote: This video was really informative and helpful and made me re-think about how important it is.
What I find funny is they can already skip to the main content by virtue of headers.  And home dude just strait up ignored that link.
One of the peeves I had watching that video is whoever wrote that web page used spacer divs, images, or something without using an aria-hidden tag. Granted, it's a university and not a professional site, but still. You'll see some of those tabs tabbed over blank space for that reason. Spacers are so 20 years ago, but if you're gonna use one because you refuse to learn Flexbox, at least use an aria tag to tell the screen reader what's up.
raddevus wrote: Thanks for sharing.
Any time buddy.
Jeremy Falcon
|
|
|
|
|
To be clear, I'm totally agreeing with you about the skip to main content link you talked about being implemented like garbage. It seems like something that would otherwise be a good idea was borked... and to top if off not even needed.
Jeremy Falcon
|
|
|
|
|
A couple of hints to developers:
* Get hold of a pair of sunglasses (unless you want to do a cleaning job on your own glasses afterwards ), and smear them with vaseline or something similar. Carrying those, go through an entire procedure of booting up your PC, reading and responding to emails, running your application with all its features, and shutting the PC down.
* For handling tunnel vision: Make a large cardboard with a small hole, at most 1/8 of the screen width. Hide the screen behind the cardboard, but allow the user to move the boar up and down and to both sides to cover the entire screen, and let users run a boot up, running your application and shutting down.
* For color blindness: Reduce the color saturation to zero, creating a greyscale image (most display card drives allow this), and try to use applications - maybe your own! - using colors to convey essential information. (You may of course use web sites converting the colors according to a specific defect, but those are usually rather slow. Converting to gray scale is certainly not as exact, but it will give a rough indication.)
* Disconnect the mouse (and other navigation tools such as touchpads or joysticks), and go through the process from boot up, starting your application and using all of its features, and closing down, using the keyboard only.
* One nice way of demonstrating how suitable your application is to a blind (or near-blind): Turn the screen to the audience so that you cannot see it, but the audience can. Then demonstrate how you handle boot up, activation of different applications including you own, and shutdown.
* Find yourself a pair of mittens, as thick as possible, and try to handle your keyboard with those.
* Assuming that it is winter, at least 15C below freezing: Go out for at least half an hour without gloves or mittens, and when you come back, go directly to your keyboard and start operating your application.
There is one test that beats them all:
Make your test team include people with different kinds of disabilities: Various sigh impairment (tunnel vision, color blindness, retina defects), various motoric defects, and so on.
A lot of developers who claim that they have adapted their software to various disabilites would (or at least should) feel ashamed when observing how people with those disabilites cope with their software!
Religious freedom is the freedom to say that two plus two make five.
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
