You know, I would almost think my work’s done if only the project manager would exclaim to me:
I say almost because my work would be done if the project manager were tearfully exclaiming it. And a burly ex-military sort.
For your consideration: The Benefits of Pissing People Off.
Although the author is talking about leadership and building, not QA specifically.
In QA, you cannot ultimately do your job without pissing people off. Granted, you can get a job in QA and hold onto it for a good long time with a get-along attitude, but you’re probably not helping to release software of any better quality than if they didn’t have you sucking up a salary.
A recent presentation illustrated the rigorous quality assurance that goes on at Facebook:
Perhaps the most interesting and revealing aspect of Robert’s talk was the discussion of Facebook’s somewhat unique development process. At the surface it appears to have the contradictory goals of: minimizing down time, scaling fast, and extremely quick code updates. First, Facebook developers are encouraged to push code often and quickly. Pushes are never delayed and applied directly to parts of the infrastructure. The idea is to quickly find issues and their impacts on the rest of system and surely fixing any bugs that would result from these frequent small changes.
Second, there is limited QA (quality assurance) teams at Facebook but lots of peer review of code.
All right, they’re looking at the code and then jamming it into production as fast as they can. And they think that is a virtue.
I’m a frequent user of Facebook, and I know how often it’s buggy. Almost daily I encounter error messages or, more infuriatingly, the damn thing just doesn’t work. I click Share, and the controls go away and the link/witty status/photo of my cute kitty is not actually shared. It drives me nuttier.
So why do I use it? Because it’s free and because it has reached a tipping point where a lot of people are using it. So Facebook, in those regards, is sort of like Microsoft, except they’re not writing anything as complicated as an operating system for disparate hardware configurations and interoperating applications plus applications themselves. Also, Facebook makes little pretense to quality.
I have no native affection for Facebook nor brand loyalty. If something better comes along, which could be something as simple as something similar that works right more frequently, I’ll hop ship and leave Facebook and its bogus applications designed to get my mobile number for billing behind.
I take it back; Facebook won’t be the next Microsoft. I think Microsoft products work well enough to keep me off of Linux or Macintosh.
And by “scientists,” I mean “Scott Adams and his readers”:
g33klady sends me word of an idea whose time hasn’t come, but it’s here anyway: Bad Code Offsets.
I have never written a bad line of code.
When I tell people that, they often scoff and offer replies like “so you’re not a programmer then?” and “let me guess, you’re a coding deity or something?” Well let me say, I am a programmer and I am not Codethulu, but in the same manner that Al Gore can fly around the world in a private jet without polluting, I have negated my bad code footprint through the purchase of Bad Code Offsets.
Nice if you’re selling coupons, but keep in mind the chief difference between carbon offsets and bad code offsets: Gore forgives, QA doesn’t.
I’ve noticed for some time that the advertisements that Hotmail serves up have leaked memory, impeded performance, and thrown normal JavaScript errors, but the one I got yesterday took the cake:
An error so egregious that Internet Explorer was dumbfounded. Now that’s an error.
A system failure bringing you down?
An FAA computer malfunction involving access to flight plans caused nationwide flight cancellations and delays Thursday.
The excuse?
Some flights were more than two hours behind schedule. Airports around the South also reported delays and cancel source of the computer software malfunction was a “packet switch” that “failed due to a database mismatch.”
Yeah, a packet switch caused by a solar flare, no doubt.
That would never happen in real life.
Article in the SD Times: Imagination, process failures doom software projects:
Scores of well-publicized software failures have taken a toll on careers, lives and resources, yet projects continue to fail at an alarming rate. Top programming experts, though, say that there are commonalities to these failures that, if avoided, can help organizations achieve greater success.
Sixty-eight percent of software projects were either challenged or failed, according to The Standish Group’s 2009 “Chaos” study, representing a “marked decrease in project success rates” for the year. Challenged projects are defined as being late, over budget, or having less than the required features and functions.
The root causes of failure are not difficult to identify. “Most cases of failure that I have seen have been in two categories: imagination and process,” said Grady Booch, chief scientist of software engineering at IBM Research.
Common process problem areas that were cited by the experts include requirements failures, failure to validate and verify requirements, failure to adhere to architecture, lack of contingency planning, failure to learn from mistakes, and the absence of best practices in developing conversion software.
An implementation of a requirement that seemingly is correct might actually be incomplete. For example, a Pennsylvania man could not receive a driver’s license after a computer system indicated that he was dead, because programmers did not include any way to “resurrect him,” said Richard Riehle, a visiting professor at the Naval Postgraduate School, a U.S defense institution in Monterey, Calif. that provides education and research programs for the United States and its allied forces.
Oddly enough, if you’ve worked QA long enough, you see these same sorts of failures over and over again. If you’re like me, you feel a little like Cassandra without the eventual comfort of the sacking of Troy to put an end to your misery. However, when every project starts becoming challenged, usually within the first couple of days or weeks or approaching the first deadline or milestone, people in the team start throwing best practices out and begin doing the cheap, ill-conceived shortcuts that lead to the problems.
There’s your failure of imagination. Failing to imagine something that works.
The Tell-a-Friend feature on the Dole Salad Super Slider contest offers a single field: friend’s e-mail address:
This works because the screen displays after you’ve entered your e-mail address and name as part of your contest entry. But what happens if you click Submit before you enter a friend’s e-mail address?
The error message mentions a friend’s first name field which is not present on the form.
The application sends the message when you enter an e-mail address, so the friend’s first name field is not actually required.
I’ll leave it to your imagination whether this represents recycled code, a late-breaking design change, or an assumption on the part of developer.
What I can tell you is that someone did not test this application thoroughly.
Remember, each of these applications is its own entity, and if you’ve done the same thing a million times before for other clients, you need to treat each application as though the others did not come before them. Your assumptions and oversights might be more glaring and showstopping than this simple problem.
Newspapers, on the other hand, are a different story. Regret the Error is a blog that collects some of the most amusing periodical corrections and flogs a book of the same name.
I think it would be awesome if developers had to issue corrections like this. Actually, I think they’re called release notes accompanying patches, but they never include a sense of responsibility for releasing the bugs in the first place.
If you want to forward a Holland America e-mail, it takes you to the Web interface where the developers have split the difference on the proper use of the possessive:
One could try to argue, as the developer probably would, that Friends name is plural because there are multiple edit boxes. However, that would apply to the e-mail address boxes, too, and Friends name as a plural should be Friends names.
Developers are so amusing when they try to manufacture grammatical rules to rationalize their mistakes.
Apparently, the United States government’s mortgaging arm is advertising on Facebook. Badly:
Keep that in mind, I suppose, when you’re participating in the program. A couple stray unrecognized characters cut-and-pasted here and there could probably blow your whole credit rating.
If you’ve QAed any number of marketing e-mails, you’ve seen this problem before: someone uses another e-mail as a guide and just puts new text in over the body, leaving a number of fragments of the old e-mail behind.
Like this one from Information Week.
Here’s alt text left behind from the renewal e-mail from which they copied this referral e-mail:
And this little bit of copy with a date and deadline for renewing and then the threat of a $200 subscription if you miss it:
You know how you combat this?
But these are best practices, not most practices.
When you sign into LinkedIn.com, note the things they got right with the redirection page:
Never mind that, let’s talk about what’s wrong. The image is sized incorrectly, and it might not be propagated across all servers since I sometimes get a broken image icon when I logged in.
Also note I did not test the click here button, so I cannot vouch for its working properly.
The little redirect pages appear in a large number of applications and Web sites, sometimes appearing only for a scant matter of milliseconds before the user is whisked away to paradise, or at least the content he sought.
That’s why these little pages are often overlooked in testing and, let’s be honest, in design and development. Someone jams a bit of text and maybe a hyperlink up there and off it goes. However, you ought to take a little look at it to make sure the words are spelled correctly and that the links are valid. You can do this by:
It’s in your Web site or application, so you’ve got to make sure they do it right.
You know what happens when you try to delete that entry? All hell breaks loose:
You know how you find that error? You just test. It stems from a basic oversight in hooking up the database to the interface, where the application lets you delete something you’re not supposed to delete.
You should never see this sort of stack trace. It leaks all sorts of insight into your database structure.
You should never see this sort of stack trace. But you probably will.
The login form on the My Toyota Web site has a login form for the narrow-minded when viewed in IE:
Enter a wide enough string, and the form pops over to the right and cannot pop back over to the left:
You’d think they’d have looked at in in IE. Well, you would if you weren’t a seasoned QA professional who expects that the designers and developers looked at it in their preferred cool browsers and didn’t care what it looked like to the hoi polloi who drives Toyotas and uses Internet Explorer.
It’s a scientific study, but from Australia:
Bad moods can actually be good for you, with an Australian study finding that being sad makes people less gullible, improves their ability to judge others and also boosts memory.
The study, authored by psychology professor Joseph Forgas at the University of New South Wales, showed that people in a negative mood were more critical of, and paid more attention to, their surroundings than happier people, who were more likely to believe anything they were told.
That’s how it seems to play out in the IT workplace, is it not?
ICANN is going to allow Web domains to use different alphabet sets:
Web sites written in Russian, Korean and other non-ASCII characters soon will be able to have their addresses displayed in the same language.
Testing on 11 pilot sites with internationalized domain names (IDN) could be completed by the end of the year, according to Kim Davies, spokesman for the Internet Corporation for Assigned Names and Numbers (ICANN). The URLs being tested in Arabic, Chinese, Greek, Japanese, Russian and a handful of other major world languages don’t cover all possible character sets, but broadly represent most countries, Davies told The Standard. Only http://, which is automatically added by browsers, is displayed in English.
Swell. That means you need to handle Web site addresses like this in your Web applications wherever you let users enter them as parts of profiles or whatnot:
Additionally, you might need to store and send e-mail to addresses like this:
I don’t want to make too big a deal of it, but this fundamentally will alter the way your applications handle e-mail addresses and URLs, so you’d better start thinking about this now and plan for whether your organization is going to explicitly limit data entry to Latin characters or if you’re going to re-engineer your applications and maybe even your databases to accommodate the new standard.
For example, say you’re forced to register at FT.com to read this story about the new change. Try to enter an address in a foreign language, and bonk:
Because if you don’t think about it now, eight weeks after the new standard is in place, your C*O is going to be freshly returned from some international conference and will discover his nice counterpart from China cannot enter his e-mail address to receive your company’s newsletter, and you need to fix it now.
You can find out more about this change at ICANN’s wiki on it.
Good luck, and may your antacid be with you.
You are currently browsing the QA Hates You weblog archives for November, 2009.
