Blog: A Bad Couple of Days

I’m home in Toronto for a day after several weeks of helping people learn to test software, and as far as I can see, the whole Web is screwed up. Here are some of the things that have happened in the last 48 hours or so.

  • A fellow on Twitter told me about an interesting Skype bug: send the string “http://:” (no quotes), and Skype hangs. Fpr me, it did more than hang; I was unable to restart Skype. I tried to update my Skype client; this was blocked by an error 1603. I tried uninstalling Skype; 1603. I tried using the Microsoft Fixit tool, which repairs corrupted Windows Registry entries (apparently some aspect of update, uninstalling, or trying to reinstall Skype leads to corrupted registry entries); still an error 1603. I went through the Registry myself, removing all the references to Skype I could find; still no joy. Eventually I was able to download a complete installation package, which finally worked; I can only speculate as to why.
  • I tweeted about the bug, copying Skype Support on the tweet. In reply, Skype support claimed that the initial problem, the http://: problem, had been fixed. (This claim appears to be true.) “We have already addressed the issue 😀 … Just update your Skype and you are good to go :)” I replied, “1603” to this. Skype Support responded, “Take a look at this Community post – maybe it will be of help”. (It wasn’t, really.) I eventually replied, “Yes, I’m reinstalled. Question, though: why a community post rather than an official, researched one from your organization?” The answer was “The Skype Community is quite adept at offering Skype Support, particularly for errors such as this that are not very common.” My reply was, “Fair enough… but why would the official support channel not be so adept?” Or at least, that’s what my reply would have been, but…
  • Twitter has blocked my account, apparently due to suspicious activity. That’s not what the iPad or iPhone clients say, though. They simply say, “Sorry, we weren’t able to send your Tweet. Would you like to retry or save your Tweet to drafts?” From the mobile Web client, I was able to enter a Tweet and apparently have the client accept the input, but no Tweet appeared, and no error message appeared either.

    Eventually, with a little detective work, I was able to determine that my account was blocked due to suspicious activity. The Web site advised me to reset my password. There are two ways to authenticate yourself: one is to have an email message sent to your registered address. I chose this option twice; the email never appeared. I opted to have a code sent to my phone via SMS. The code appeared almost immediately.I entered it into the Password Reset page. I was startled to see the message “We couldn’t find your account with that information.” I repeated the process twice more with the same result.

    On the fourth time, of course, I was told that I had tried to reset my password too many times, and that I would have to wait an hour. I waited an hour and ten minutes, and tried again. My account remained locked. A button offered the opportunity to “contact support”. This isn’t exactly a means of contacting any human support person, but a set of pages offering suggestions and some options for troubleshooting.

    Eventually I found a form that (apparently) affords a means of contacting people. The confirmation page noted “We are usually able to respond within a few days, but some issues may take longer. Please check your email inbox for an email from Twitter Support.” But remember… if Twitter had attempted to contact me via email, that didn’t work.

  • Someone contacted me for help via LinkedIn. I went to LinkedIn’s internal email client, and started a reply to him (I’ll anonymize his name here). “Hi, Jules…”, I typed, and press the Enter key. The Enter key had no effect. End of attempt to reply. Some software change has been made since the last time I answered a personal message on LinkedIn. Did anyone try this after the change was made?
  • Amnesty International sent me some mail on a campaign they’re running, which I agree with. I clicked on the “Take Action” button, which took me to a form soliciting my name and address, which Chrome’s auto-fill settings supplied automatically. Upon pressing Send, an error message appeared: “PAF violation: Insufficient address data.” My address data was entirely correct; I’ve used this page and that address dozens of times before. I left the browser window open. About half an hour later, I clicked on the button again, without having changed the data. The site accepted the data this time. I don’t know why it failed, and I don’t know why it started working again.

At the end of the day, these things aren’t life-or-death problems. What worries me more is that software being developed for life-or-death contexts may be developed in something like the same way. That software—because it is subject to a value-destroying problem based on a single misplaced bit—often may be just as fragile, just as unreliable, just as vulnerable. It frightens me that the company developing the Google Car is the same company that developed Google Buzz. It annoys and frustrates me that organizations that supposedly provide a service have chipped away at the idea of live, real-time customer support.

So, I’ll see you on Twitter. Eventually. Maybe.

Want to know more? Learn about Rapid Software Testing classes here.

5 responses to “A Bad Couple of Days”

  1. I too have same concern, whether product is tested (if its not peculiar issues to user) – or just releases with updates ?

  2. Thomas Toth says:

    Working in the health industry I can see that not everything is tested that way. Some less, some more.

    We’ve recently introduced SIL levels (safety integrity levels) for stories so that the business can indicate and differentiate the different levels of risk associated with certain functionality. Based on the risk the V&V team then decide how much testing is required to cover that functionality. So some stories get less test coverage, some get more.

  3. Clive says:

    That seems pretty typical to me, there’s not many weeks I don’t run into some kind of issue like that with a major site. Commercial but not what you think of as immediately public facing seem to be the worst, like sending a secure message to the bank or trying to pay your major socal utility bill but finding IE isn’t supported (with no indication until you try chrome or Firefox)…seems like that’s the future.. ask college kids how some software they use works.. they can’t tell you, they just get it to work in some fashion and that seems good enough to them, as that’s what their used to. The last time when i got the null pointer issue paying my phone bill the payment didn’t go through, but the most recent time i got the null pointer issue it processed just fine!

    Then when you contact support it’s your issue not theirs. (Smile)

  4. JCD says:

    It appears the blog commenting system thought I wanted to submit while I was typing … and I lost all my text, even when I clicked back. Very on topic, sadly. So rather than rewriting it, I’ll just say it is nice to see other people noticing a trend I wrote about back in Jan of last year: http://www.stickyminds.com/article/why-testers-need-get-used-change

    – JCD

  5. Greg Mooney says:

    The Skype error still happens with the quotes as well. Actually, I was able to restart (close the window and relaunch) Skype without the quotes but when I used the quotes, I had to end the process tree.

    Michael replies: That’s interesting. If it happens after you’ve updated, that’s really interesting.

Leave a Reply

Your email address will not be published. Required fields are marked *