Thursday, November 26, 2009

Do Not Discard My Data

I just wasted 40 minutes trying to post a comment to an IBM developerworks blog posting from David Chadwick.

That wasted 40 minutes was a flawed attempt to describe the simple changes I think could be made to David's list of things which he believes exploratory testing is "not". I believe with relatively simple wording changes, I could modify each of his "not" descriptions to instead be descriptions of high value, useful exploratory tests. The wording changes are so small that I believe they hint David may not yet understand exploratory testing and how to apply it. However, this posting is not about exploratory testing, it is about a site that lost my data, twice.

My first comment was lost after 20 minutes of writing, thinking, and editing. The comment was lost because I clicked the "Add Comment" link, entered my comment, then pressed the "submit" button. The page which was returned politely informed me that my comment was rejected, and listed several possible reasons for the rejection. Unfortunately, IT DISCARDED MY DATA.

Don't discard my data! It frustrates me as a user and makes me unwilling to return to the site. If I must be logged in to submit a comment, force the login before accepting my input.

I registered on developerworks, logged in, and clicked the "add comment" link again. I added a short, dummy comment to assure that I was now able to add comments. I was able to add the comment.

I wrote a new response (I assume it was a little better than the first comment, since second drafts are commonly better than first drafts). After about 20 minutes of working on that response, I clicked the "submit" link. My comment was rejected again, with the same list of possible reasons for the rejection. Unfortunately, IT DISCARDED MY DATA AGAIN.

Don't discard my data! It makes me feel stupid, and then I need to remind myself that I'm not stupid, the software which should be working for me is instead making me do the work.

I assume the second failure was either due to my inserting URL's into the text, or due to the length of the text I was trying to post. I don't know which it is, and I'm frustrated enough with the developerworks site to not care which it is.

I gave up on trying to post a useful comment. I left a short note that my comments had been rejected twice, and if the author wanted my comments, he would need to send me e-mail.


TestyRedhead said...

Totally agreed. I've had to get in the habit of select all and copy of a long comment to have a backup. If your website makes a plain text editor look attractive or the human use a backup, you've violated user trust in the worst way and deserve flaming.

Maybe they need some actual testing off of the happy path to find these bugs? Or naw, they got "full code coverage" and didn't consider what wasn't in the code or the user experience.

This ends when users stop putting up with it. We must demand better. Thanks for your post. This drives me mad as well.

Cem Kaner said...

IBM/Rational's undoubtedly thoroughly-automated-regression-tested developerworks site allowed you to post a comment.

That' more than I got. Instead, the site gave me uninformative error messages that blocked me from posting comments and sometimes blocked me from even seeing the original blog post. I varied browsers and security settings on the browsers to try to get past this, but that resulted in variation-of-failure rather than actually letting me reliably see the blog post (let alone commenting on it).

Oh well.

Maybe they need to do some exploratory testing...