IT Werkz Sometimes

Is scrum a cure for death? Nope but it is mighty, mighty important to some people.

I know one guy who was great at testing but got sacked because he wasn’t any good at scrum. Huh. How good do you need to be at scrum to be any good or no good at it. You can teach people scrum in about 5 minutes as all you need to do is teach them how the daily scrum works, the pre-planning meetings, the post-sprint review meeting and the product and testing backlog. It is spoken about as if it’s the most important thing in the world but it makes no difference to me. I still test software the same way as if I’m on a non-scrum project. I’ve actually had a conversation with someone who’s keen on scrum about whether we do scrum to generate software or we write software to do scrum.

What is good about the version of scrum we practice is that the test conditions – called test confirmations – are tacked onto the back of the user story at about the same time as the coder is doing his bit. I check out the user story and derive the test confirmations and add them to the end of the user story. The coders soon realise that the user stories have got the test confirmations and obviously start reading them and realise all of the grisly tests I’m going to do. This has the effect of them hardening their code so that it deals with my test confirmations, and even better, they start running some of my tests as their unit tests. Talk about a win-win situation. They find most of the bugs that I would have found before I even get sight of the system, so by the time I do get to play with the software most of the bugs have already been sorted out. Of course I still test all the test confirmations but most pass OK.

So what have we done here? My test confirmations are acting as a secondary interpretation of the requirements, almost a confirmation of them and for some reason they are now taken more seriously by the developers. Without the developers seeing the test confirmations, say they were written in a separate document, then their testing wouldn’t have been as deep and they would have handed over a system that worked OK but I’d tear it to pieces.

That part of our scrum variation is worth its weight in gold.

I’ve read quite a lot about SEO recently and one of the things that’s important for good SEO is to get a domain with relevant keywords in the name.

I can understand that OK but if you’re creating a web site for system testing it ain’t easy as it appears that most of the obvious domain names are already registered. If I hunt around enough and mash various testing words together I can find some domains that are available but the names are uninspiring if not downright boring, and they also resemble a lot of other existing domains. So not very memorable.

I was told that domains with hyphens in are displayed on search results pages as without the hyphen so is even more user friendly when displayed. I’ve also read to not use domains with hyphens in as they are harder to remember or to say i.e. www.software hyphen testing dot com. I’ve registered one that was ok and is with the new .co TLD but that might just cause confusion about where I am based.

I could just create a new domain with no relevant keywords in it at all. The good thing about that is that I can register a domain that could be far more memorable but will need a lot more marketing as the nature of the site will not be obvious from the domain name.

It works for Loop11.com and Amazon.

On the PHP/MySQL course last weekend we wrote a very simple page to read some user input – name – and update a database with it. Can’t get much simpler than that except that that was the first time I had ever written any PHP to update a database.

Anyway I gave it a go and using the lecturers displayed example I keyed that in and butchered that to write to my own database. I did manage to get the thing to work. Once. The second time I ran it I got some db duplicate problem and that was caused by me not setting an increment setting on the database id field. Without that being set the second record was trying to be inserted, not updated, with the same non-incremented id.

On Sunday I thought I’d better see if I could get the same code, which I’d saved on a USB stick, working from home. Did it work? Did it heck. The first page was form.php which displayed the name field for the user to enter data into and a submit control. When that was clicked on control should have passed to the addme.php page which in turn called _dbconnect.php, to connect to the database and then control was returned to addme.php to actually insert the data into the database and finally display a confirmation message.

So what happened? I started XAMPP and then Apache and MySQL and the form.php page loaded OK but entering any data and clicking on submit resulted in either an error page or squat happening. Think, think. Oh yeah, I might have MySQL installed and running but I hadn’t created a database like the one on the course. I also noticed that form.php was looking for addme.php on the H: drive, which is where it was on the course PC, and it should have been the C: drive.

I amended that to get it to point at my htdocs folder and created a database with the same name as the addme.php expected. Created the one table in it that it wanted, with a couple of columns and reran form.php. Again nada.

I stuck a (PHP) print_r into the first line of the addme.php page to see if it was even started and what was displayed was ‘array()’. Looked kinda empty to me. Or is that because the print_r() command moves the array pointer to the end and therefore there was nothing to display in the array? Apparently using reset() will move the pointer back to the beginning. Question is what was the form or what parameters did I apply to the print_r command that I unleashed.

Jeez, I’m hacking like a lunatic.