In general, how much of a project timescale would you say goes to testing and bug fixing? Roughly.
I’ve been quoted ~30 days development time for a project that we’re having done, but there is an additional 24 days for test and debug. Personally I think that allowing an additional 1/3 of the development time for testing and bug fixing is reasonable (ie an additional 10 days in this case) but nearly doubling the timescale from 30 to 54 days is way too much.
I appreciate that it depends on the project, but what do you guys find? I’m looking for general, rough figures, off the tops of your heads. No need for any evidence to back up any claims. In this particular case it’s PHP/MySQL work, but it shouldn’t really be relevant.
What do you think?
I think its 100% dependent on the project. Nothing should be set in stone when it comes to this. In its production environment, what is the turn over in time to launch any bug fixes. What is the cost of a failure once its in production?
These are fair points and of course are considerations. We’re trying to keep control of a very tight development schedule. We simply can’t afford for it to take 54 man days (we only get 15 a month so we’re looking nearly 4 months for a simple RESTful API implementation). It feels gratuitous, quoting 54 days, even though it is a maximum
I’d ask about agile processes, continuious delivery and stuff like that – no reason to separate the phases aside from a final layer of testing at the end.
That’s a tough one to answer. In my personal experiences, I’ve had bug testing completed in a matter of a couple of days. Other projects had bugs ongoing for weeks. I guess it all depends on the development team that is working on your project and how capable their skills are.
Yeah it really does depend. A small admin panel managing user data/user access may take a few days to test. If however I’m testing on a project that involves credit cards and PCI compliance, it could take weeks. Of course if the project is larger, it could take longer depending on the level of security involved and the feature-set requested.
Well, there is no such thing as “Testing Enough”. There’s always bug somewhere no matter what software you write. You can only fix bugs with given time. IMO, it depends on priority. Most of my project is “Deadline over Quality”. I never worked on a project that’s “Quality over Deadline”…quite sad but oh well. If you’re based on Quality over Deadline then I advice you to Test Driven Development. This will ensure that you tested 100% code coverage and constantly build automated test suites. Also, put time for Code Review as well.