Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Skill levels also help determine the number of bugs which will be found. Testing on contracts, at medium skill levels, it looked to me to be in the ballpark of 50 bugs per unit of code. You can test this by looking at your contracts history and checking how many bugs the client found (and charged you for) and adding it to how many you know you removed in Beta.
For my own projects, with rather limited, inpromtu testing, it seemed to be a little higher. That was with one person (my founder) doing everything on a 2D Editor and then on an Audio Tool. All of the "relative" skills were at about 70% for my projects. They were both basic tools for the mid-80s, and I took out 350 bugs for the 2D Editor and 450 for the Audio tool. Both projects came out with Great quality. Support team found an additional 500 or so bugs between the two of them (I didn't really write any of this down).
As an aside, you'll see the number of bugs begin to slow down with your team. At first it will rise rapidly, and after several months it will slow to a crawl. Because it is an unknown, that is why I use the guideline. Otherwise I might never release my products. I've had products with 150k users generate 200 bugs (simple tools like 2D Editor) and another one generate 1500 bugs (that one was an OS). Bugs won't kill your sales, as long as you're fixing them, and even if you're not it is really unclear just how much it impacts sales.
I know that isn't really all that helpful, but a lot of it is really a mystery, to me at least. I can only make observations and suppositions from them. I have not done any hardcore testing of the mechanic, and maybe someone else has, but this is how I do mine. I don't time my releases based on bugs found, but rather by other software companies releases, but that is another topic completely.
From my example above, early 2D Editors require ~10 units of code and Audio Tools require slilghtly more, maybe 14 units. Targets using the guideline mentioned earlier would suggest that the 2D Editor will have around 500 total bugs, while the Audio Tool will have around 700 bugs. I released at 350 and 450 respectively for a total of 800 bugs found during Beta. My support team found an additional 500 (approximately) bugs between them for a grand total of 1300 bugs. The guideline told me I would have around 1200 bugs total between them, so it's pretty accurate.
It is not cost effective to catch all bugs (and I think not possible).
When catching bugs slows down considerably, I release the product.
And yes, users always find extra bugs.
I suppose you could fix all the bugs in Beta until it just completely stops, but it will probably take a very, very long time to happen. It's simply not cost effective to let it go on forever in Beta to achieve that goal. Generally speaking I keep a product in beta for 3-6 months depending on what kind of software it is, and when I have decided to release it based on the "Upcoming Releases" search for that type of product. Never release a product the same month as another competing product. If you can't beat them to market by a couple of months because your'e not ready, let it linger until 3 or 4 months after their release. I normally aim for May-June or November-December release dates. I also never announce a release date until I go into Beta, because changing your release date can be problematic.