There are quite a lot of reasons, but these come to mind first:
There's a shockingly large number of projects out there that use "borrowed" code. This is in the sense that they may use code from open source projects (and didn't distribute the code back as stipulated in the license), code that they've inherited from other companies that may or may not have been part of a legal acquirement or through the much heard "Corporate Espionage", which sounds sexier than it really is.
Usually, it's just an intern or someone else who finds out the default password to source control and/or CVS server and sends out a few feelers to someone interested. Next thing you know, he plugs in a thumb drive and now he's a free agent with a plane ticket and fancier watch.
Early games especially were like this since being first often meant being successful at courting investors (this was before the video game crash of '83).
Source: I once did a freelance job ages ago where the dev team manager, who used to work at a gaming company, freely admitted that it happened quite a lot in the industry. Maybe it's still happening.
Then there are cases where the source is outright appalling in the number of hacks, re-hacks, back-hacks and any number of ducktape and hope holding libraries together. No one of sound mind would ever let anyone else see that mess. Let alone give the impression that some of that code survives in some form in modern software.
Can you imagine how many vulnerabilities exist? For all we know, it may be possible to do a lot more than install Linux on a Gamecube. http://www.gc-linux.org/wiki/Main_Page
Of course there are also the times where "works of art", as it were, are lost to history. As jbri said, there are times where large projects often are destroyed for silly reasons like crashes or even recycling.
I've bought several used hard SCSI drives back in the day on eBay that may have belonged to one such company (SCSI was expensive, yo!) I have no idea what the project was or what I may have in my hands, but it looked like a whole lot of assembly, a bunch of esoteric C (which might as well be assembly) and some proprietary language code, that resembles a cross between Pascal and Python to my untrained eye, and I don't think anyone outside knew how or where to compile.
Now did they copy this stuff before selling it off? Who knows. This was after the dot com crash, so I have no idea if even the original owners were involved in the sale since many of those companies' assets were liquidated in a very short period of time. And we know how delicate they can be when money is at stake.
Then there are issues with licensing (kinda related to the first point) and already mentioned by people here. Open source, as samarudge said, is complicated. If your legalese isn't kosher, you can expect to lose a buttload of cash in a lawsuit(s) even if you win. Especially with the threat of patent trolls around, you can bet fewer companies are willing to risk OS-ing their code.
There's a shockingly large number of projects out there that use "borrowed" code. This is in the sense that they may use code from open source projects (and didn't distribute the code back as stipulated in the license), code that they've inherited from other companies that may or may not have been part of a legal acquirement or through the much heard "Corporate Espionage", which sounds sexier than it really is.
Usually, it's just an intern or someone else who finds out the default password to source control and/or CVS server and sends out a few feelers to someone interested. Next thing you know, he plugs in a thumb drive and now he's a free agent with a plane ticket and fancier watch.
Early games especially were like this since being first often meant being successful at courting investors (this was before the video game crash of '83).
Source: I once did a freelance job ages ago where the dev team manager, who used to work at a gaming company, freely admitted that it happened quite a lot in the industry. Maybe it's still happening.
Then there are cases where the source is outright appalling in the number of hacks, re-hacks, back-hacks and any number of ducktape and hope holding libraries together. No one of sound mind would ever let anyone else see that mess. Let alone give the impression that some of that code survives in some form in modern software.
Can you imagine how many vulnerabilities exist? For all we know, it may be possible to do a lot more than install Linux on a Gamecube. http://www.gc-linux.org/wiki/Main_Page
Of course there are also the times where "works of art", as it were, are lost to history. As jbri said, there are times where large projects often are destroyed for silly reasons like crashes or even recycling.
I've bought several used hard SCSI drives back in the day on eBay that may have belonged to one such company (SCSI was expensive, yo!) I have no idea what the project was or what I may have in my hands, but it looked like a whole lot of assembly, a bunch of esoteric C (which might as well be assembly) and some proprietary language code, that resembles a cross between Pascal and Python to my untrained eye, and I don't think anyone outside knew how or where to compile.
Now did they copy this stuff before selling it off? Who knows. This was after the dot com crash, so I have no idea if even the original owners were involved in the sale since many of those companies' assets were liquidated in a very short period of time. And we know how delicate they can be when money is at stake.
Then there are issues with licensing (kinda related to the first point) and already mentioned by people here. Open source, as samarudge said, is complicated. If your legalese isn't kosher, you can expect to lose a buttload of cash in a lawsuit(s) even if you win. Especially with the threat of patent trolls around, you can bet fewer companies are willing to risk OS-ing their code.