[eluser]jedd[/eluser]
If you provide software, and do not have anything in writing, signed by the client, that says after 30 days it will stop working unless he pays you money - then you are still exposed if the software stops working in 30 days.
Here's how it will look from a legal point of view.
You have supplied software to this person. The person relies upon this software. You engineered it in such a way as to make it stop working. There is nothing in writing that makes that agreement clear - the only agreement that is clear is that you provided this software. I think, but again I emphasise that IANAL, that you are exposed in this scenario - primarily because of the lack of a written, signed agreement, and the implicit agreement that could be inferred by you providing the software in the first instance (which in turn implies certain things).
As has been mentioned, you're using PHP so your code is visible.
I've been pondering this since my first posting a few days ago, and I came to the conclusion that if, as a hypothetical exercise, I found myself in a similar situation, where I was not motivated enough to back out of the arrangement, but wary about proceeding, I might adjust the way that I developed the code.
Because there was a risk of not being paid, I might reasonably be inclined to spend less time working on the project - this means I'd likely not have time to do any in-line documentation. I'd also probably be more prone to making off-by-one and edge- and corner- case errors that wouldn't necessarily be found and fixed during initial testing. I'd probably, because I was eager to provide maximum performance for the client at minimal cost (to me) - and again this is quite a reasonable motivation for anyone in my position - try to run some automatic code optimisers. So, with the intent of reducing the file size and consequently improving load times, I'd not only rip out any existing wasteful comment lines, I'd then run the code through said optimiser. Sometimes they have names like 'php-obfuscator' - but their intent is clearly to maximise performance by trimming all your variable names to just one letter, reducing wasteful white-space, and so on. This would be quite a defensible move - as I say, you're attempting to optimise the performance of the code for the benefit of the client, in the cheapest way possible for you because you are concerned about payment.
I'd likely do these optimisations just after I'd forked the codebase, just before handing the code over to the client. If they ever wanted proper code optimisations, or indeed any modifications done to it in the future, they'd probably need to access the much slower and full-of-useless-comment-lines fork that I kept in my cvs.
At which point it quite rightly becomes a negotiation issue, not a technical one, as to what is needed for them to gain access to that codebase.
Just pondering out loud on this one.
