Re:"BitC isn't going to work in its current form"
by _delirium on Saturday March 24, 2012 @04:17PM
(#3750584)
They have fairly different goals, though. BitC's main goal was to be a verifiable systems language, so it makes some sense that a lot of effort up front was put into investigating the "verifiable" part. Rather than building libraries and trying to get it to catch on, the effort was focused on trying to figure out if it could be designed in a way that simultaneously supported low-level, systems-ish programming while also supporting verification. If the answer on the verifiability part turned out to be no, then that would mean it wouldn't necessarily be worth building the rest (or trying to get people to use it), since that was the point of the project existing in the first place.
[ Parent
| on HN ]
Re:"BitC isn't going to work in its current form"
by tedunangst on Saturday March 24, 2012 @05:52PM
(#3750854)
Let's define success as new + useful. You can build something new then make it useful, or you can start with useful and add something new.
I suggest that the second approach is better, because we already have a good sense of how to measure useful, so you will always know where you are. Keep your eyes on the prize, as they say.
Re:"BitC isn't going to work in its current form"
by _delirium on Saturday March 24, 2012 @07:19PM
(#3751117)
If the goal was to make a C-like language that's verifiable, though, and you end up with a C-like language that isn't verifiable, in what sense is that useful? The whole usefulness of the project, at least as it was conceived, hinged on it adding some notion of verifiability to C. Otherwise it's just a C clone with smaller installed base, worse tool support, etc.
Re:"BitC isn't going to work in its current form"
by tedunangst on Saturday March 24, 2012 @10:46PM
(#3751543)
The whole usefulness of the project, at least as it was conceived, hinged on it adding some notion of verifiability to C.
That is exactly what I am saying. It sounds like they trimmed away too much of C in the process and didn't check to make sure those parts could be added back. I haven't really followed the whole saga though.
Re:"BitC isn't going to work in its current form"
by eli_gottlieb on Saturday March 24, 2012 @08:23PM
(#3751268)
There are degrees of verifiability. A C-like language with imperfect but stronger verifiability (for example, with a strong, formalized type system) is still more verifiable (admits the proof of more theorems via static analysis) than C.