I wonder how does Software Foundations compare against other more specialized classics in the field like [1-3].
[1] https://mitpress.mit.edu/books/principles-model-checking
I've not read or worked through the other books (wow they are expensive), but I'm in general not very keen on model checking as an approach. Tools like Alloy or NuSMV can be useful when designing or investigating algorithms, but they always do a bounded search and can't be applied directly to implementations. I personally don't confer that much trustworthyness to systems that appeal to model checking tools for their correctness.
I've seen The Little Prover, but am somewhat wary of it. On the other hand, I enjoyed The Little Schemer very much. I might take a look at it next month.
There are model checkers that work directly on implementations, like NASA's JPF.
> I personally don't confer that much trustworthyness to systems that appeal to model checking tools for their correctness.
Given that the going rate for proofs is roughly 1-1.5 days per one line of code, which grows superlinearly with the code size, I think we have no choice...
In any event, pretty much all safety-critical software that human lives depend on (even many, many lives) is verified with model-checking in the best case. Use of proofs in verifying safety-critical systems is very rare, as even those domains can rarely afford it.
Also, remember that no one really needs 100% proof, as no other, non-software component in the system guarantees 100% reliability (not that model-checking can give us probability bounds).
Proofs are useful -- because of their total coverage -- when verifying components that will be used in many different projects, like compilers, kernels and libraries, and even then only when they are sufficiently small to be doable within, say, 5-10 years by a dedicated team. A one-time gigantic job will pay for itself by being used (and hopefully paid for) by many customers over many years.
I think I also have a different conception of proof than you're naming. You don't need to prove your entire program. 10 small proofs strung together can go an enormous distance with low weight.
I have a very small sub-algorithm that I might try proving in TLA+ just to see how it goes.
[1]: I couldn't find the original TLA+ proof online, but the author of the Raft protocol said that he struggled even with partial machine-checked proofs.
People prove specific algorithms (distributed, concurrent or sequential) with TLA+, but once the spec is of a real system, things get much, much harder to prove.
I think I disagree a bit wrt model checking, which is very useful in some specialized domains. I have had quite a lot of success with it when verifying complex real time systems, specified as timed automata. We then generated code from this specifications. It'd be challenging to use any other technique to achieve such an exhaustive verification.
Both [1-2] can be found online, and [2] is particularly enlightening! I followed a course by the authors, and implemented a big chunk of the book. It's really awesome to implement static analyzers by interpreting programs with a more abstract set of types.
That sounds like an excellent use of model checking tools. Most approaches I've seen involve creating a model of some system separate from the implementation, which I think is risky. Still better than nothing, by far.
I'll definitely have to check those books out now. Thanks for the recommendations!
Really? I know a lot of programmers that were excited for TLP and were very disappointed. I haven't even gotten a lukewarm review from anyone yet.
How did you organise reading such a book? Any recommendations?
I wonder if I should go Isabelle, Coq or Agda. The former is more suitable to my personal prover needs, as I like classical logic and nominal techniques, but I have a feeling that Coq is easier to get into because there is more literature. Agda is the cleanest system, but lacks proof automation.
What we did was meet once a week to discuss the reading and the exercises, or work through the exercises together. During the week we'd read the next few sections and try the exercises. Especially at first it was really slow and took a lot of effort as we weren't very comfortable with that form of logic (most of us had only previously taken an undergrad intro to discrete math class).
I personally like Isabelle very much, and if I were to start this with a new group of people I'd do it in Isabelle. Its Prover IDE in the form of Isabelle/jEdit is much nicer than Proof General or CoqIDE. I also think its presentation is a bit more immediately understandable for someone familiar with classical logic but not necessarily type theory. Isabelle/HOL is very natural from that perspective, imo.
I agree that Isabelle's jEdit integration is nice. But what would be a polished and up-to-date introduction? I think "Concrete Semantics" is a bit heavy going for a complete beginner. No detailed explanation of proof rules, tactics etc.
It assumes some knowledge of basic functional programming, but builds up proving-related jargon quite quickly through examples with some interspersed explanations.
It doesn't discuss structured proofs with Isar, and instead focuses on the tactical style, but I think this is an advantage - it's simpler that way, and structured proofs really only benefit human readers. They're somewhat confusing to write if you don't already have a good grasp of the entire system and what is possible.
The length of that tutorial was really intimidating to me at first, but it's quite approachable and isn't all that dense. It does take some time to work through it enough to become comfortable with the system.
What was really hampering my grasping Isabelle/HOL was the lack of a description of the overall structure of the Isabelle/HOL system. In particular, there wasn't really a good account that explains the distinction between the rules of HOL and the meta-rules of Isabelle, and how they relate. This is hand-waved away in the tutorial. Clearly the authors don't want to burden the beginner with this material. Maybe that's the right choice ... they have a lot of experience teaching this material.
A consequence of this choice is that the whole tutorial did not (and still doesn't as far as I can gather from a quick browse) include a complete listing of the logical rules (and meta rules)! As the structure of Isabelle/HOL and its (meta-)rules are not explicitly given, it becomes hard to see how tactics work. Under what conditions does it make sense to use tactic X or tactic Y or Z? Without saying what the rules are and how tactics work on them, how do you know how to proceed when Isabelle/HOL gets stuck?
If you stick only to the examples given in the tutorials and in "Concrete Semantics" this is not so much of a problem, but as soon as you work on your own proofs, you are your own. I found this to be a major stumbling block.
I wonder if Coq is better documented in this regard. Maybe not, ITPs are just complex beasts.
Another free book, using the Isabelle proof assistant, is http://concrete-semantics.org/. It's also very good.