UseAuto: Theory and Practice of Automation in Coq Proofs

In a machine-checked proof, every single detail has to be justified. This can result in huge proof scripts. Fortunately, Coq comes with a proof-search mechanism and with several decision procedures that enable the system to automatically synthesize simple pieces of proof. Automation is very powerful when set up appropriately. The purpose of this chapter is to explain the basics of working of automation. The chapter is organized in two parts. The first part focuses on a general mechanism called "proof search." In short, proof search consists in naively trying to apply lemmas and assumptions in all possible ways. The second part describes "decision procedures", which are tactics that are very good at solving proof obligations that fall in some particular fragment of the logic of Coq. Many of the examples used in this chapter consist of small lemmas that have been made up to illustrate particular aspects of automation. These examples are completely independent from the rest of the Software Foundations course. This chapter also contains some bigger examples which are used to explain how to use automation in realistic proofs. These examples are taken from other chapters of the course (mostly from STLC), and the proofs that we present make use of the tactics from the library LibTactics.v, which is presented in the chapter UseTactics.

Basic Features of Proof Search

The idea of proof search is to replace a sequence of tactics applying lemmas and assumptions with a call to a single tactic, for example auto. This form of proof automation saves a lot of effort. It typically leads to much shorter proof scripts, and to scripts that are typically more robust to change. If one makes a little change to a definition, a proof that exploits automation probably won't need to be modified at all. Of course, using too much automation is a bad idea. When a proof script no longer records the main arguments of a proof, it becomes difficult to fix it when it gets broken after a change in a definition. Overall, a reasonable use of automation is generally a big win, as it saves a lot of time both in building proof scripts and in subsequently maintaining those proof scripts.

Strength of Proof Search

We are going to study four proof-search tactics: auto, eauto, iauto and jauto. The tactics auto and eauto are builtin in Coq. The tactic iauto is a shorthand for the builtin tactic try solve [intuition eauto]. The tactic jauto is defined in the library LibTactics, and simply performs some preprocessing of the goal before calling eauto. The goal of this chapter is to explain the general principles of proof search and to give rule of thumbs for guessing which of the four tactics mentioned above is best suited for solving a given goal. Proof search is a compromise between efficiency and expressiveness, that is, a tradeoff between how complex goals the tactic can solve and how much time the tactic requires for terminating. The tactic auto builds proofs only by using the basic tactics reflexivity, assumption, and apply. The tactic eauto can also exploit eapply. The tactic jauto extends eauto by being able to open conjunctions and existentials that occur in the context. The tactic iauto is able to deal with conjunctions, disjunctions, and negation in a quite clever way; however it is not able to open existentials from the context. Also, iauto usually becomes very slow when the goal involves several disjunctions. Note that proof search tactics never perform any rewriting step (tactics rewrite, subst), nor any case analysis on an arbitrary data structure or predicate (tactics destruct and inversion), nor any proof by induction (tactic induction). So, proof search is really intended to automate the final steps from the various branches of a proof. It is not able to discover the overall structure of a proof.

Basics

The tactic auto is able to solve a goal that can be proved using a sequence of intros, apply, assumption, and reflexivity. Two examples follow. The first one shows the ability for auto to call reflexivity at any time. In fact, calling reflexivity is always the first thing that auto tries to do.

The second example illustrates a proof where a sequence of two calls to apply are needed. The goal is to prove that if Q n implies P n for any n and if Q n holds for any n, then P 2 holds.

We can ask auto to tell us what proof it came up with, by invoking info_auto in place of auto.

The tactic auto can invoke apply but not eapply. So, auto cannot exploit lemmas whose instantiation cannot be directly deduced from the proof goal. To exploit such lemmas, one needs to invoke the tactic eauto, which is able to call eapply. In the following example, the first hypothesis asserts that P n is true when Q m is true for some m, and the goal is to prove that Q 1 implies P 2. This implication follows direction from the hypothesis by instantiating m as the value 1. The following proof script shows that eauto successfully solves the goal, whereas auto is not able to do so.

Remark: Again, we can use info_eauto to see what proof eauto comes up with.

Conjunctions

So far, we've seen that eauto is stronger than auto in the sense that it can deal with eapply. In the same way, we are going to see how jauto and iauto are stronger than auto and eauto in the sense that they provide better support for conjunctions. The tactics auto and eauto can prove a goal of the form F ∧ F', where F and F' are two propositions, as soon as both F and F' can be proved in the current context. An example follows.

However, when an assumption is a conjunction, auto and eauto are not able to exploit this conjunction. It can be quite surprising at first that eauto can prove very complex goals but that it fails to prove that F ∧ F' implies F. The tactics iauto and jauto are able to decompose conjunctions from the context. Here is an example.

The tactic jauto is implemented by first calling a pre-processing tactic called jauto_set, and then calling eauto. So, to understand how jauto works, one can directly call the tactic jauto_set.

Next is a more involved goal that can be solved by iauto and jauto.

The strategy of iauto and jauto is to run a global analysis of the top-level conjunctions, and then call eauto. For this reason, those tactics are not good at dealing with conjunctions that occur as the conclusion of some universally quantified hypothesis. The following example illustrates a general weakness of Coq proof search mechanisms.

This situation is slightly disappointing, since automation is able to prove the following goal, which is very similar. The only difference is that the universal quantification has been distributed over the conjunction.

Disjunctions

The tactics auto and eauto can handle disjunctions that occur in the goal.

However, only iauto is able to automate reasoning on the disjunctions that appear in the context. For example, iauto can prove that F ∨ F' entails F' ∨ F.

More generally, iauto can deal with complex combinations of conjunctions, disjunctions, and negations. Here is an example.

However, the ability of iauto to automatically perform a case analysis on disjunctions comes with a downside: iauto may be very slow. If the context involves several hypotheses with disjunctions, iauto typically generates an exponential number of subgoals on which eauto is called. One major advantage of jauto compared with iauto is that it never spends time performing this kind of case analyses.

Existentials

The tactics eauto, iauto, and jauto can prove goals whose conclusion is an existential. For example, if the goal is ∃ x, f x, the tactic eauto introduces an existential variable, say ?25, in place of x. The remaining goal is f ?25, and eauto tries to solve this goal, allowing itself to instantiate ?25 with any appropriate value. For example, if an assumption f 2 is available, then the variable ?25 gets instantiated with 2 and the goal is solved, as shown below.

A major strength of jauto over the other proof search tactics is that it is able to exploit the existentially-quantified hypotheses, i.e., those of the form ∃ x, P.

Negation

The tactics auto and eauto suffer from some limitations with respect to the manipulation of negations, mostly related to the fact that negation, written ¬ P, is defined as P → False but that the unfolding of this definition is not performed automatically. Consider the following example.

For this reason, the tactics iauto and jauto systematically invoke unfold not in × as part of their pre-processing. So, they are able to solve the previous goal right away.

We will come back later on to the behavior of proof search with respect to the unfolding of definitions.

Equalities

Coq's proof-search feature is not good at exploiting equalities. It can do very basic operations, like exploiting reflexivity and symmetry, but that's about it. Here is a simple example that auto can solve, by first calling symmetry and then applying the hypothesis.

To automate more advanced reasoning on equalities, one should rather try to use the tactic congruence, which is presented at the end of this chapter in the "Decision Procedures" section.

How Proof Search Works

Search Depth

The tactic auto works as follows. It first tries to call reflexivity and assumption. If one of these calls solves the goal, the job is done. Otherwise auto tries to apply the most recently introduced assumption that can be applied to the goal without producing and error. This application produces subgoals. There are two possible cases. If the sugboals produced can be solved by a recursive call to auto, then the job is done. Otherwise, if this application produces at least one subgoal that auto cannot solve, then auto starts over by trying to apply the second most recently introduced assumption. It continues in a similar fashion until it finds a proof or until no assumption remains to be tried. It is very important to have a clear idea of the backtracking process involved in the execution of the auto tactic; otherwise its behavior can be quite puzzling. For example, auto is not able to solve the following triviality.

The reason auto fails to solve the goal is because there are too many conjunctions. If there had been only five of them, auto would have successfully solved the proof, but six is too many. The tactic auto limits the number of lemmas and hypotheses that can be applied in a proof, so as to ensure that the proof search eventually terminates. By default, the maximal number of steps is five. One can specify a different bound, writing for example auto 6 to search for a proof involving at most six steps. For example, auto 6 would solve the previous lemma. (Similarly, one can invoke eauto 6 or intuition eauto 6.) The argument n of auto n is called the "search depth." The tactic auto is simply defined as a shorthand for auto 5. The behavior of auto n can be summarized as follows. It first tries to solve the goal using reflexivity and assumption. If this fails, it tries to apply a hypothesis (or a lemma that has been registered in the hint database), and this application produces a number of sugoals. The tactic auto (n-1) is then called on each of those subgoals. If all the subgoals are solved, the job is completed, otherwise auto n tries to apply a different hypothesis. During the process, auto n calls auto (n-1), which in turn might call auto (n-2), and so on. The tactic auto 0 only tries reflexivity and assumption, and does not try to apply any lemma. Overall, this means that when the maximal number of steps allowed has been exceeded, the auto tactic stops searching and backtracks to try and investigate other paths. The following lemma admits a unique proof that involves exactly three steps. So, auto n proves this goal iff n is greater than three.

We can generalize the example by introducing an assumption asserting that P k is derivable from P (k-1) for all k, and keep the assumption P 0. The tactic auto, which is the same as auto 5, is able to derive P k for all values of k less than 5. For example, it can prove P 4.

However, to prove P 5, one needs to call at least auto 6.

Because auto looks for proofs at a limited depth, there are cases where auto can prove a goal F and can prove a goal F' but cannot prove F ∧ F'. In the following example, auto can prove P 4 but it is not able to prove P 4 ∧ P 4, because the splitting of the conjunction consumes one proof step. To prove the conjunction, one needs to increase the search depth, using at least auto 6.

Backtracking

In the previous section, we have considered proofs where at each step there was a unique assumption that auto could apply. In general, auto can have several choices at every step. The strategy of auto consists of trying all of the possibilities (using a depth-first search exploration). To illustrate how automation works, we are going to extend the previous example with an additional assumption asserting that P k is also derivable from P (k+1). Adding this hypothesis offers a new possibility that auto could consider at every step. There exists a special command that one can use for tracing all the steps that proof-search considers. To view such a trace, one should write debug eauto. (For some reason, the command debug auto does not exist, so we have to use the command debug eauto instead.)

The output message produced by debug eauto is as follows.

    depth=5 
    depth=4 apply H3
    depth=3 apply H3
    depth=3 exact H1

The depth indicates the value of n with which eauto n is called. The tactics shown in the message indicate that the first thing that eauto has tried to do is to apply H3. The effect of applying H3 is to replace the goal P 2 with the goal P 1. Then, again, H3 has been applied, changing the goal P 1 into P 0. At that point, the goal was exactly the hypothesis H1. It seems that eauto was quite lucky there, as it never even tried to use the hypothesis H2 at any time. The reason is that auto always tries to use the most recently introduced hypothesis first, and H3 is a more recent hypothesis than H2 in the goal. So, let's permute the hypotheses H2 and H3 and see what happens.

This time, the output message suggests that the proof search investigates many possibilities. Replacing debug eauto with info_eauto, we observe that the proof that eauto comes up with is actually not the simplest one. apply H2; apply H3; apply H3; apply H3; exact H1 This proof goes through the proof obligation P 3, even though it is not any useful. The following tree drawing describes all the goals that automation has been through.

    |5||4||3||2||1||0| -- below, tabulation indicates the depth

    [P 2]
    -> [P 3]
       -> [P 4]
          -> [P 5]
             -> [P 6]
                -> [P 7]
                -> [P 5] 
             -> [P 4]
                -> [P 5]
                -> [P 3] 
          --> [P 3]
             -> [P 4]
                -> [P 5]
                -> [P 3] 
             -> [P 2]
                -> [P 3]
                -> [P 1] 
       -> [P 2]
          -> [P 3]
             -> [P 4]
                -> [P 5]
                -> [P 3] 
             -> [P 2]
                -> [P 3]
                -> [P 1] 
          -> [P 1]
             -> [P 2]
                -> [P 3]
                -> [P 1] 
             -> [P 0]
                -> !! Done !! 

The first few lines read as follows. To prove P 2, eauto 5 has first tried to apply H2, producing the subgoal P 3. To solve it, eauto 4 has tried again to apply H2, producing the goal P 4. Similarly, the search goes through P 5, P 6 and P 7. When reaching P 7, the tactic eauto 0 is called but as it is not allowed to try and apply any lemma, it fails. So, we come back to the goal P 6, and try this time to apply hypothesis H3, producing the subgoal P 5. Here again, eauto 0 fails to solve this goal. The process goes on and on, until backtracking to P 3 and trying to apply H2 three times in a row, going through P 2 and P 1 and P 0. This search tree explains why eauto came up with a proof starting with apply H2.

Adding Hints

By default, auto (and eauto) only tries to apply the hypotheses that appear in the proof context. There are two possibilities for telling auto to exploit a lemma that have been proved previously: either adding the lemma as an assumption just before calling auto, or adding the lemma as a hint, so that it can be used by every calls to auto. The first possibility is useful to have auto exploit a lemma that only serves at this particular point. To add the lemma as hypothesis, one can type generalize mylemma; intros, or simply lets: mylemma (the latter requires LibTactics.v). The second possibility is useful for lemmas that need to be exploited several times. The syntax for adding a lemma as a hint is Hint Resolve mylemma. For example, the lemma asserting than any number is less than or equal to itself, ∀ x, x ≤ x, called Le.le_refl in the Coq standard library, can be added as a hint as follows.

A convenient shorthand for adding all the constructors of an inductive datatype as hints is the command Hint Constructors mydatatype. Warning: some lemmas, such as transitivity results, should not be added as hints as they would very badly affect the performance of proof search. The description of this problem and the presentation of a general work-around for transitivity lemmas appear further on.

Integration of Automation in Tactics

The library "LibTactics" introduces a convenient feature for invoking automation after calling a tactic. In short, it suffices to add the symbol star (×) to the name of a tactic. For example, apply× H is equivalent to apply H; auto_star, where auto_star is a tactic that can be defined as needed. By default, auto_star first tries to solve the goal using auto, and if this does not succeed then it tries to call jauto. Even though jauto is strictly stronger than auto, it makes sense to call auto first: when auto succeeds it may save a lot of time, and when auto fails to prove the goal, it fails very quickly. The definition of auto_star, which determines the meaning of the star symbol, can be modified whenever needed. Simply write: Ltac auto_star ::= a_new_definition. ]] Observe the use of ::= instead of :=, which indicates that the tactic is being rebound to a new definition. So, the default definition is as follows.

Nearly all standard Coq tactics and all the tactics from "LibTactics" can be called with a star symbol. For example, one can invoke subst×, destruct× H, inverts× H, lets× I: H x, specializes× H x, and so on... There are two notable exceptions. The tactic auto× is just another name for the tactic auto_star. And the tactic apply× H calls eapply H (or the more powerful applys H if needed), and then calls auto_star. Note that there is no eapply× H tactic, use apply× H instead. In large developments, it can be convenient to use two degrees of automation. Typically, one would use a fast tactic, like auto, and a slower but more powerful tactic, like jauto. To allow for a smooth coexistence of the two form of automation, LibTactics.v also defines a "tilde" version of tactics, like apply¬ H, destruct¬ H, subst¬, auto¬ and so on. The meaning of the tilde symbol is described by the auto_tilde tactic, whose default implementation is auto.

In the examples that follow, only auto_star is needed.

Examples of Use of Automation

Let's see how to use proof search in practice on the main theorems of the "Software Foundations" course, proving in particular results such as determinism, preservation and progress.

Determinism

Recall the original proof of the determinism lemma for the IMP language, shown below.

Exercise: rewrite this proof using auto whenever possible. (The solution uses auto 9 times.)

In fact, using automation is not just a matter of calling auto in place of one or two other tactics. Using automation is about rethinking the organization of sequences of tactics so as to minimize the effort involved in writing and maintaining the proof. This process is eased by the use of the tactics from LibTactics.v. So, before trying to optimize the way automation is used, let's first rewrite the proof of determinism:

• use introv H instead of intros x H,
• use gen x instead of generalize dependent x,
• use inverts H instead of inversion H; subst,
• use tryfalse to handle contradictions, and get rid of the cases where beval st b1 = true and beval st b1 = false both appear in the context,
• stop using ceval_cases to label subcases.

To obtain a nice clean proof script, we have to remove the calls assert (st' = st'0). Such a tactic invokation is not nice because it refers to some variables whose name has been automatically generated. This kind of tactics tend to be very brittle. The tactic assert (st' = st'0) is used to assert the conclusion that we want to derive from the induction hypothesis. So, rather than stating this conclusion explicitly, we are going to ask Coq to instantiate the induction hypothesis, using automation to figure out how to instantiate it. The tactic forwards, described in LibTactics.v precisely helps with instantiating a fact. So, let's see how it works out on our example.

To polish the proof script, it remains to factorize the calls to auto, using the star symbol. The proof of determinism can then be rewritten in only four lines, including no more than 10 tactics.

Preservation for STLC

Consider the proof of perservation of STLC, shown below. This proof already uses eauto through the triple-dot mechanism.

Exercise: rewrite this proof using tactics from LibTactics and calling automation using the star symbol rather than the triple-dot notation. More precisely, make use of the tactics inverts× and applys× to call auto× after a call to inverts or to applys. The solution is three lines long.

Progress for STLC

Consider the proof of the progress theorem.

Exercise: optimize the above proof. Hint: make use of destruct× and inverts×. The solution consists of 10 short lines.

BigStep and SmallStep

Consider the proof relating a small-step reduction judgment to a big-step reduction judgment.

Our goal is to optimize the above proof. It is generally easier to isolate inductions into separate lemmas. So, we are going to first prove an intermediate result that consists of the judgment over which the induction is being performed. Exercise: prove the following result, using tactics introv, induction and subst, and apply×. The solution is 3 lines long.

Exercise: using the lemma above, simplify the proof of the result multistep__eval. You should use the tactics introv, inverts, split× and apply×. The solution is 2 lines long.

If we try to combine the two proofs into a single one, we will likely fail, because of a limitation of the induction tactic. Indeed, this tactic looses information when applied to a predicate whose arguments are not reduced to variables, such as t ==>* (C n). You will thus need to use the more powerful tactic called dependent induction. This tactic is available only after importing the Program library, as shown below.

Exercise: prove the lemma multistep__eval without invoking the lemma multistep_eval_ind, that is, by inlining the proof by induction involved in multistep_eval_ind, using the tactic dependent induction instead of induction. The solution is 5 lines long.

Preservation for STLCRef

The proof of preservation for STLCRef can be found in chapter References. It contains 58 lines (not counting the labelling of cases). The optimized proof script is more than twice shorter. The following material explains how to build the optimized proof script. The resulting optimized proof script for the preservation theorem appears afterwards.

Let's come back to the proof case that was hard to optimize. The difficulty comes from the statement of nth_eq_snoc, which takes the form nth (length l) (snoc l x) d = x. This lemma is hard to exploit because its first argument, length l, mentions a list l that has to be exactly the same as the l occuring in snoc l x. In practice, the first argument is often a natural number n that is provably equal to length l yet that is not syntactically equal to length l. There is a simple fix for making nth_eq_snoc easy to apply: introduce the intermediate variable n explicitly, so that the goal becomes nth n (snoc l x) d = x, with a premise asserting n = length l.

The proof case for ref from the preservation theorem then becomes much easier to prove, because rewrite nth_eq_snoc' now succeeds.

The optimized proof of preservation is summarized next.

Progress for STLCRef

The proof of progress for STLCRef can be found in chapter References. It contains 53 lines and the optimized proof script is, here again, half the length.

Subtyping

Consider the inversion lemma for typing judgment of abstractions in a type system with subtyping.

Exercise: optimize the proof script, using introv, lets and inverts×. In particular, you will find it useful to replace the pattern apply K in H. destruct H as I with lets I: K H. The solution is 4 lines.

The lemma substitution_preserves_typing has already been used to illustrate the working of lets and applys in chapter UseTactics. Optimize further this proof using automation (with the star symbol), and using the tactic cases_if'. The solution is 33 lines, including the Case instructions (21 lines without them).

Advanced Topics in Proof Search

Stating Lemmas in the Right Way

Due to its depth-first strategy, eauto can get exponentially slower as the depth search increases, even when a short proof exists. In general, to make proof search run reasonably fast, one should avoid using a depth search greater than 5 or 6. Moreover, one should try to minimize the number of applicable lemmas, and usually put first the hypotheses whose proof usefully instantiates the existential variables. In fact, the ability for eauto to solve certain goals actually depends on the order in which the hypotheses are stated. This point is illustrated through the following example, in which P is a predicate on natural numbers. This predicate is such that P n holds for any n as soon as P m holds for at least one m different from zero. The goal is to prove that P 2 implies P 1. When the hypothesis about P is stated in the form ∀ n m, P m → m ≠ 0 → P n, then eauto works. However, with ∀ n m, m ≠ 0 → P m → P n, the tactic eauto fails.

It is very important to understand that the hypothesis ∀ n m, P m → m ≠ 0 → P n is eauto-friendly, whereas ∀ n m, m ≠ 0 → P m → P n really isn't. Guessing a value of m for which P m holds and then checking that m ≠ 0 holds works well because there are few values of m for which P m holds. So, it is likely that eauto comes up with the right one. On the other hand, guessing a value of m for which m ≠ 0 and then checking that P m holds does not work well, because there are many values of m that satisfy m ≠ 0 but not P m.

Unfolding of Definitions During Proof-Search

The use of intermediate definitions is generally encouraged in a formal development as it usually leads to more concise and more readable statements. Yet, definitions can make it a little harder to automate proofs. The problem is that it is not obvious for a proof search mechanism to know when definitions need to be unfolded. Note that a naive strategy that consists in unfolding all definitions before calling proof search does not scale up to large proofs, so we avoid it. This section introduces a few techniques for avoiding to manually unfold definitions before calling proof search. To illustrate the treatment of definitions, let P be an abstract predicate on natural numbers, and let myFact be a definition denoting the proposition P x holds for any x less than or equal to 3.

Proving that myFact under the assumption that P x holds for any x should be trivial. Yet, auto fails to prove it unless we unfold the definition of myFact explicitly.

To automate the unfolding of definitions that appear as proof obligation, one can use the command Hint Unfold myFact to tell Coq that it should always try to unfold myFact when myFact appears in the goal.

This time, automation is able to see through the definition of myFact.

However, the Hint Unfold mechanism only works for unfolding definitions that appear in the goal. In general, proof search does not unfold definitions from the context. For example, assume we want to prove that P 3 holds under the assumption that True → myFact.

There is actually one exception to the previous rule: a constant occuring in an hypothesis is automatically unfolded if the hypothesis can be directly applied to the current goal. For example, auto can prove myFact → P 3, as illustrated below.

Automation for Proving Absurd Goals

In this section, we'll see that lemmas concluding on a negation are generally not useful as hints, and that lemmas whose conclusion is False can be useful hints but having too many of them makes proof search inefficient. We'll also see a practical work-around to the efficiency issue. Consider the following lemma, which asserts that a number less than or equal to 3 is not greater than 3.

Equivalently, one could state that a number greater than three is not less than or equal to 3.

In fact, both statements are equivalent to a third one stating that x ≤ 3 and x > 3 are contradictory, in the sense that they imply False.

The following investigation aim at figuring out which of the three statments is the most convenient with respect to proof automation. The following material is enclosed inside a Section, so as to restrict the scope of the hints that we are adding. In other words, after the end of the section, the hints added within the section will no longer be active.

Let's try to add the first lemma, le_not_gt, as hint, and see whether we can prove that the proposition ∃ x, x ≤ 3 ∧ x > 3 is absurd.

The lemma gt_not_le is symmetric to le_not_gt, so it will not be any better. The third lemma, le_gt_false, is a more useful hint, because it concludes on False, so proof search will try to apply it when the current goal is False.

In summary, a lemma of the form H1 → H2 → False is a much more effective hint than H1 → ¬ H2, even though the two statments are equivalent up to the definition of the negation symbol ¬. That said, one should be careful with adding lemmas whose conclusion is False as hint. The reason is that whenever reaching the goal False, the proof search mechanism will potentially try to apply all the hints whose conclusion is False before applying the appropriate one.

Adding lemmas whose conclusion is False as hint can be, locally, a very effective solution. However, this approach does not scale up for global hints. For most practical applications, it is reasonable to give the name of the lemmas to be exploited for deriving a contradiction. The tactic false H, provided by LibTactics serves that purpose: false H replaces the goal with False and calls eapply H. Its behavior is described next. Observe that any of the three statements le_not_gt, gt_not_le or le_gt_false can be used.

In the above example, false le_gt_false; eauto proves the goal, but false le_gt_false; auto does not, because auto does not correctly instantiate the existential variable. Note that false× le_gt_false would not work either, because the star symbol tries to call auto first. So, there are two possibilities for completing the proof: either call false le_gt_false; eauto, or call false× (le_gt_false 3).

Automation for Transitivity Lemmas

Some lemmas should never be added as hints, because they would very badly slow down proof search. The typical example is that of transitivity results. This section describes the problem and presents a general workaround. Consider a subtyping relation, written subtype S T, that relates two object S and T of type typ. Assume that this relation has been proved reflexive and transitive. The corresponding lemmas are named subtype_refl and subtype_trans.

Adding reflexivity as hint is generally a good idea, so let's add reflexivity of subtyping as hint.

Adding transitivity as hint is generally a bad idea. To understand why, let's add it as hint and see what happens. Because we cannot remove hints once we've added them, we are going to open a "Section," so as to restrict the scope of the transitivity hint to that section.

Now, consider the goal ∀ S T, subtype S T, which clearly has no hope of being solved. Let's call eauto on this goal.

Note that after closing the section, the hint subtype_trans is no longer active.

In the previous example, the proof search has spent a lot of time trying to apply transitivity and reflexivity in every possible way. Its process can be summarized as follows. The first goal is subtype S T. Since reflexivity does not apply, eauto invokes transitivity, which produces two subgoals, subtype S ?X and subtype ?X T. Solving the first subgoal, subtype S ?X, is straightforward, it suffices to apply reflexivity. This unifies ?X with S. So, the second sugoal, subtype ?X T, becomes subtype S T, which is exactly what we started from... The problem with the transitivity lemma is that it is applicable to any goal concluding on a subtyping relation. Because of this, eauto keeps trying to apply it even though it most often doesn't help to solve the goal. So, one should never add a transitivity lemma as a hint for proof search. There is a general workaround for having automation to exploit transitivity lemmas without giving up on efficiency. This workaround relies on a powerful mechanism called "external hint." This mechanism allows to manually describe the condition under which a particular lemma should be tried out during proof search. For the case of transitivity of subtyping, we are going to tell Coq to try and apply the transitivity lemma on a goal of the form subtype S U only when the proof context already contains an assumption either of the form subtype S T or of the form subtype T U. In other words, we only apply the transitivity lemma when there is some evidence that this application might help. To set up this "external hint," one has to write the following.

This hint declaration can be understood as follows.

• "Hint Extern" introduces the hint.
• The number "1" corresponds to a priority for proof search. It doesn't matter so much what priority is used in practice.
• The pattern subtype ?S ?U describes the kind of goal on which the pattern should apply. The question marks are used to indicate that the variables ?S and ?U should be bound to some value in the rest of the hint description.
• The construction match goal with ... end tries to recognize patterns in the goal, or in the proof context, or both.
• The first pattern is H: subtype S ?T |- _. It indices that the context should contain an hypothesis H of type subtype S ?T, where S has to be the same as in the goal, and where ?T can have any value.
• The symbol |- _ at the end of H: subtype S ?T |- _ indicates that we do not impose further condition on how the proof obligation has to look like.
• The branch ⇒ apply (@subtype_trans S T U) that follows indicates that if the goal has the form subtype S U and if there exists an hypothesis of the form subtype S T, then we should try and apply transitivity lemma instantiated on the arguments S, T and U. (Note: the symbol @ in front of subtype_trans is only actually needed when the "Implicit Arguments" feature is activated.)
• The other branch, which corresponds to an hypothesis of the form H: subtype ?T U is symmetrical.

Note: the same external hint can be reused for any other transitive relation, simply by renaming subtype into the name of that relation. Let us see an example illustrating how the hint works.

We may also check that the new external hint does not suffer from the complexity blow up.

Decision Procedures

A decision procedure is able to solve proof obligations whose statement admits a particular form. This section describes three useful decision procedures. The tactic omega handles goals involving arithmetic and inequalities, but not general multiplications. The tactic ring handles goals involving arithmetic, including multiplications, but does not support inequalities. The tactic congruence is able to prove equalities and inequalities by exploiting equalities available in the proof context.

Omega

The tactic omega supports natural numbers (type nat) as well as integers (type Z, available by including the module ZArith). It supports addition, substraction, equalities and inequalities. Before using omega, one needs to import the module Omega, as follows.

Here is an example. Let x and y be two natural numbers (they cannot be negative). Assume y is less than 4, assume x+x+1 is less than y, and assume x is not zero. Then, it must be the case that x is equal to one.

Another example: if z is the mean of x and y, and if the difference between x and y is at most 4, then the difference between x and z is at most 2.

One can proof False using omega if the mathematical facts from the context are contradictory. In the following example, the constraints on the values x and y cannot be all satisfied in the same time.

Note: omega can prove a goal by contradiction only if its conclusion is reduced False. The tactic omega always fails when the conclusion is an arbitrary proposition P, even though False implies any proposition P (by ex_falso_quodlibet).

Ring

Compared with omega, the tactic ring adds support for multiplications, however it gives up the ability to reason on inequations. Moreover, it supports only integers (type Z) and not natural numbers (type nat). Here is an example showing how to use ring.

Congruence

The tactic congruence is able to exploit equalities from the proof context in order to automatically perform the rewriting operations necessary to establish a goal. It is slightly more powerful than the tactic subst, which can only handle equalities of the form x = e where x is a variable and e an expression.

Moreover, congruence is able to exploit universally quantified equalities, for example ∀ a, g a = h a.

Next is an example where congruence is very useful.

The tactic congruence is able to prove a contradiction if the goal entails an equality that contradicts an inequality available in the proof context.

One of the strengths of congruence is that it is a very fast tactic. So, one should not hesitate to invoke it wherever it might help.

Summary

Let us summarize the main automation tactics available.

• auto automatically applies reflexivity, assumption, and apply.
• eauto moreover tries eapply, and in particular can instantiate existentials in the conclusion.
• iauto extends eauto with support for negation, conjunctions, and disjunctions. However, its support for disjunction can make it exponentially slow.
• jauto extends eauto with support for negation, conjunctions, and existential at the head of hypothesis.
• congruence helps reasoning about equalities and inequalities.
• omega proves arithmetic goals with equalities and inequalities, but it does not support multiplication.
• ring proves arithmetic goals with multiplications, but does not support inequalities.

In order to set up automation appropriately, keep in mind the following rule of thumbs:

• automation is all about balance: not enough automation makes proofs not very robust on change, whereas too much automation makes proofs very hard to fix when they break.
• if a lemma is not goal directed (i.e., some of its variables do not occur in its conclusion), then the premises need to be ordered in such a way that proving the first premises maximizes the chances of correctly instantiating the variables that do not occur in the conclusion.
• a lemma whose conclusion is False should only be added as a local hint, i.e., as a hint within the current section.
• a transitivity lemma should never be considered as hint; if automation of transitivity reasoning is really necessary, an Extern Hint needs to be set up.
• a definition usually needs to be accompanied with a Hint Unfold.

Becoming a master in the black art of automation certainly requires some investment, however this investment will pay off very quickly.